Merge branch 'master' into chulwoo-dec12-2015

Polymorphic lattices & various small updates

.gitignore

@@ -94,7 +94,7 @@ Thumbs.db
 
 # build directory #
 ###################
-build/*
+build*/*
 
 # IDE related files #
 #####################

.travis.yml

@@ -0,0 +1,79 @@
+language: cpp
+
+matrix:
+  include:
+    - os:        osx
+      osx_image: xcode7.2
+      compiler: clang
+    - os:        osx
+      osx_image: xcode7.2
+      compiler: gcc
+      env: VERSION=-5
+    - compiler: gcc
+      addons:
+        apt:
+          sources:
+            - ubuntu-toolchain-r-test
+          packages:
+            - g++-4.9
+            - libmpfr-dev
+            - libgmp-dev
+            - libmpc-dev
+            - binutils-dev
+      env: VERSION=-4.9
+    - compiler: gcc
+      addons:
+        apt:
+          sources:
+            - ubuntu-toolchain-r-test
+          packages:
+            - g++-5
+            - libmpfr-dev
+            - libgmp-dev
+            - libmpc-dev
+            - binutils-dev
+      env: VERSION=-5
+    - compiler: clang
+      addons:
+        apt:
+          sources:
+            - ubuntu-toolchain-r-test
+            - llvm-toolchain-precise-3.7
+          packages:
+            - clang-3.7
+            - libmpfr-dev
+            - libgmp-dev
+            - libmpc-dev
+            - binutils-dev
+      env: VERSION=-3.7
+    - compiler: clang
+      addons:
+        apt:
+          sources:
+            - ubuntu-toolchain-r-test
+            - llvm-toolchain-precise-3.8
+          packages:
+            - clang-3.8
+            - libmpfr-dev
+            - libgmp-dev
+            - libmpc-dev
+            - binutils-dev
+      env: VERSION=-3.8
+      
+before_install:
+    - if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then brew update; fi
+    - if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then brew install libmpc; fi
+    - if [[ "$TRAVIS_OS_NAME" == "osx" ]] && [[ "$CC" == "gcc" ]]; then brew install gcc5; fi
+    
+install:
+    - export CC=$CC$VERSION
+    - export CXX=$CXX$VERSION
+    - if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then export LDFLAGS='-L/usr/local/lib'; fi
+    
+script:
+    - ./scripts/reconfigure_script
+    - mkdir build
+    - cd build
+    - ../configure CXXFLAGS="-msse4.2 -O3 -std=c++11" LIBS="-lmpfr -lgmp" --enable-precision=single --enable-simd=SSE4 --enable-comms=none
+    - make -j4
+    - ./benchmarks/Benchmark_dwf --threads 1

AUTHORS

@@ -1,5 +1,4 @@
-Peter Boyle
-Azusa Yamaguchi
-Intel Parallel Computing Centre @ Higgs Centre for Theoretical Physics
-University of Edinburgh
-Scotland, UK
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: Peter Boyle <peterboyle@MacBook-Pro.local>
+Author: paboyle <paboyle@ph.ed.ac.uk>

COPYING

@@ -1,622 +1,281 @@
-                    GNU GENERAL PUBLIC LICENSE
-                       Version 3, 29 June 2007
+                   GNU GENERAL PUBLIC LICENSE
+                       Version 2, June 1991
 
- Copyright (C) 2007 Free Software Foundation, Inc. <http://fsf.org/>
+ Copyright (C) 1989, 1991 Free Software Foundation, Inc.,
+ 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  Everyone is permitted to copy and distribute verbatim copies
  of this license document, but changing it is not allowed.
 
                             Preamble
 
-  The GNU General Public License is a free, copyleft license for
-software and other kinds of works.
-
-  The licenses for most software and other practical works are designed
-to take away your freedom to share and change the works.  By contrast,
-the GNU General Public License is intended to guarantee your freedom to
-share and change all versions of a program--to make sure it remains free
-software for all its users.  We, the Free Software Foundation, use the
-GNU General Public License for most of our software; it applies also to
-any other work released this way by its authors.  You can apply it to
+  The licenses for most software are designed to take away your
+freedom to share and change it.  By contrast, the GNU General Public
+License is intended to guarantee your freedom to share and change free
+software--to make sure the software is free for all its users.  This
+General Public License applies to most of the Free Software
+Foundation's software and to any other program whose authors commit to
+using it.  (Some other Free Software Foundation software is covered by
+the GNU Lesser General Public License instead.)  You can apply it to
 your programs, too.
 
   When we speak of free software, we are referring to freedom, not
 price.  Our General Public Licenses are designed to make sure that you
 have the freedom to distribute copies of free software (and charge for
-them if you wish), that you receive source code or can get it if you
-want it, that you can change the software or use pieces of it in new
-free programs, and that you know you can do these things.
+this service if you wish), that you receive source code or can get it
+if you want it, that you can change the software or use pieces of it
+in new free programs; and that you know you can do these things.
 
-  To protect your rights, we need to prevent others from denying you
-these rights or asking you to surrender the rights.  Therefore, you have
-certain responsibilities if you distribute copies of the software, or if
-you modify it: responsibilities to respect the freedom of others.
+  To protect your rights, we need to make restrictions that forbid
+anyone to deny you these rights or to ask you to surrender the rights.
+These restrictions translate to certain responsibilities for you if you
+distribute copies of the software, or if you modify it.
 
   For example, if you distribute copies of such a program, whether
-gratis or for a fee, you must pass on to the recipients the same
-freedoms that you received.  You must make sure that they, too, receive
-or can get the source code.  And you must show them these terms so they
-know their rights.
+gratis or for a fee, you must give the recipients all the rights that
+you have.  You must make sure that they, too, receive or can get the
+source code.  And you must show them these terms so they know their
+rights.
 
-  Developers that use the GNU GPL protect your rights with two steps:
-(1) assert copyright on the software, and (2) offer you this License
-giving you legal permission to copy, distribute and/or modify it.
+  We protect your rights with two steps: (1) copyright the software, and
+(2) offer you this license which gives you legal permission to copy,
+distribute and/or modify the software.
 
-  For the developers' and authors' protection, the GPL clearly explains
-that there is no warranty for this free software.  For both users' and
-authors' sake, the GPL requires that modified versions be marked as
-changed, so that their problems will not be attributed erroneously to
-authors of previous versions.
+  Also, for each author's protection and ours, we want to make certain
+that everyone understands that there is no warranty for this free
+software.  If the software is modified by someone else and passed on, we
+want its recipients to know that what they have is not the original, so
+that any problems introduced by others will not reflect on the original
+authors' reputations.
 
-  Some devices are designed to deny users access to install or run
-modified versions of the software inside them, although the manufacturer
-can do so.  This is fundamentally incompatible with the aim of
-protecting users' freedom to change the software.  The systematic
-pattern of such abuse occurs in the area of products for individuals to
-use, which is precisely where it is most unacceptable.  Therefore, we
-have designed this version of the GPL to prohibit the practice for those
-products.  If such problems arise substantially in other domains, we
-stand ready to extend this provision to those domains in future versions
-of the GPL, as needed to protect the freedom of users.
-
-  Finally, every program is threatened constantly by software patents.
-States should not allow patents to restrict development and use of
-software on general-purpose computers, but in those that do, we wish to
-avoid the special danger that patents applied to a free program could
-make it effectively proprietary.  To prevent this, the GPL assures that
-patents cannot be used to render the program non-free.
+  Finally, any free program is threatened constantly by software
+patents.  We wish to avoid the danger that redistributors of a free
+program will individually obtain patent licenses, in effect making the
+program proprietary.  To prevent this, we have made it clear that any
+patent must be licensed for everyone's free use or not licensed at all.
 
   The precise terms and conditions for copying, distribution and
 modification follow.
 
-                       TERMS AND CONDITIONS
-
-  0. Definitions.
-
-  "This License" refers to version 3 of the GNU General Public License.
-
-  "Copyright" also means copyright-like laws that apply to other kinds of
-works, such as semiconductor masks.
-
-  "The Program" refers to any copyrightable work licensed under this
-License.  Each licensee is addressed as "you".  "Licensees" and
-"recipients" may be individuals or organizations.
-
-  To "modify" a work means to copy from or adapt all or part of the work
-in a fashion requiring copyright permission, other than the making of an
-exact copy.  The resulting work is called a "modified version" of the
-earlier work or a work "based on" the earlier work.
-
-  A "covered work" means either the unmodified Program or a work based
-on the Program.
-
-  To "propagate" a work means to do anything with it that, without
-permission, would make you directly or secondarily liable for
-infringement under applicable copyright law, except executing it on a
-computer or modifying a private copy.  Propagation includes copying,
-distribution (with or without modification), making available to the
-public, and in some countries other activities as well.
-
-  To "convey" a work means any kind of propagation that enables other
-parties to make or receive copies.  Mere interaction with a user through
-a computer network, with no transfer of a copy, is not conveying.
-
-  An interactive user interface displays "Appropriate Legal Notices"
-to the extent that it includes a convenient and prominently visible
-feature that (1) displays an appropriate copyright notice, and (2)
-tells the user that there is no warranty for the work (except to the
-extent that warranties are provided), that licensees may convey the
-work under this License, and how to view a copy of this License.  If
-the interface presents a list of user commands or options, such as a
-menu, a prominent item in the list meets this criterion.
-
-  1. Source Code.
-
-  The "source code" for a work means the preferred form of the work
-for making modifications to it.  "Object code" means any non-source
-form of a work.
-
-  A "Standard Interface" means an interface that either is an official
-standard defined by a recognized standards body, or, in the case of
-interfaces specified for a particular programming language, one that
-is widely used among developers working in that language.
-
-  The "System Libraries" of an executable work include anything, other
-than the work as a whole, that (a) is included in the normal form of
-packaging a Major Component, but which is not part of that Major
-Component, and (b) serves only to enable use of the work with that
-Major Component, or to implement a Standard Interface for which an
-implementation is available to the public in source code form.  A
-"Major Component", in this context, means a major essential component
-(kernel, window system, and so on) of the specific operating system
-(if any) on which the executable work runs, or a compiler used to
-produce the work, or an object code interpreter used to run it.
-
-  The "Corresponding Source" for a work in object code form means all
-the source code needed to generate, install, and (for an executable
-work) run the object code and to modify the work, including scripts to
-control those activities.  However, it does not include the work's
-System Libraries, or general-purpose tools or generally available free
-programs which are used unmodified in performing those activities but
-which are not part of the work.  For example, Corresponding Source
-includes interface definition files associated with source files for
-the work, and the source code for shared libraries and dynamically
-linked subprograms that the work is specifically designed to require,
-such as by intimate data communication or control flow between those
-subprograms and other parts of the work.
-
-  The Corresponding Source need not include anything that users
-can regenerate automatically from other parts of the Corresponding
-Source.
-
-  The Corresponding Source for a work in source code form is that
-same work.
-
-  2. Basic Permissions.
-
-  All rights granted under this License are granted for the term of
-copyright on the Program, and are irrevocable provided the stated
-conditions are met.  This License explicitly affirms your unlimited
-permission to run the unmodified Program.  The output from running a
-covered work is covered by this License only if the output, given its
-content, constitutes a covered work.  This License acknowledges your
-rights of fair use or other equivalent, as provided by copyright law.
-
-  You may make, run and propagate covered works that you do not
-convey, without conditions so long as your license otherwise remains
-in force.  You may convey covered works to others for the sole purpose
-of having them make modifications exclusively for you, or provide you
-with facilities for running those works, provided that you comply with
-the terms of this License in conveying all material for which you do
-not control copyright.  Those thus making or running the covered works
-for you must do so exclusively on your behalf, under your direction
-and control, on terms that prohibit them from making any copies of
-your copyrighted material outside their relationship with you.
-
-  Conveying under any other circumstances is permitted solely under
-the conditions stated below.  Sublicensing is not allowed; section 10
-makes it unnecessary.
-
-  3. Protecting Users' Legal Rights From Anti-Circumvention Law.
-
-  No covered work shall be deemed part of an effective technological
-measure under any applicable law fulfilling obligations under article
-11 of the WIPO copyright treaty adopted on 20 December 1996, or
-similar laws prohibiting or restricting circumvention of such
-measures.
-
-  When you convey a covered work, you waive any legal power to forbid
-circumvention of technological measures to the extent such circumvention
-is effected by exercising rights under this License with respect to
-the covered work, and you disclaim any intention to limit operation or
-modification of the work as a means of enforcing, against the work's
-users, your or third parties' legal rights to forbid circumvention of
-technological measures.
-
-  4. Conveying Verbatim Copies.
-
-  You may convey verbatim copies of the Program's source code as you
-receive it, in any medium, provided that you conspicuously and
-appropriately publish on each copy an appropriate copyright notice;
-keep intact all notices stating that this License and any
-non-permissive terms added in accord with section 7 apply to the code;
-keep intact all notices of the absence of any warranty; and give all
-recipients a copy of this License along with the Program.
-
-  You may charge any price or no price for each copy that you convey,
-and you may offer support or warranty protection for a fee.
-
-  5. Conveying Modified Source Versions.
-
-  You may convey a work based on the Program, or the modifications to
-produce it from the Program, in the form of source code under the
-terms of section 4, provided that you also meet all of these conditions:
-
-    a) The work must carry prominent notices stating that you modified
-    it, and giving a relevant date.
-
-    b) The work must carry prominent notices stating that it is
-    released under this License and any conditions added under section
-    7.  This requirement modifies the requirement in section 4 to
-    "keep intact all notices".
-
-    c) You must license the entire work, as a whole, under this
-    License to anyone who comes into possession of a copy.  This
-    License will therefore apply, along with any applicable section 7
-    additional terms, to the whole of the work, and all its parts,
-    regardless of how they are packaged.  This License gives no
-    permission to license the work in any other way, but it does not
-    invalidate such permission if you have separately received it.
-
-    d) If the work has interactive user interfaces, each must display
-    Appropriate Legal Notices; however, if the Program has interactive
-    interfaces that do not display Appropriate Legal Notices, your
-    work need not make them do so.
-
-  A compilation of a covered work with other separate and independent
-works, which are not by their nature extensions of the covered work,
-and which are not combined with it such as to form a larger program,
-in or on a volume of a storage or distribution medium, is called an
-"aggregate" if the compilation and its resulting copyright are not
-used to limit the access or legal rights of the compilation's users
-beyond what the individual works permit.  Inclusion of a covered work
-in an aggregate does not cause this License to apply to the other
-parts of the aggregate.
-
-  6. Conveying Non-Source Forms.
-
-  You may convey a covered work in object code form under the terms
-of sections 4 and 5, provided that you also convey the
-machine-readable Corresponding Source under the terms of this License,
-in one of these ways:
-
-    a) Convey the object code in, or embodied in, a physical product
-    (including a physical distribution medium), accompanied by the
-    Corresponding Source fixed on a durable physical medium
-    customarily used for software interchange.
-
-    b) Convey the object code in, or embodied in, a physical product
-    (including a physical distribution medium), accompanied by a
-    written offer, valid for at least three years and valid for as
-    long as you offer spare parts or customer support for that product
-    model, to give anyone who possesses the object code either (1) a
-    copy of the Corresponding Source for all the software in the
-    product that is covered by this License, on a durable physical
-    medium customarily used for software interchange, for a price no
-    more than your reasonable cost of physically performing this
-    conveying of source, or (2) access to copy the
-    Corresponding Source from a network server at no charge.
-
-    c) Convey individual copies of the object code with a copy of the
-    written offer to provide the Corresponding Source.  This
-    alternative is allowed only occasionally and noncommercially, and
-    only if you received the object code with such an offer, in accord
-    with subsection 6b.
-
-    d) Convey the object code by offering access from a designated
-    place (gratis or for a charge), and offer equivalent access to the
-    Corresponding Source in the same way through the same place at no
-    further charge.  You need not require recipients to copy the
-    Corresponding Source along with the object code.  If the place to
-    copy the object code is a network server, the Corresponding Source
-    may be on a different server (operated by you or a third party)
-    that supports equivalent copying facilities, provided you maintain
-    clear directions next to the object code saying where to find the
-    Corresponding Source.  Regardless of what server hosts the
-    Corresponding Source, you remain obligated to ensure that it is
-    available for as long as needed to satisfy these requirements.
-
-    e) Convey the object code using peer-to-peer transmission, provided
-    you inform other peers where the object code and Corresponding
-    Source of the work are being offered to the general public at no
-    charge under subsection 6d.
-
-  A separable portion of the object code, whose source code is excluded
-from the Corresponding Source as a System Library, need not be
-included in conveying the object code work.
-
-  A "User Product" is either (1) a "consumer product", which means any
-tangible personal property which is normally used for personal, family,
-or household purposes, or (2) anything designed or sold for incorporation
-into a dwelling.  In determining whether a product is a consumer product,
-doubtful cases shall be resolved in favor of coverage.  For a particular
-product received by a particular user, "normally used" refers to a
-typical or common use of that class of product, regardless of the status
-of the particular user or of the way in which the particular user
-actually uses, or expects or is expected to use, the product.  A product
-is a consumer product regardless of whether the product has substantial
-commercial, industrial or non-consumer uses, unless such uses represent
-the only significant mode of use of the product.
-
-  "Installation Information" for a User Product means any methods,
-procedures, authorization keys, or other information required to install
-and execute modified versions of a covered work in that User Product from
-a modified version of its Corresponding Source.  The information must
-suffice to ensure that the continued functioning of the modified object
-code is in no case prevented or interfered with solely because
-modification has been made.
-
-  If you convey an object code work under this section in, or with, or
-specifically for use in, a User Product, and the conveying occurs as
-part of a transaction in which the right of possession and use of the
-User Product is transferred to the recipient in perpetuity or for a
-fixed term (regardless of how the transaction is characterized), the
-Corresponding Source conveyed under this section must be accompanied
-by the Installation Information.  But this requirement does not apply
-if neither you nor any third party retains the ability to install
-modified object code on the User Product (for example, the work has
-been installed in ROM).
-
-  The requirement to provide Installation Information does not include a
-requirement to continue to provide support service, warranty, or updates
-for a work that has been modified or installed by the recipient, or for
-the User Product in which it has been modified or installed.  Access to a
-network may be denied when the modification itself materially and
-adversely affects the operation of the network or violates the rules and
-protocols for communication across the network.
-
-  Corresponding Source conveyed, and Installation Information provided,
-in accord with this section must be in a format that is publicly
-documented (and with an implementation available to the public in
-source code form), and must require no special password or key for
-unpacking, reading or copying.
-
-  7. Additional Terms.
-
-  "Additional permissions" are terms that supplement the terms of this
-License by making exceptions from one or more of its conditions.
-Additional permissions that are applicable to the entire Program shall
-be treated as though they were included in this License, to the extent
-that they are valid under applicable law.  If additional permissions
-apply only to part of the Program, that part may be used separately
-under those permissions, but the entire Program remains governed by
-this License without regard to the additional permissions.
-
-  When you convey a copy of a covered work, you may at your option
-remove any additional permissions from that copy, or from any part of
-it.  (Additional permissions may be written to require their own
-removal in certain cases when you modify the work.)  You may place
-additional permissions on material, added by you to a covered work,
-for which you have or can give appropriate copyright permission.
-
-  Notwithstanding any other provision of this License, for material you
-add to a covered work, you may (if authorized by the copyright holders of
-that material) supplement the terms of this License with terms:
-
-    a) Disclaiming warranty or limiting liability differently from the
-    terms of sections 15 and 16 of this License; or
-
-    b) Requiring preservation of specified reasonable legal notices or
-    author attributions in that material or in the Appropriate Legal
-    Notices displayed by works containing it; or
-
-    c) Prohibiting misrepresentation of the origin of that material, or
-    requiring that modified versions of such material be marked in
-    reasonable ways as different from the original version; or
-
-    d) Limiting the use for publicity purposes of names of licensors or
-    authors of the material; or
-
-    e) Declining to grant rights under trademark law for use of some
-    trade names, trademarks, or service marks; or
-
-    f) Requiring indemnification of licensors and authors of that
-    material by anyone who conveys the material (or modified versions of
-    it) with contractual assumptions of liability to the recipient, for
-    any liability that these contractual assumptions directly impose on
-    those licensors and authors.
-
-  All other non-permissive additional terms are considered "further
-restrictions" within the meaning of section 10.  If the Program as you
-received it, or any part of it, contains a notice stating that it is
-governed by this License along with a term that is a further
-restriction, you may remove that term.  If a license document contains
-a further restriction but permits relicensing or conveying under this
-License, you may add to a covered work material governed by the terms
-of that license document, provided that the further restriction does
-not survive such relicensing or conveying.
-
-  If you add terms to a covered work in accord with this section, you
-must place, in the relevant source files, a statement of the
-additional terms that apply to those files, or a notice indicating
-where to find the applicable terms.
-
-  Additional terms, permissive or non-permissive, may be stated in the
-form of a separately written license, or stated as exceptions;
-the above requirements apply either way.
-
-  8. Termination.
-
-  You may not propagate or modify a covered work except as expressly
-provided under this License.  Any attempt otherwise to propagate or
-modify it is void, and will automatically terminate your rights under
-this License (including any patent licenses granted under the third
-paragraph of section 11).
-
-  However, if you cease all violation of this License, then your
-license from a particular copyright holder is reinstated (a)
-provisionally, unless and until the copyright holder explicitly and
-finally terminates your license, and (b) permanently, if the copyright
-holder fails to notify you of the violation by some reasonable means
-prior to 60 days after the cessation.
-
-  Moreover, your license from a particular copyright holder is
-reinstated permanently if the copyright holder notifies you of the
-violation by some reasonable means, this is the first time you have
-received notice of violation of this License (for any work) from that
-copyright holder, and you cure the violation prior to 30 days after
-your receipt of the notice.
-
-  Termination of your rights under this section does not terminate the
-licenses of parties who have received copies or rights from you under
-this License.  If your rights have been terminated and not permanently
-reinstated, you do not qualify to receive new licenses for the same
-material under section 10.
-
-  9. Acceptance Not Required for Having Copies.
-
-  You are not required to accept this License in order to receive or
-run a copy of the Program.  Ancillary propagation of a covered work
-occurring solely as a consequence of using peer-to-peer transmission
-to receive a copy likewise does not require acceptance.  However,
-nothing other than this License grants you permission to propagate or
-modify any covered work.  These actions infringe copyright if you do
-not accept this License.  Therefore, by modifying or propagating a
-covered work, you indicate your acceptance of this License to do so.
-
-  10. Automatic Licensing of Downstream Recipients.
-
-  Each time you convey a covered work, the recipient automatically
-receives a license from the original licensors, to run, modify and
-propagate that work, subject to this License.  You are not responsible
-for enforcing compliance by third parties with this License.
-
-  An "entity transaction" is a transaction transferring control of an
-organization, or substantially all assets of one, or subdividing an
-organization, or merging organizations.  If propagation of a covered
-work results from an entity transaction, each party to that
-transaction who receives a copy of the work also receives whatever
-licenses to the work the party's predecessor in interest had or could
-give under the previous paragraph, plus a right to possession of the
-Corresponding Source of the work from the predecessor in interest, if
-the predecessor has it or can get it with reasonable efforts.
-
-  You may not impose any further restrictions on the exercise of the
-rights granted or affirmed under this License.  For example, you may
-not impose a license fee, royalty, or other charge for exercise of
-rights granted under this License, and you may not initiate litigation
-(including a cross-claim or counterclaim in a lawsuit) alleging that
-any patent claim is infringed by making, using, selling, offering for
-sale, or importing the Program or any portion of it.
-
-  11. Patents.
-
-  A "contributor" is a copyright holder who authorizes use under this
-License of the Program or a work on which the Program is based.  The
-work thus licensed is called the contributor's "contributor version".
-
-  A contributor's "essential patent claims" are all patent claims
-owned or controlled by the contributor, whether already acquired or
-hereafter acquired, that would be infringed by some manner, permitted
-by this License, of making, using, or selling its contributor version,
-but do not include claims that would be infringed only as a
-consequence of further modification of the contributor version.  For
-purposes of this definition, "control" includes the right to grant
-patent sublicenses in a manner consistent with the requirements of
+                    GNU GENERAL PUBLIC LICENSE
+   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
+
+  0. This License applies to any program or other work which contains
+a notice placed by the copyright holder saying it may be distributed
+under the terms of this General Public License.  The "Program", below,
+refers to any such program or work, and a "work based on the Program"
+means either the Program or any derivative work under copyright law:
+that is to say, a work containing the Program or a portion of it,
+either verbatim or with modifications and/or translated into another
+language.  (Hereinafter, translation is included without limitation in
+the term "modification".)  Each licensee is addressed as "you".
+
+Activities other than copying, distribution and modification are not
+covered by this License; they are outside its scope.  The act of
+running the Program is not restricted, and the output from the Program
+is covered only if its contents constitute a work based on the
+Program (independent of having been made by running the Program).
+Whether that is true depends on what the Program does.
+
+  1. You may copy and distribute verbatim copies of the Program's
+source code as you receive it, in any medium, provided that you
+conspicuously and appropriately publish on each copy an appropriate
+copyright notice and disclaimer of warranty; keep intact all the
+notices that refer to this License and to the absence of any warranty;
+and give any other recipients of the Program a copy of this License
+along with the Program.
+
+You may charge a fee for the physical act of transferring a copy, and
+you may at your option offer warranty protection in exchange for a fee.
+
+  2. You may modify your copy or copies of the Program or any portion
+of it, thus forming a work based on the Program, and copy and
+distribute such modifications or work under the terms of Section 1
+above, provided that you also meet all of these conditions:
+
+    a) You must cause the modified files to carry prominent notices
+    stating that you changed the files and the date of any change.
+
+    b) You must cause any work that you distribute or publish, that in
+    whole or in part contains or is derived from the Program or any
+    part thereof, to be licensed as a whole at no charge to all third
+    parties under the terms of this License.
+
+    c) If the modified program normally reads commands interactively
+    when run, you must cause it, when started running for such
+    interactive use in the most ordinary way, to print or display an
+    announcement including an appropriate copyright notice and a
+    notice that there is no warranty (or else, saying that you provide
+    a warranty) and that users may redistribute the program under
+    these conditions, and telling the user how to view a copy of this
+    License.  (Exception: if the Program itself is interactive but
+    does not normally print such an announcement, your work based on
+    the Program is not required to print an announcement.)
+
+These requirements apply to the modified work as a whole.  If
+identifiable sections of that work are not derived from the Program,
+and can be reasonably considered independent and separate works in
+themselves, then this License, and its terms, do not apply to those
+sections when you distribute them as separate works.  But when you
+distribute the same sections as part of a whole which is a work based
+on the Program, the distribution of the whole must be on the terms of
+this License, whose permissions for other licensees extend to the
+entire whole, and thus to each and every part regardless of who wrote it.
+
+Thus, it is not the intent of this section to claim rights or contest
+your rights to work written entirely by you; rather, the intent is to
+exercise the right to control the distribution of derivative or
+collective works based on the Program.
+
+In addition, mere aggregation of another work not based on the Program
+with the Program (or with a work based on the Program) on a volume of
+a storage or distribution medium does not bring the other work under
+the scope of this License.
+
+  3. You may copy and distribute the Program (or a work based on it,
+under Section 2) in object code or executable form under the terms of
+Sections 1 and 2 above provided that you also do one of the following:
+
+    a) Accompany it with the complete corresponding machine-readable
+    source code, which must be distributed under the terms of Sections
+    1 and 2 above on a medium customarily used for software interchange; or,
+
+    b) Accompany it with a written offer, valid for at least three
+    years, to give any third party, for a charge no more than your
+    cost of physically performing source distribution, a complete
+    machine-readable copy of the corresponding source code, to be
+    distributed under the terms of Sections 1 and 2 above on a medium
+    customarily used for software interchange; or,
+
+    c) Accompany it with the information you received as to the offer
+    to distribute corresponding source code.  (This alternative is
+    allowed only for noncommercial distribution and only if you
+    received the program in object code or executable form with such
+    an offer, in accord with Subsection b above.)
+
+The source code for a work means the preferred form of the work for
+making modifications to it.  For an executable work, complete source
+code means all the source code for all modules it contains, plus any
+associated interface definition files, plus the scripts used to
+control compilation and installation of the executable.  However, as a
+special exception, the source code distributed need not include
+anything that is normally distributed (in either source or binary
+form) with the major components (compiler, kernel, and so on) of the
+operating system on which the executable runs, unless that component
+itself accompanies the executable.
+
+If distribution of executable or object code is made by offering
+access to copy from a designated place, then offering equivalent
+access to copy the source code from the same place counts as
+distribution of the source code, even though third parties are not
+compelled to copy the source along with the object code.
+
+  4. You may not copy, modify, sublicense, or distribute the Program
+except as expressly provided under this License.  Any attempt
+otherwise to copy, modify, sublicense or distribute the Program is
+void, and will automatically terminate your rights under this License.
+However, parties who have received copies, or rights, from you under
+this License will not have their licenses terminated so long as such
+parties remain in full compliance.
+
+  5. You are not required to accept this License, since you have not
+signed it.  However, nothing else grants you permission to modify or
+distribute the Program or its derivative works.  These actions are
+prohibited by law if you do not accept this License.  Therefore, by
+modifying or distributing the Program (or any work based on the
+Program), you indicate your acceptance of this License to do so, and
+all its terms and conditions for copying, distributing or modifying
+the Program or works based on it.
+
+  6. Each time you redistribute the Program (or any work based on the
+Program), the recipient automatically receives a license from the
+original licensor to copy, distribute or modify the Program subject to
+these terms and conditions.  You may not impose any further
+restrictions on the recipients' exercise of the rights granted herein.
+You are not responsible for enforcing compliance by third parties to
 this License.
 
-  Each contributor grants you a non-exclusive, worldwide, royalty-free
-patent license under the contributor's essential patent claims, to
-make, use, sell, offer for sale, import and otherwise run, modify and
-propagate the contents of its contributor version.
-
-  In the following three paragraphs, a "patent license" is any express
-agreement or commitment, however denominated, not to enforce a patent
-(such as an express permission to practice a patent or covenant not to
-sue for patent infringement).  To "grant" such a patent license to a
-party means to make such an agreement or commitment not to enforce a
-patent against the party.
-
-  If you convey a covered work, knowingly relying on a patent license,
-and the Corresponding Source of the work is not available for anyone
-to copy, free of charge and under the terms of this License, through a
-publicly available network server or other readily accessible means,
-then you must either (1) cause the Corresponding Source to be so
-available, or (2) arrange to deprive yourself of the benefit of the
-patent license for this particular work, or (3) arrange, in a manner
-consistent with the requirements of this License, to extend the patent
-license to downstream recipients.  "Knowingly relying" means you have
-actual knowledge that, but for the patent license, your conveying the
-covered work in a country, or your recipient's use of the covered work
-in a country, would infringe one or more identifiable patents in that
-country that you have reason to believe are valid.
-
-  If, pursuant to or in connection with a single transaction or
-arrangement, you convey, or propagate by procuring conveyance of, a
-covered work, and grant a patent license to some of the parties
-receiving the covered work authorizing them to use, propagate, modify
-or convey a specific copy of the covered work, then the patent license
-you grant is automatically extended to all recipients of the covered
-work and works based on it.
-
-  A patent license is "discriminatory" if it does not include within
-the scope of its coverage, prohibits the exercise of, or is
-conditioned on the non-exercise of one or more of the rights that are
-specifically granted under this License.  You may not convey a covered
-work if you are a party to an arrangement with a third party that is
-in the business of distributing software, under which you make payment
-to the third party based on the extent of your activity of conveying
-the work, and under which the third party grants, to any of the
-parties who would receive the covered work from you, a discriminatory
-patent license (a) in connection with copies of the covered work
-conveyed by you (or copies made from those copies), or (b) primarily
-for and in connection with specific products or compilations that
-contain the covered work, unless you entered into that arrangement,
-or that patent license was granted, prior to 28 March 2007.
-
-  Nothing in this License shall be construed as excluding or limiting
-any implied license or other defenses to infringement that may
-otherwise be available to you under applicable patent law.
-
-  12. No Surrender of Others' Freedom.
-
-  If conditions are imposed on you (whether by court order, agreement or
+  7. If, as a consequence of a court judgment or allegation of patent
+infringement or for any other reason (not limited to patent issues),
+conditions are imposed on you (whether by court order, agreement or
 otherwise) that contradict the conditions of this License, they do not
-excuse you from the conditions of this License.  If you cannot convey a
-covered work so as to satisfy simultaneously your obligations under this
-License and any other pertinent obligations, then as a consequence you may
-not convey it at all.  For example, if you agree to terms that obligate you
-to collect a royalty for further conveying from those to whom you convey
-the Program, the only way you could satisfy both those terms and this
-License would be to refrain entirely from conveying the Program.
+excuse you from the conditions of this License.  If you cannot
+distribute so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you
+may not distribute the Program at all.  For example, if a patent
+license would not permit royalty-free redistribution of the Program by
+all those who receive copies directly or indirectly through you, then
+the only way you could satisfy both it and this License would be to
+refrain entirely from distribution of the Program.
 
-  13. Use with the GNU Affero General Public License.
+If any portion of this section is held invalid or unenforceable under
+any particular circumstance, the balance of the section is intended to
+apply and the section as a whole is intended to apply in other
+circumstances.
 
-  Notwithstanding any other provision of this License, you have
-permission to link or combine any covered work with a work licensed
-under version 3 of the GNU Affero General Public License into a single
-combined work, and to convey the resulting work.  The terms of this
-License will continue to apply to the part which is the covered work,
-but the special requirements of the GNU Affero General Public License,
-section 13, concerning interaction through a network will apply to the
-combination as such.
+It is not the purpose of this section to induce you to infringe any
+patents or other property right claims or to contest validity of any
+such claims; this section has the sole purpose of protecting the
+integrity of the free software distribution system, which is
+implemented by public license practices.  Many people have made
+generous contributions to the wide range of software distributed
+through that system in reliance on consistent application of that
+system; it is up to the author/donor to decide if he or she is willing
+to distribute software through any other system and a licensee cannot
+impose that choice.
 
-  14. Revised Versions of this License.
+This section is intended to make thoroughly clear what is believed to
+be a consequence of the rest of this License.
 
-  The Free Software Foundation may publish revised and/or new versions of
-the GNU General Public License from time to time.  Such new versions will
+  8. If the distribution and/or use of the Program is restricted in
+certain countries either by patents or by copyrighted interfaces, the
+original copyright holder who places the Program under this License
+may add an explicit geographical distribution limitation excluding
+those countries, so that distribution is permitted only in or among
+countries not thus excluded.  In such case, this License incorporates
+the limitation as if written in the body of this License.
+
+  9. The Free Software Foundation may publish revised and/or new versions
+of the General Public License from time to time.  Such new versions will
 be similar in spirit to the present version, but may differ in detail to
 address new problems or concerns.
 
-  Each version is given a distinguishing version number.  If the
-Program specifies that a certain numbered version of the GNU General
-Public License "or any later version" applies to it, you have the
-option of following the terms and conditions either of that numbered
-version or of any later version published by the Free Software
-Foundation.  If the Program does not specify a version number of the
-GNU General Public License, you may choose any version ever published
-by the Free Software Foundation.
+Each version is given a distinguishing version number.  If the Program
+specifies a version number of this License which applies to it and "any
+later version", you have the option of following the terms and conditions
+either of that version or of any later version published by the Free
+Software Foundation.  If the Program does not specify a version number of
+this License, you may choose any version ever published by the Free Software
+Foundation.
 
-  If the Program specifies that a proxy can decide which future
-versions of the GNU General Public License can be used, that proxy's
-public statement of acceptance of a version permanently authorizes you
-to choose that version for the Program.
+  10. If you wish to incorporate parts of the Program into other free
+programs whose distribution conditions are different, write to the author
+to ask for permission.  For software which is copyrighted by the Free
+Software Foundation, write to the Free Software Foundation; we sometimes
+make exceptions for this.  Our decision will be guided by the two goals
+of preserving the free status of all derivatives of our free software and
+of promoting the sharing and reuse of software generally.
 
-  Later license versions may give you additional or different
-permissions.  However, no additional obligations are imposed on any
-author or copyright holder as a result of your choosing to follow a
-later version.
+                            NO WARRANTY
 
-  15. Disclaimer of Warranty.
+  11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
+FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
+OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
+PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
+OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
+TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
+PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
+REPAIR OR CORRECTION.
 
-  THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
-APPLICABLE LAW.  EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
-HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
-OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
-THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
-PURPOSE.  THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
-IS WITH YOU.  SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
-ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
-
-  16. Limitation of Liability.
-
-  IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
-WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
-THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
-GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
-USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
-DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
-PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
-EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
-SUCH DAMAGES.
-
-  17. Interpretation of Sections 15 and 16.
-
-  If the disclaimer of warranty and limitation of liability provided
-above cannot be given local legal effect according to their terms,
-reviewing courts shall apply local law that most closely approximates
-an absolute waiver of all civil liability in connection with the
-Program, unless a warranty or assumption of liability accompanies a
-copy of the Program in return for a fee.
+  12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
+REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
+INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
+OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
+TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
+YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
+PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGES.
 
                      END OF TERMS AND CONDITIONS
 
@@ -628,15 +287,15 @@ free software which everyone can redistribute and change under these terms.
 
   To do so, attach the following notices to the program.  It is safest
 to attach them to the start of each source file to most effectively
-state the exclusion of warranty; and each file should have at least
+convey the exclusion of warranty; and each file should have at least
 the "copyright" line and a pointer to where the full notice is found.
 
     <one line to give the program's name and a brief idea of what it does.>
     Copyright (C) <year>  <name of author>
 
-    This program is free software: you can redistribute it and/or modify
+    This program is free software; you can redistribute it and/or modify
     it under the terms of the GNU General Public License as published by
-    the Free Software Foundation, either version 3 of the License, or
+    the Free Software Foundation; either version 2 of the License, or
     (at your option) any later version.
 
     This program is distributed in the hope that it will be useful,
@@ -644,31 +303,38 @@ the "copyright" line and a pointer to where the full notice is found.
     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
     GNU General Public License for more details.
 
-    You should have received a copy of the GNU General Public License
-    along with this program.  If not, see <http://www.gnu.org/licenses/>.
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
 
 Also add information on how to contact you by electronic and paper mail.
 
-  If the program does terminal interaction, make it output a short
-notice like this when it starts in an interactive mode:
+If the program is interactive, make it output a short notice like this
+when it starts in an interactive mode:
 
-    <program>  Copyright (C) <year>  <name of author>
-    This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    Gnomovision version 69, Copyright (C) year name of author
+    Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
     This is free software, and you are welcome to redistribute it
     under certain conditions; type `show c' for details.
 
 The hypothetical commands `show w' and `show c' should show the appropriate
-parts of the General Public License.  Of course, your program's commands
-might be different; for a GUI interface, you would use an "about box".
+parts of the General Public License.  Of course, the commands you use may
+be called something other than `show w' and `show c'; they could even be
+mouse-clicks or menu items--whatever suits your program.
 
-  You should also get your employer (if you work as a programmer) or school,
-if any, to sign a "copyright disclaimer" for the program, if necessary.
-For more information on this, and how to apply and follow the GNU GPL, see
-<http://www.gnu.org/licenses/>.
+You should also get your employer (if you work as a programmer) or your
+school, if any, to sign a "copyright disclaimer" for the program, if
+necessary.  Here is a sample; alter the names:
+
+  Yoyodyne, Inc., hereby disclaims all copyright interest in the program
+  `Gnomovision' (which makes passes at compilers) written by James Hacker.
+
+  <signature of Ty Coon>, 1 April 1989
+  Ty Coon, President of Vice
+
+This General Public License does not permit incorporating your program into
+proprietary programs.  If your program is a subroutine library, you may
+consider it more useful to permit linking proprietary applications with the
+library.  If this is what you want to do, use the GNU Lesser General
+Public License instead of this License.
 
-  The GNU General Public License does not permit incorporating your program
-into proprietary programs.  If your program is a subroutine library, you
-may consider it more useful to permit linking proprietary applications with
-the library.  If this is what you want to do, use the GNU Lesser General
-Public License instead of this License.  But first, please read
-<http://www.gnu.org/philosophy/why-not-lgpl.html>.

LICENSE

@@ -1,7 +1,7 @@
-GNU GENERAL PUBLIC LICENSE
+                   GNU GENERAL PUBLIC LICENSE
                        Version 2, June 1991
 
- Copyright (C) 1989, 1991 Free Software Foundation, Inc., <http://fsf.org/>
+ Copyright (C) 1989, 1991 Free Software Foundation, Inc.,
  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  Everyone is permitted to copy and distribute verbatim copies
  of this license document, but changing it is not allowed.
@@ -290,8 +290,8 @@ to attach them to the start of each source file to most effectively
 convey the exclusion of warranty; and each file should have at least
 the "copyright" line and a pointer to where the full notice is found.
 
-    {description}
-    Copyright (C) {year}  {fullname}
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
 
     This program is free software; you can redistribute it and/or modify
     it under the terms of the GNU General Public License as published by
@@ -329,7 +329,7 @@ necessary.  Here is a sample; alter the names:
   Yoyodyne, Inc., hereby disclaims all copyright interest in the program
   `Gnomovision' (which makes passes at compilers) written by James Hacker.
 
-  {signature of Ty Coon}, 1 April 1989
+  <signature of Ty Coon>, 1 April 1989
   Ty Coon, President of Vice
 
 This General Public License does not permit incorporating your program into

README.md

@@ -1,6 +1,8 @@
-# Grid
+# Grid [![Build Status](https://travis-ci.org/paboyle/Grid.svg?branch=master)](https://travis-ci.org/paboyle/Grid)
 Data parallel C++ mathematical object library
 
+Last update 2015/7/30
+
 This library provides data parallel C++ container classes with internal memory layout
 that is transformed to map efficiently to SIMD architectures. CSHIFT facilities
 are provided, similar to HPF and cmfortran, and user control is given over the mapping of
@@ -20,16 +22,16 @@ optimally use MPI, OpenMP and SIMD parallelism under the hood. This is a signifi
 for most programmers.
 
 The layout transformations are parametrised by the SIMD vector length. This adapts according to the architecture.
-Presently SSE4 (128 bit) AVX, AVX2 (256 bit) and IMCI and AVX512 (512 bit) targets are supported.
+Presently SSE4 (128 bit) AVX, AVX2 (256 bit) and IMCI and AVX512 (512 bit) targets are supported (ARM NEON on the way).
 
 These are presented as 
 
-  vRealF, vRealD, vComplexF, vComplexD 
+     vRealF, vRealD, vComplexF, vComplexD 
 
 internal vector data types. These may be useful in themselves for other programmers.
 The corresponding scalar types are named
 
-  RealF, RealD, ComplexF, ComplexD
+     RealF, RealD, ComplexF, ComplexD
 
 MPI, OpenMP, and SIMD parallelism are present in the library.
 
@@ -39,12 +41,18 @@ are examples:
 
      ./configure CXX=clang++ CXXFLAGS="-std=c++11 -O3 -msse4" --enable-simd=SSE4
 
-     ./configure CXX=clang++ CXXFLAGS="-std=c++11 -O3 -mavx" --enable-simd=AVX1
+     ./configure CXX=clang++ CXXFLAGS="-std=c++11 -O3 -mavx" --enable-simd=AVX
 
      ./configure CXX=clang++ CXXFLAGS="-std=c++11 -O3 -mavx2" --enable-simd=AVX2
 
      ./configure CXX=icpc CXXFLAGS="-std=c++11 -O3 -mmic" --enable-simd=AVX512 --host=none
      
+Note: Before running configure it could be necessary to execute the script 
+       
+       script/filelist
+
+
      
 For developers:
 Use reconfigure_script in the scripts/ directory to create the autotools environment 
+

TODO

@@ -1,5 +1,27 @@
 TODO:
 ---------------
+
+* Forces; the UdSdU  term in gauge force term is half of what I think it should
+  be. This is a consequence of taking ONLY the first term in:
+
+  dSg/dt = dU/dt dSdU + dUdag/dt dSdUdag
+
+  in the fermion force.
+
+  Now, S_mom = - tr Pmu Pmu      ; Pmu anti-herm
+
+                                  .
+       d Smom/dt = - 2.0 tr Pmu Pmu   = - dSg/dt = - tr Pmu [Umu dSdUmu + UmuDag dSdUmuDag]
+
+           .
+       => Pmu =  Umu dSdUmu
+
+       Where the norm is half expected.
+
+  This means we must double the force in the Test_xxx_force routines, and is the origin of the factor of two.
+  This 2x is applied by hand in the fermion routines and in the Test_rect_force routine.
+
+
 Policies:
 
 * Link smearing/boundary conds; Policy class based implementation ; framework more in place

benchmarks/Benchmark_comms.cc

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./benchmarks/Benchmark_comms.cc
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include <Grid.h>
 
 using namespace std;

benchmarks/Benchmark_dwf.cc

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./benchmarks/Benchmark_dwf.cc
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include <Grid.h>
 
 using namespace std;
@@ -16,15 +44,21 @@ struct scal {
     Gamma::GammaT
   };
 
+bool overlapComms = false;
+
 int main (int argc, char ** argv)
 {
   Grid_init(&argc,&argv);
 
+  if( GridCmdOptionExists(argv,argv+argc,"--asynch") ){
+    overlapComms = true;
+  }
+
   int threads = GridThread::GetThreads();
   std::cout<<GridLogMessage << "Grid is setup to use "<<threads<<" threads"<<std::endl;
 
   std::vector<int> latt4 = GridDefaultLatt();
-  const int Ls=16;
+  const int Ls=8;
   GridCartesian         * UGrid   = SpaceTimeGrid::makeFourDimGrid(GridDefaultLatt(), GridDefaultSimd(Nd,vComplex::Nsimd()),GridDefaultMpi());
   GridRedBlackCartesian * UrbGrid = SpaceTimeGrid::makeFourDimRedBlackGrid(UGrid);
   GridCartesian         * FGrid   = SpaceTimeGrid::makeFiveDimGrid(Ls,UGrid);
@@ -79,10 +113,16 @@ int main (int argc, char ** argv)
 
   RealD mass=0.1;
   RealD M5  =1.8;
-  DomainWallFermionR Dw(Umu,*FGrid,*FrbGrid,*UGrid,*UrbGrid,mass,M5);
+
+  typename DomainWallFermionR::ImplParams params; 
+  params.overlapCommsCompute = overlapComms;
+  
+  RealD NP = UGrid->_Nprocessors;
+
+  DomainWallFermionR Dw(Umu,*FGrid,*FrbGrid,*UGrid,*UrbGrid,mass,M5,params);
   
   std::cout<<GridLogMessage << "Calling Dw"<<std::endl;
-  int ncall=10000;
+  int ncall=100;
   {
     double t0=usecond();
     for(int i=0;i<ncall;i++){
@@ -97,11 +137,13 @@ int main (int argc, char ** argv)
     std::cout<<GridLogMessage << "norm result "<< norm2(result)<<std::endl;
     std::cout<<GridLogMessage << "norm ref    "<< norm2(ref)<<std::endl;
     std::cout<<GridLogMessage << "mflop/s =   "<< flops/(t1-t0)<<std::endl;
+    std::cout<<GridLogMessage << "mflop/s per node =  "<< flops/(t1-t0)/NP<<std::endl;
     err = ref-result; 
     std::cout<<GridLogMessage << "norm diff   "<< norm2(err)<<std::endl;
     Dw.Report();
   }
 
+  exit(0);
 
   if (1)
   { // Naive wilson dag implementation
@@ -154,8 +196,8 @@ int main (int argc, char ** argv)
     double flops=(1344.0*volume*ncall)/2;
 
     std::cout<<GridLogMessage << "Deo mflop/s =   "<< flops/(t1-t0)<<std::endl;
+    std::cout<<GridLogMessage << "Deo mflop/s per node   "<< flops/(t1-t0)/NP<<std::endl;
   }
-
   Dw.DhopEO(src_o,r_e,DaggerNo);
   Dw.DhopOE(src_e,r_o,DaggerNo);
   Dw.Dhop  (src  ,result,DaggerNo);

benchmarks/Benchmark_memory_asynch.cc

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./benchmarks/Benchmark_memory_asynch.cc
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include <Grid.h>
 
 using namespace std;

benchmarks/Benchmark_memory_bandwidth.cc

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./benchmarks/Benchmark_memory_bandwidth.cc
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include <Grid.h>
 
 using namespace std;

benchmarks/Benchmark_su3.cc

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./benchmarks/Benchmark_su3.cc
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: Peter Boyle <peterboyle@Peters-MacBook-Pro-2.local>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include <Grid.h>
 
 using namespace std;

benchmarks/Benchmark_wilson.cc

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./benchmarks/Benchmark_wilson.cc
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include <Grid.h>
 
 using namespace std;
@@ -16,10 +44,15 @@ struct scal {
     Gamma::GammaT
   };
 
+bool overlapComms = false;
+
 int main (int argc, char ** argv)
 {
   Grid_init(&argc,&argv);
 
+  if( GridCmdOptionExists(argv,argv+argc,"--asynch") ){
+    overlapComms = true;
+  }
 
   std::vector<int> latt_size   = GridDefaultLatt();
   std::vector<int> simd_layout = GridDefaultSimd(Nd,vComplex::Nsimd());
@@ -57,11 +90,12 @@ int main (int argc, char ** argv)
   Complex cone(1.0,0.0);
   for(int nn=0;nn<Nd;nn++){
     random(pRNG,U[nn]);
-    if(0) {
-      if (nn==-1) { U[nn]=zero; std::cout<<GridLogMessage << "zeroing gauge field in dir "<<nn<<std::endl; }
-      else       { U[nn] = cone;std::cout<<GridLogMessage << "unit gauge field in dir "<<nn<<std::endl; }
+    if(1) {
+      if (nn!=2) { U[nn]=zero; std::cout<<GridLogMessage << "zeroing gauge field in dir "<<nn<<std::endl; }
+      //      else       { U[nn]= cone;std::cout<<GridLogMessage << "unit gauge field in dir "<<nn<<std::endl; }
+      else       { std::cout<<GridLogMessage << "random gauge field in dir "<<nn<<std::endl; }
     }
-    pokeIndex<LorentzIndex>(Umu,U[nn],nn);
+    PokeIndex<LorentzIndex>(Umu,U[nn],nn);
   }
 #endif
 
@@ -87,7 +121,11 @@ int main (int argc, char ** argv)
   }
   ref = -0.5*ref;
   RealD mass=0.1;
-  WilsonFermionR Dw(Umu,Grid,RBGrid,mass);
+
+  typename WilsonFermionR::ImplParams params; 
+  params.overlapCommsCompute = overlapComms;
+
+  WilsonFermionR Dw(Umu,Grid,RBGrid,mass,params);
   
   std::cout<<GridLogMessage << "Calling Dw"<<std::endl;
   int ncall=1000;

benchmarks/Benchmark_zmm.cc

@@ -0,0 +1,172 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./tests/Test_zmm.cc
+
+    Copyright (C) 2015
+
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
+#include <Grid.h>
+#include <PerfCount.h>
+
+
+using namespace Grid;
+using namespace Grid::QCD;
+
+
+int bench(std::ofstream &os, std::vector<int> &latt4,int Ls);
+
+int main(int argc,char **argv)
+{
+  Grid_init(&argc,&argv);
+  std::ofstream os("zmm.dat");
+
+  os << "#V Ls Lxy Lzt C++ Asm OMP L1 " <<std::endl;
+  for(int L=4;L<=32;L+=4){
+    for(int m=1;m<=2;m++){
+      for(int Ls=8;Ls<=16;Ls+=8){
+	std::vector<int> grid({L,L,m*L,m*L});
+	for(int i=0;i<4;i++) { 
+	  std::cout << grid[i]<<"x";
+	}
+	std::cout << Ls<<std::endl;
+	bench(os,grid,Ls);
+      }
+    }
+  }
+}
+
+int bench(std::ofstream &os, std::vector<int> &latt4,int Ls)
+{
+
+  GridCartesian         * UGrid   = SpaceTimeGrid::makeFourDimGrid(latt4, GridDefaultSimd(Nd,vComplex::Nsimd()),GridDefaultMpi());
+  GridRedBlackCartesian * UrbGrid = SpaceTimeGrid::makeFourDimRedBlackGrid(UGrid);
+  GridCartesian         * FGrid   = SpaceTimeGrid::makeFiveDimGrid(Ls,UGrid);
+  GridRedBlackCartesian * FrbGrid = SpaceTimeGrid::makeFiveDimRedBlackGrid(Ls,UGrid);
+
+  std::vector<int> simd_layout = GridDefaultSimd(Nd,vComplex::Nsimd());
+  std::vector<int> mpi_layout  = GridDefaultMpi();
+  int threads = GridThread::GetThreads();
+
+  std::vector<int> seeds4({1,2,3,4});
+  std::vector<int> seeds5({5,6,7,8});
+
+  GridSerialRNG sRNG; sRNG.SeedFixedIntegers(seeds4);
+
+  LatticeFermion src (FGrid);
+  LatticeFermion tmp (FGrid);
+  LatticeFermion srce(FrbGrid);
+
+  LatticeFermion resulto(FrbGrid); resulto=zero;
+  LatticeFermion resulta(FrbGrid); resulta=zero;
+  LatticeFermion junk(FrbGrid); junk=zero;
+  LatticeFermion diff(FrbGrid); 
+  LatticeGaugeField Umu(UGrid);
+
+  double mfc, mfa, mfo, mfl1;
+
+  GridParallelRNG          RNG4(UGrid);  RNG4.SeedFixedIntegers(seeds4);
+  GridParallelRNG          RNG5(FGrid);  RNG5.SeedFixedIntegers(seeds5);
+  random(RNG5,src);
+#if 1
+  random(RNG4,Umu);
+#else
+  int mmu=2;
+  std::vector<LatticeColourMatrix> U(4,UGrid);
+  for(int mu=0;mu<Nd;mu++){
+    U[mu] = PeekIndex<LorentzIndex>(Umu,mu);
+    if ( mu!=mmu ) U[mu] = zero;
+    if ( mu==mmu ) U[mu] = 1.0;
+    PokeIndex<LorentzIndex>(Umu,U[mu],mu);
+  }
+#endif
+ pickCheckerboard(Even,srce,src);
+
+  RealD mass=0.1;
+  RealD M5  =1.8;
+  DomainWallFermionR Dw(Umu,*FGrid,*FrbGrid,*UGrid,*UrbGrid,mass,M5);
+
+  std::cout<<GridLogMessage << "Calling Dw"<<std::endl;
+  int ncall=50;
+  double t0=usecond();
+  for(int i=0;i<ncall;i++){
+    Dw.DhopOE(srce,resulto,0);
+  }
+  double t1=usecond();
+
+  double volume=Ls;  for(int mu=0;mu<Nd;mu++) volume=volume*latt4[mu];
+  double flops=1344*volume/2;
+
+  mfc = flops*ncall/(t1-t0);
+  std::cout<<GridLogMessage << "Called C++ Dw"<< " mflop/s =   "<< mfc<<std::endl;
+
+  QCD::WilsonFermion5DStatic::AsmOptDslash=1;
+  t0=usecond();
+  for(int i=0;i<ncall;i++){
+    Dw.DhopOE(srce,resulta,0);
+  }
+  t1=usecond();
+  mfa = flops*ncall/(t1-t0);
+  std::cout<<GridLogMessage << "Called ASM Dw"<< " mflop/s =   "<< mfa<<std::endl;
+  /*
+  int dag=DaggerNo;
+  t0=usecond();
+  for(int i=0;i<1;i++){
+    Dw.DhopInternalOMPbench(Dw.StencilEven,Dw.LebesgueEvenOdd,Dw.UmuOdd,srce,resulta,dag);
+  }
+  t1=usecond();
+  mfo = flops*100/(t1-t0);
+  std::cout<<GridLogMessage << "Called ASM-OMP Dw"<< " mflop/s =   "<< mfo<<std::endl;
+
+  t0=usecond();
+  for(int i=0;i<1;i++){
+    Dw.DhopInternalL1bench(Dw.StencilEven,Dw.LebesgueEvenOdd,Dw.UmuOdd,srce,resulta,dag);
+  }
+  t1=usecond();
+  mfl1= flops*100/(t1-t0);
+  std::cout<<GridLogMessage << "Called ASM-L1 Dw"<< " mflop/s =   "<< mfl1<<std::endl;
+  os << latt4[0]*latt4[1]*latt4[2]*latt4[3]<< " "<<Ls<<" "<< latt4[0] <<" " <<latt4[2]<< " "
+     << mfc<<" "
+     << mfa<<" "
+     << mfo<<" "
+     << mfl1<<std::endl;
+  */
+
+#if 0
+  for(int i=0;i< PerformanceCounter::NumTypes(); i++ ){
+    Dw.DhopOE(srce,resulta,0);
+    PerformanceCounter Counter(i);
+    Counter.Start();
+    Dw.DhopOE(srce,resulta,0);
+    Counter.Stop();
+    Counter.Report();
+  }
+#endif
+  //resulta = (-0.5) * resulta;
+
+  diff = resulto-resulta;
+  std::cout<<GridLogMessage << "diff "<< norm2(diff)<<std::endl;
+  std::cout<<std::endl;
+  return 0;
+}
+
+

benchmarks/Make.inc

@@ -1,5 +1,5 @@
 
-bin_PROGRAMS = Benchmark_comms Benchmark_dwf Benchmark_memory_asynch Benchmark_memory_bandwidth Benchmark_su3 Benchmark_wilson
+bin_PROGRAMS = Benchmark_comms Benchmark_dwf Benchmark_memory_asynch Benchmark_memory_bandwidth Benchmark_su3 Benchmark_wilson Benchmark_zmm
 
 
 Benchmark_comms_SOURCES=Benchmark_comms.cc
@@ -25,3 +25,7 @@ Benchmark_su3_LDADD=-lGrid
 Benchmark_wilson_SOURCES=Benchmark_wilson.cc
 Benchmark_wilson_LDADD=-lGrid
 
+
+Benchmark_zmm_SOURCES=Benchmark_zmm.cc
+Benchmark_zmm_LDADD=-lGrid
+

benchmarks/simple_su3_expr.cc

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./benchmarks/simple_su3_expr.cc
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include <Grid.h>
 
 using namespace std;

benchmarks/simple_su3_test.cc

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./benchmarks/simple_su3_test.cc
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include <Grid.h>
 
 using namespace std;

configure.ac

@@ -53,25 +53,21 @@ echo
 echo Checking libraries 
 echo :::::::::::::::::::::::::::::::::::::::::::
 
-
 AC_CHECK_FUNCS([gettimeofday])
 
-#AC_CHECK_LIB([gmp],[__gmpf_init],,
-#        [AC_MSG_ERROR(GNU Multiple Precision GMP library was not found in your system.
-#Please install or provide the correct path to your installation
-#Info at: http://www.gmplib.org)])
 
-#AC_CHECK_LIB([mpfr],[mpfr_init],,
-#        [AC_MSG_ERROR(GNU Multiple Precision MPFR library was not found in your system.
-#Please install or provide the correct path to your installation
-#Info at: http://www.mpfr.org/)])
+#
+# SIMD instructions selection
+#
 
 AC_ARG_ENABLE([simd],[AC_HELP_STRING([--enable-simd=SSE4|AVX|AVXFMA4|AVX2|AVX512|IMCI],\
 	[Select instructions to be SSE4.0, AVX 1.0, AVX 2.0+FMA, AVX 512, IMCI])],\
-	[ac_SIMD=${enable_simd}],[ac_SIMD=AVX2])
+	[ac_SIMD=${enable_simd}],[ac_SIMD=DEBUG])
 
 supported=no
 
+ac_ZMM=no;
+
 case ${ac_SIMD} in
      SSE4)
        echo Configuring for SSE4
@@ -113,11 +109,13 @@ case ${ac_SIMD} in
        echo Configuring for AVX512 
        AC_DEFINE([AVX512],[1],[AVX512 Intrinsics for Knights Landing] )
        supported="cross compilation"
+       ac_ZMM=yes;
      ;;
      IMCI)
        echo Configuring for IMCI
        AC_DEFINE([IMCI],[1],[IMCI Intrinsics for Knights Corner] )
        supported="cross compilation"
+       ac_ZMM=no;
      ;;
      NEONv8)
        echo Configuring for experimental ARMv8a support 
@@ -133,6 +131,17 @@ case ${ac_SIMD} in
      ;;
 esac
 
+case ${ac_ZMM} in
+yes)
+	echo Enabling ZMM source code
+;;
+no)
+	echo Disabling ZMM source code
+;;
+esac
+
+AM_CONDITIONAL(BUILD_ZMM,[ test "X${ac_ZMM}X" == "XyesX" ])
+
 AC_ARG_ENABLE([precision],[AC_HELP_STRING([--enable-precision=single|double],[Select default word size of Real])],[ac_PRECISION=${enable_precision}],[ac_PRECISION=double])
 case ${ac_PRECISION} in
      single)
@@ -145,6 +154,10 @@ case ${ac_PRECISION} in
      ;;
 esac
 
+#
+# Comms selection
+#
+
 AC_ARG_ENABLE([comms],[AC_HELP_STRING([--enable-comms=none|mpi],[Select communications])],[ac_COMMS=${enable_comms}],[ac_COMMS=none])
 
 case ${ac_COMMS} in
@@ -156,14 +169,39 @@ case ${ac_COMMS} in
        echo Configuring for MPI communications
        AC_DEFINE([GRID_COMMS_MPI],[1],[GRID_COMMS_MPI] )
      ;;
+     shmem)
+       echo Configuring for SHMEM communications
+       AC_DEFINE([GRID_COMMS_SHMEM],[1],[GRID_COMMS_SHMEM] )
+     ;;
      *)
      AC_MSG_ERROR([${ac_COMMS} unsupported --enable-comms option]); 
      ;;
 esac
 
+AM_CONDITIONAL(BUILD_COMMS_SHMEM,[ test "X${ac_COMMS}X" == "XshmemX" ])
 AM_CONDITIONAL(BUILD_COMMS_MPI,[ test "X${ac_COMMS}X" == "XmpiX" ])
 AM_CONDITIONAL(BUILD_COMMS_NONE,[ test "X${ac_COMMS}X" == "XnoneX" ])
 
+#
+# RNG selection
+#
+AC_ARG_ENABLE([rng],[AC_HELP_STRING([--enable-rng=ranlux48|mt19937],\
+	[Select Random Number Generator to be used])],\
+	[ac_RNG=${enable_rng}],[ac_RNG=ranlux48])
+case ${ac_RNG} in
+     ranlux48)
+     AC_DEFINE([RNG_RANLUX],[1],[RNG_RANLUX] )
+     ;;
+     mt19937)
+     AC_DEFINE([RNG_MT19937],[1],[RNG_MT19937] )
+     ;;
+     *)
+     AC_MSG_ERROR([${ac_RNG} unsupported --enable-rng option]); 
+     ;;
+esac
+#
+# Chroma regression tests
+#
 AC_ARG_ENABLE([chroma],[AC_HELP_STRING([--enable-chroma],[Expect chroma compiled under c++11 ])],ac_CHROMA=yes,ac_CHROMA=no)
 
 case ${ac_CHROMA} in
@@ -180,6 +218,26 @@ esac
 
 AM_CONDITIONAL(BUILD_CHROMA_REGRESSION,[ test "X${ac_CHROMA}X" == "XyesX" ])
 
+#
+# Lapack
+#
+AC_ARG_ENABLE([lapack],[AC_HELP_STRING([--enable-lapack],[Enable lapack yes/no ])],[ac_LAPACK=${enable_lapack}],[ac_LAPACK=no])
+
+case ${ac_LAPACK} in
+     yes)
+       echo Enabling lapack
+     ;;
+     no)
+       echo Disabling lapack
+     ;;
+     *)
+       echo Enabling lapack at ${ac_LAPACK}
+     ;;
+esac
+
+AM_CONDITIONAL(USE_LAPACK,[ test "X${ac_LAPACK}X" != "XnoX" ])
+AM_CONDITIONAL(USE_LAPACK_LIB,[ test "X${ac_LAPACK}X" != "XyesX" ])
+
 ###################################################################
 # Checks for doxygen support
 # if present enables the "make doxyfile" command
@@ -219,8 +277,11 @@ The following features are enabled:
 - graphs and diagrams           : `if test "x$enable_dot" = xyes; then echo yes; else echo no; fi`
 - Supported SIMD flags          : $SIMD_FLAGS
 ----------------------------------------------------------
-- enabled simd support          : ${ac_SIMD}   (supported: $supported )
+- enabled simd support          : ${ac_SIMD}   (config macro says supported: $supported )
 - communications type           : ${ac_COMMS}
+- default precision             : ${ac_PRECISION}
+- RNG choice                    : ${ac_RNG} 
+- LAPACK	                : ${ac_LAPACK} 
 
 
 "

gcc-bug-report/broken.cc

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./gcc-bug-report/broken.cc
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include <vector>
 #include <complex>
 #include <type_traits>

lib/Algorithms.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Algorithms.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_ALGORITHMS_H
 #define GRID_ALGORITHMS_H
 

lib/AlignedAllocator.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/AlignedAllocator.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_ALIGNED_ALLOCATOR_H
 #define GRID_ALIGNED_ALLOCATOR_H
 
@@ -8,11 +36,18 @@
 #include <malloc.h>
 #endif
 
-#include <immintrin.h>
 #ifdef HAVE_MM_MALLOC_H
 #include <mm_malloc.h>
 #endif
 
+#ifdef GRID_COMMS_SHMEM
+extern "C" { 
+#include <mpp/shmem.h>
+extern void * shmem_align(size_t, size_t);
+extern void  shmem_free(void *);
+}
+#endif
+
 namespace Grid {
 
 ////////////////////////////////////////////////////////////////////
@@ -44,21 +79,59 @@ public:
 
   size_type  max_size() const throw() { return size_t(-1) / sizeof(_Tp); }
 
-  pointer allocate(size_type __n, const void* = 0)
+  pointer allocate(size_type __n, const void* _p= 0)
   { 
+#ifdef GRID_COMMS_SHMEM
+
+    _Tp *ptr = (_Tp *) shmem_align(__n*sizeof(_Tp),64);
+
+
+#define PARANOID_SYMMETRIC_HEAP
+#ifdef PARANOID_SYMMETRIC_HEAP
+    static void * bcast;
+    static long  psync[_SHMEM_REDUCE_SYNC_SIZE];
+
+    bcast = (void *) ptr;
+    shmem_broadcast32((void *)&bcast,(void *)&bcast,sizeof(void *)/4,0,0,0,shmem_n_pes(),psync);
+
+    if ( bcast != ptr ) {
+      std::printf("inconsistent alloc pe %d %lx %lx \n",shmem_my_pe(),bcast,ptr);std::fflush(stdout);
+      BACKTRACEFILE();
+      exit(0);
+    }
+
+    assert( bcast == (void *) ptr);
+
+#endif 
+#else
+
 #ifdef HAVE_MM_MALLOC_H
     _Tp * ptr = (_Tp *) _mm_malloc(__n*sizeof(_Tp),128);
 #else
     _Tp * ptr = (_Tp *) memalign(128,__n*sizeof(_Tp));
 #endif
+
+#endif
+    _Tp tmp;
+#undef FIRST_TOUCH_OPTIMISE
+#ifdef FIRST_TOUCH_OPTIMISE
+#pragma omp parallel for 
+  for(int i=0;i<__n;i++){
+    ptr[i]=tmp;
+  }
+#endif 
     return ptr;
   }
 
   void deallocate(pointer __p, size_type) { 
+#ifdef GRID_COMMS_SHMEM
+    shmem_free((void *)__p);
+#else
 #ifdef HAVE_MM_MALLOC_H
     _mm_free((void *)__p); 
 #else
     free((void *)__p);
+#endif
 #endif
   }
   void construct(pointer __p, const _Tp& __val) { };

lib/Cartesian.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Cartesian.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_CARTESIAN_H
 #define GRID_CARTESIAN_H
 

lib/Communicator.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Communicator.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_COMMUNICATOR_H
 #define GRID_COMMUNICATOR_H
 

lib/Config.h.in

@@ -1,174 +0,0 @@
-/* lib/Config.h.in.  Generated from configure.ac by autoheader.  */
-
-/* AVX Intrinsics */
-#undef AVX1
-
-/* AVX2 Intrinsics */
-#undef AVX2
-
-/* AVX512 Intrinsics for Knights Landing */
-#undef AVX512
-
-/* AVX Intrinsics with FMA4 */
-#undef AVXFMA4
-
-/* EMPTY_SIMD only for DEBUGGING */
-#undef EMPTY_SIMD
-
-/* GRID_COMMS_MPI */
-#undef GRID_COMMS_MPI
-
-/* GRID_COMMS_NONE */
-#undef GRID_COMMS_NONE
-
-/* GRID_DEFAULT_PRECISION is DOUBLE */
-#undef GRID_DEFAULT_PRECISION_DOUBLE
-
-/* GRID_DEFAULT_PRECISION is SINGLE */
-#undef GRID_DEFAULT_PRECISION_SINGLE
-
-/* Support Altivec instructions */
-#undef HAVE_ALTIVEC
-
-/* Support AVX (Advanced Vector Extensions) instructions */
-#undef HAVE_AVX
-
-/* Support AVX2 (Advanced Vector Extensions 2) instructions */
-#undef HAVE_AVX2
-
-/* Define to 1 if you have the declaration of `be64toh', and to 0 if you
-   don't. */
-#undef HAVE_DECL_BE64TOH
-
-/* Define to 1 if you have the declaration of `ntohll', and to 0 if you don't.
-   */
-#undef HAVE_DECL_NTOHLL
-
-/* Define to 1 if you have the <endian.h> header file. */
-#undef HAVE_ENDIAN_H
-
-/* Define to 1 if you have the <execinfo.h> header file. */
-#undef HAVE_EXECINFO_H
-
-/* Support FMA3 (Fused Multiply-Add) instructions */
-#undef HAVE_FMA
-
-/* Define to 1 if you have the `gettimeofday' function. */
-#undef HAVE_GETTIMEOFDAY
-
-/* Define to 1 if you have the <gmp.h> header file. */
-#undef HAVE_GMP_H
-
-/* Define to 1 if you have the <inttypes.h> header file. */
-#undef HAVE_INTTYPES_H
-
-/* Define to 1 if you have the <malloc.h> header file. */
-#undef HAVE_MALLOC_H
-
-/* Define to 1 if you have the <malloc/malloc.h> header file. */
-#undef HAVE_MALLOC_MALLOC_H
-
-/* Define to 1 if you have the <memory.h> header file. */
-#undef HAVE_MEMORY_H
-
-/* Support mmx instructions */
-#undef HAVE_MMX
-
-/* Define to 1 if you have the <mm_malloc.h> header file. */
-#undef HAVE_MM_MALLOC_H
-
-/* Support SSE (Streaming SIMD Extensions) instructions */
-#undef HAVE_SSE
-
-/* Support SSE2 (Streaming SIMD Extensions 2) instructions */
-#undef HAVE_SSE2
-
-/* Support SSE3 (Streaming SIMD Extensions 3) instructions */
-#undef HAVE_SSE3
-
-/* Support SSSE4.1 (Streaming SIMD Extensions 4.1) instructions */
-#undef HAVE_SSE4_1
-
-/* Support SSSE4.2 (Streaming SIMD Extensions 4.2) instructions */
-#undef HAVE_SSE4_2
-
-/* Support SSSE3 (Supplemental Streaming SIMD Extensions 3) instructions */
-#undef HAVE_SSSE3
-
-/* Define to 1 if you have the <stdint.h> header file. */
-#undef HAVE_STDINT_H
-
-/* Define to 1 if you have the <stdlib.h> header file. */
-#undef HAVE_STDLIB_H
-
-/* Define to 1 if you have the <strings.h> header file. */
-#undef HAVE_STRINGS_H
-
-/* Define to 1 if you have the <string.h> header file. */
-#undef HAVE_STRING_H
-
-/* Define to 1 if you have the <sys/stat.h> header file. */
-#undef HAVE_SYS_STAT_H
-
-/* Define to 1 if you have the <sys/types.h> header file. */
-#undef HAVE_SYS_TYPES_H
-
-/* Define to 1 if you have the <unistd.h> header file. */
-#undef HAVE_UNISTD_H
-
-/* IMCI Intrinsics for Knights Corner */
-#undef IMCI
-
-/* NEON ARMv8 Experimental support */
-#undef NEONv8
-
-/* Name of package */
-#undef PACKAGE
-
-/* Define to the address where bug reports for this package should be sent. */
-#undef PACKAGE_BUGREPORT
-
-/* Define to the full name of this package. */
-#undef PACKAGE_NAME
-
-/* Define to the full name and version of this package. */
-#undef PACKAGE_STRING
-
-/* Define to the one symbol short name of this package. */
-#undef PACKAGE_TARNAME
-
-/* Define to the home page for this package. */
-#undef PACKAGE_URL
-
-/* Define to the version of this package. */
-#undef PACKAGE_VERSION
-
-/* SSE4 Intrinsics */
-#undef SSE4
-
-/* Define to 1 if you have the ANSI C header files. */
-#undef STDC_HEADERS
-
-/* Version number of package */
-#undef VERSION
-
-/* Define for Solaris 2.5.1 so the uint32_t typedef from <sys/synch.h>,
-   <pthread.h>, or <semaphore.h> is not used. If the typedef were allowed, the
-   #define below would cause a syntax error. */
-#undef _UINT32_T
-
-/* Define for Solaris 2.5.1 so the uint64_t typedef from <sys/synch.h>,
-   <pthread.h>, or <semaphore.h> is not used. If the typedef were allowed, the
-   #define below would cause a syntax error. */
-#undef _UINT64_T
-
-/* Define to `unsigned int' if <sys/types.h> does not define. */
-#undef size_t
-
-/* Define to the type of an unsigned integer type of width exactly 32 bits if
-   such a type exists and the standard includes do not define it. */
-#undef uint32_t
-
-/* Define to the type of an unsigned integer type of width exactly 64 bits if
-   such a type exists and the standard includes do not define it. */
-#undef uint64_t

lib/Cshift.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Cshift.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef _GRID_CSHIFT_H_
 #define _GRID_CSHIFT_H_
 
@@ -10,4 +37,8 @@
 #ifdef GRID_COMMS_MPI
 #include <cshift/Cshift_mpi.h>
 #endif 
+
+#ifdef GRID_COMMS_SHMEM
+#include <cshift/Cshift_mpi.h> // uses same implementation of communicator
+#endif 
 #endif

lib/Grid.h

@@ -1,3 +1,32 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Grid.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: azusayamaguchi <ayamaguc@YAMAKAZE.local>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 //
 //  Grid.h
 //  simd
@@ -33,10 +62,12 @@
 #include <serialisation/Serialisation.h>
 #include <Config.h>
 #include <Timer.h>
+#include <PerfCount.h>
 #include <Log.h>
 #include <AlignedAllocator.h>
 #include <Simd.h>
 #include <Threads.h>
+#include <Lexicographic.h>
 #include <Communicator.h> 
 #include <Cartesian.h>    
 #include <Tensors.h>      
@@ -44,11 +75,14 @@
 #include <Cshift.h>       
 #include <Stencil.h>      
 #include <Algorithms.h>   
-#include <qcd/QCD.h>
 #include <parallelIO/BinaryIO.h>
+#include <qcd/QCD.h>
 #include <parallelIO/NerscIO.h>
-
 #include <Init.h>
 
+#include <qcd/hmc/NerscCheckpointer.h>
+#include <qcd/hmc/HmcRunner.h>
+
+
 
 #endif

lib/Init.cc

@@ -1,3 +1,33 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Init.cc
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: Peter Boyle <peterboyle@MacBook-Pro.local>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 /****************************************************************************/
 /* pab: Signal magic. Processor state dump is x86-64 specific               */
 /****************************************************************************/
@@ -15,12 +45,6 @@
 #include <algorithm>
 #include <iterator>
 
-#define __X86_64
-
-#ifdef HAVE_EXECINFO_H
-#include <execinfo.h>
-#endif
-
 namespace Grid {
 
 //////////////////////////////////////////////////////
@@ -120,6 +144,10 @@ void GridParseLayout(char **argv,int argc,
   }
   if( GridCmdOptionExists(argv,argv+argc,"--threads") ){
     std::vector<int> ompthreads(0);
+#ifndef GRID_OMP
+    std::cout << GridLogWarning << "'--threads' option used but Grid was"
+              << " not compiled with thread support" << std::endl;
+#endif
     arg= GridCmdOptionPayload(argv,argv+argc,"--threads");
     GridCmdOptionIntVector(arg,ompthreads);
     assert(ompthreads.size()==1);
@@ -144,9 +172,8 @@ std::string GridCmdVectorIntToString(const std::vector<int> & vec){
 /////////////////////////////////////////////////////////
 void Grid_init(int *argc,char ***argv)
 {
-#ifdef GRID_COMMS_MPI
-  MPI_Init(argc,argv);
-#endif
+  CartesianCommunicator::Init(argc,argv);
+
   // Parse command line args.
 
   GridLogger::StopWatch.Start();
@@ -164,9 +191,10 @@ void Grid_init(int *argc,char ***argv)
     std::cout<<GridLogMessage<<"--debug-stdout  : print stdout from EVERY node"<<std::endl;    
     std::cout<<GridLogMessage<<"--decomposition : report on default omp,mpi and simd decomposition"<<std::endl;    
     std::cout<<GridLogMessage<<"--mpi n.n.n.n   : default MPI decomposition"<<std::endl;    
-    std::cout<<GridLogMessage<<"--omp n         : default number of OMP threads"<<std::endl;    
+    std::cout<<GridLogMessage<<"--threads n     : default number of OMP threads"<<std::endl;
     std::cout<<GridLogMessage<<"--grid n.n.n.n  : default Grid size"<<std::endl;    
-    std::cout<<GridLogMessage<<"--log list      : comma separted list of streams from Error,Warning,Message,Performance,Iterative,Debug"<<std::endl;    
+    std::cout<<GridLogMessage<<"--log list      : comma separted list of streams from Error,Warning,Message,Performance,Iterative,Integrator,Debug"<<std::endl;
+    exit(EXIT_SUCCESS);
   }
 
   if( GridCmdOptionExists(*argv,*argv+*argc,"--log") ){
@@ -207,7 +235,37 @@ void Grid_init(int *argc,char ***argv)
     std::cout<<GridLogMessage<<"\tvComplexD      : "<<sizeof(vComplexD)*8 <<"bits ; " <<GridCmdVectorIntToString(GridDefaultSimd(4,vComplexD::Nsimd()))<<std::endl;
   }
 
-
+  std::cout <<std::endl;
+  std::cout <<Logger::RED  << "__|__|__|__|__"<<             "|__|__|_"<<Logger::PURPLE<<"_|__|__|"<<                "__|__|__|__|__"<<std::endl; 
+  std::cout <<Logger::RED  << "__|__|__|__|__"<<             "|__|__|_"<<Logger::PURPLE<<"_|__|__|"<<                "__|__|__|__|__"<<std::endl; 
+  std::cout <<Logger::RED  << "__|__|  |  |  "<<             "|  |  | "<<Logger::PURPLE<<" |  |  |"<<                "  |  |  | _|__"<<std::endl; 
+  std::cout <<Logger::RED  << "__|__         "<<             "        "<<Logger::PURPLE<<"        "<<                "          _|__"<<std::endl; 
+  std::cout <<Logger::RED  << "__|_  "<<Logger::GREEN<<" GGGG   "<<Logger::RED<<" RRRR   "<<Logger::BLUE  <<" III    "<<Logger::PURPLE<<"DDDD  "<<Logger::PURPLE<<"    _|__"<<std::endl;
+  std::cout <<Logger::RED  << "__|_  "<<Logger::GREEN<<"G       "<<Logger::RED<<" R   R  "<<Logger::BLUE  <<"  I     "<<Logger::PURPLE<<"D   D "<<Logger::PURPLE<<"    _|__"<<std::endl;
+  std::cout <<Logger::RED  << "__|_  "<<Logger::GREEN<<"G       "<<Logger::RED<<" R   R  "<<Logger::BLUE  <<"  I     "<<Logger::PURPLE<<"D    D"<<Logger::PURPLE<<"    _|__"<<std::endl;
+  std::cout <<Logger::BLUE << "__|_  "<<Logger::GREEN<<"G  GG   "<<Logger::RED<<" RRRR   "<<Logger::BLUE  <<"  I     "<<Logger::PURPLE<<"D    D"<<Logger::GREEN <<"    _|__"<<std::endl;
+  std::cout <<Logger::BLUE << "__|_  "<<Logger::GREEN<<"G   G   "<<Logger::RED<<" R  R   "<<Logger::BLUE  <<"  I     "<<Logger::PURPLE<<"D   D "<<Logger::GREEN <<"    _|__"<<std::endl;
+  std::cout <<Logger::BLUE << "__|_  "<<Logger::GREEN<<" GGGG   "<<Logger::RED<<" R   R  "<<Logger::BLUE  <<" III    "<<Logger::PURPLE<<"DDDD  "<<Logger::GREEN <<"    _|__"<<std::endl;
+  std::cout <<Logger::BLUE << "__|__         "<<             "        "<<Logger::GREEN <<"        "<<                "          _|__"<<std::endl; 
+  std::cout <<Logger::BLUE << "__|__|__|__|__"<<             "|__|__|_"<<Logger::GREEN <<"_|__|__|"<<                "__|__|__|__|__"<<std::endl; 
+  std::cout <<Logger::BLUE << "__|__|__|__|__"<<             "|__|__|_"<<Logger::GREEN <<"_|__|__|"<<                "__|__|__|__|__"<<std::endl; 
+  std::cout <<Logger::BLUE << "  |  |  |  |  "<<             "|  |  | "<<Logger::GREEN <<" |  |  |"<<                "  |  |  |  |  "<<std::endl; 
+  std::cout << std::endl;
+  std::cout << std::endl;
+  std::cout <<Logger::YELLOW<< std::endl;
+  std::cout << "Copyright (C) 2015 Peter Boyle, Azusa Yamaguchi, Guido Cossu, Antonin Portelli and other authors"<<std::endl;
+  std::cout << "Colours by Tadahito Boyle "<<std::endl;
+  std::cout << std::endl;
+  std::cout << "This program is free software; you can redistribute it and/or modify"<<std::endl;
+  std::cout << "it under the terms of the GNU General Public License as published by"<<std::endl;
+  std::cout << "the Free Software Foundation; either version 2 of the License, or"<<std::endl;
+  std::cout << "(at your option) any later version."<<std::endl;
+  std::cout << std::endl;
+  std::cout << "This program is distributed in the hope that it will be useful,"<<std::endl;
+  std::cout << "but WITHOUT ANY WARRANTY; without even the implied warranty of"<<std::endl;
+  std::cout << "MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the"<<std::endl;
+  std::cout << "GNU General Public License for more details."<<std::endl;
+  std::cout << Logger::BLACK <<std::endl;
 }
 
   
@@ -224,7 +282,6 @@ double usecond(void) {
   return 1.0*tv.tv_usec + 1.0e6*tv.tv_sec;
 }
 
-#define _NBACKTRACE (256)
 void * Grid_backtrace_buffer[_NBACKTRACE];
 
 void Grid_sa_signal_handler(int sig,siginfo_t *si,void * ptr)
@@ -234,13 +291,13 @@ void Grid_sa_signal_handler(int sig,siginfo_t *si,void * ptr)
   printf("         code %d\n",si->si_code);
 
   // Linux/Posix
-#ifdef __linux__ 
+#ifdef __linux__
   // And x86 64bit
-    ucontext_t * uc= (ucontext_t *)ptr;
+#ifdef __x86_64__
+  ucontext_t * uc= (ucontext_t *)ptr;
   struct sigcontext *sc = (struct sigcontext *)&uc->uc_mcontext;
   printf("  instruction %llx\n",(unsigned long long)sc->rip);
 #define REG(A)  printf("  %s %lx\n",#A,sc-> A);
-
   REG(rdi);
   REG(rsi);
   REG(rbp);
@@ -261,17 +318,15 @@ void Grid_sa_signal_handler(int sig,siginfo_t *si,void * ptr)
   REG(r14);
   REG(r15);
 #endif
-#ifdef HAVE_EXECINFO_H
-  int symbols    = backtrace        (Grid_backtrace_buffer,_NBACKTRACE);
-  char **strings = backtrace_symbols(Grid_backtrace_buffer,symbols);
-  for (int i = 0; i < symbols; i++){
-    printf ("%s\n", strings[i]);
-  }
 #endif
+  BACKTRACE();
   exit(0);
   return;
 };
-
+#ifdef GRID_FPE
+#define _GNU_SOURCE
+#include <fenv.h>
+#endif
 void Grid_debug_handler_init(void)
 {
   struct sigaction sa,osa;
@@ -280,5 +335,9 @@ void Grid_debug_handler_init(void)
   sa.sa_flags    = SA_SIGINFO;
   sigaction(SIGSEGV,&sa,NULL);
   sigaction(SIGTRAP,&sa,NULL);
+#ifdef GRID_FPE
+  feenableexcept( FE_INVALID|FE_OVERFLOW|FE_DIVBYZERO);
+  sigaction(SIGFPE,&sa,NULL);
+#endif
 }
 }

lib/Init.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Init.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_INIT_H
 #define GRID_INIT_H
 
@@ -21,6 +49,8 @@ namespace Grid {
   std::string GridCmdOptionPayload(char ** begin, char ** end, const std::string & option);
   bool        GridCmdOptionExists(char** begin, char** end, const std::string& option);
   std::string GridCmdVectorIntToString(const std::vector<int> & vec);
+  void GridCmdOptionCSL(std::string str,std::vector<std::string> & vec);
+  void GridCmdOptionIntVector(std::string &str,std::vector<int> & vec);
 
   void GridParseLayout(char **argv,int argc,
 		       std::vector<int> &latt,

lib/Lattice.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Lattice.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_H
 #define GRID_LATTICE_H
 

lib/Lexicographic.h

@@ -0,0 +1,32 @@
+#ifndef GRID_LEXICOGRAPHIC_H
+#define GRID_LEXICOGRAPHIC_H
+
+
+namespace Grid{
+
+  class Lexicographic {
+  public:
+
+    static inline void CoorFromIndex (std::vector<int>& coor,int index,std::vector<int> &dims){
+      int nd= dims.size();
+      coor.resize(nd);
+      for(int d=0;d<nd;d++){
+	coor[d] = index % dims[d];
+	index   = index / dims[d];
+      }
+    }
+
+    static inline void IndexFromCoor (std::vector<int>& coor,int &index,std::vector<int> &dims){
+      int nd=dims.size();
+      int stride=1;
+      index=0;
+      for(int d=0;d<nd;d++){
+	index = index+stride*coor[d];
+	stride=stride*dims[d];
+      }
+    }
+
+  };
+
+}
+#endif

lib/Log.cc

@@ -1,16 +1,67 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Log.cc
+
+    Copyright (C) 2015
+
+Author: Antonin Portelli <antonin.portelli@me.com>
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include <Grid.h>
 
 namespace Grid {
 
 GridStopWatch Logger::StopWatch;
 std::ostream  Logger::devnull(0);
+std::string Logger::BLACK("\033[30m");
+std::string Logger::RED("\033[31m");
+std::string Logger::GREEN("\033[32m");
+std::string Logger::YELLOW("\033[33m");
+std::string Logger::BLUE("\033[34m");
+std::string Logger::PURPLE("\033[35m");
+std::string Logger::CYAN("\033[36m");
+std::string Logger::WHITE("\033[37m");
+std::string Logger::NORMAL("\033[0;39m");
+std::string EMPTY("");
 
-GridLogger GridLogError      (1,"Error");
-GridLogger GridLogWarning    (1,"Warning");
-GridLogger GridLogMessage    (1,"Message");
-GridLogger GridLogDebug      (1,"Debug");
-GridLogger GridLogPerformance(1,"Performance");
-GridLogger GridLogIterative  (1,"Iterative");
+#if 0  
+  GridLogger GridLogError      (1,"Error",Logger::RED);
+  GridLogger GridLogWarning    (1,"Warning",Logger::YELLOW);
+  GridLogger GridLogMessage    (1,"Message",Logger::BLACK);
+  GridLogger GridLogDebug      (1,"Debug",Logger::PURPLE);
+  GridLogger GridLogPerformance(1,"Performance",Logger::GREEN);
+  GridLogger GridLogIterative  (1,"Iterative",Logger::BLUE);
+  GridLogger GridLogIntegrator (1,"Integrator",Logger::BLUE);
+#else
+  GridLogger GridLogError      (1,"Error",EMPTY);
+  GridLogger GridLogWarning    (1,"Warning",EMPTY);
+  GridLogger GridLogMessage    (1,"Message",EMPTY);
+  GridLogger GridLogDebug      (1,"Debug",EMPTY);
+  GridLogger GridLogPerformance(1,"Performance",EMPTY);
+  GridLogger GridLogIterative  (1,"Iterative",EMPTY);
+  GridLogger GridLogIntegrator (1,"Integrator",EMPTY);
+#endif
 
 void GridLogConfigure(std::vector<std::string> &logstreams)
 {
@@ -20,6 +71,20 @@ void GridLogConfigure(std::vector<std::string> &logstreams)
   GridLogIterative.Active(0);
   GridLogDebug.Active(0);
   GridLogPerformance.Active(0);
+  GridLogIntegrator.Active(0);
+
+  int blackAndWhite = 1;
+  if(blackAndWhite){
+    Logger::BLACK = std::string("");
+    Logger::RED    =Logger::BLACK;
+    Logger::GREEN  =Logger::BLACK;
+    Logger::YELLOW =Logger::BLACK;
+    Logger::BLUE   =Logger::BLACK;
+    Logger::PURPLE =Logger::BLACK;
+    Logger::CYAN   =Logger::BLACK;
+    Logger::WHITE  =Logger::BLACK;
+    Logger::NORMAL =Logger::BLACK;
+  }
 
   for(int i=0;i<logstreams.size();i++){
     if ( logstreams[i]== std::string("Error")       ) GridLogError.Active(1);
@@ -28,6 +93,7 @@ void GridLogConfigure(std::vector<std::string> &logstreams)
     if ( logstreams[i]== std::string("Iterative")   ) GridLogIterative.Active(1);
     if ( logstreams[i]== std::string("Debug")       ) GridLogDebug.Active(1);
     if ( logstreams[i]== std::string("Performance") ) GridLogPerformance.Active(1);
+    if ( logstreams[i]== std::string("Integrator" ) ) GridLogIntegrator.Active(1);
   }
 }
 
@@ -36,13 +102,16 @@ void GridLogConfigure(std::vector<std::string> &logstreams)
 ////////////////////////////////////////////////////////////
 void Grid_quiesce_nodes(void)
 {
+  int me=0;
 #ifdef GRID_COMMS_MPI
-  int me;
   MPI_Comm_rank(MPI_COMM_WORLD,&me);
+#endif
+#ifdef GRID_COMMS_SHMEM
+  me = shmem_my_pe();
+#endif
   if ( me ) { 
     std::cout.setstate(std::ios::badbit);
   }
-#endif
 }
 
 void Grid_unquiesce_nodes(void)
@@ -52,11 +121,6 @@ void Grid_unquiesce_nodes(void)
 #endif
 }
 
-std::ostream& operator<< (std::ostream& stream, const GridTime& time)
-{
-  stream << time.count()<<" ms";
-  return stream;
-}
 
 }
 

lib/Log.h

@@ -1,21 +1,65 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Log.h
+
+    Copyright (C) 2015
+
+Author: Antonin Portelli <antonin.portelli@me.com>
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LOG_H
 #define GRID_LOG_H
+
+#ifdef HAVE_EXECINFO_H
+#include <execinfo.h>
+#endif
+
 namespace Grid {
 
 // Dress the output; use std::chrono for time stamping via the StopWatch class
+int Rank(void); // used for early stage debug before library init
 
-std::ostream& operator<< (std::ostream& stream, const GridTime& time);
 
 class Logger {
 protected:
     int active;
-    std::string name, topName;
+    std::string name, topName, COLOUR;
 public:
     static GridStopWatch StopWatch;
     static std::ostream devnull;
+
+    static std::string BLACK;
+    static std::string RED  ;
+    static std::string GREEN;
+    static std::string YELLOW;
+    static std::string BLUE  ;
+    static std::string PURPLE;
+    static std::string CYAN  ;
+    static std::string WHITE ;
+    static std::string NORMAL;
     
-    Logger(std::string topNm, int on, std::string nm)
-    : active(on), name(nm), topName(topNm) {};
+ Logger(std::string topNm, int on, std::string nm,std::string col)
+   : active(on), name(nm), topName(topNm), COLOUR(col) {};
     
     void Active(int on) {active = on;};
     int  isActive(void) {return active;};
@@ -25,9 +69,10 @@ public:
             StopWatch.Stop();
             GridTime now = StopWatch.Elapsed();
             StopWatch.Start();
-            stream << std::setw(8) << std::left << log.topName << " : ";
-            stream << std::setw(12) << std::left << log.name << " : ";
-            stream << now << " : ";
+            stream << BLACK <<std::setw(8) << std::left << log.topName << BLACK<< " : ";
+            stream << log.COLOUR <<std::setw(11)  << log.name << BLACK << " : ";
+            stream << YELLOW <<std::setw(6) << now <<BLACK << " : " ;
+            stream << log.COLOUR;
             return stream;
         } else { 
             return devnull;
@@ -38,7 +83,7 @@ public:
     
 class GridLogger: public Logger {
 public:
-  GridLogger(int on, std::string nm): Logger("Grid", on, nm){};
+ GridLogger(int on, std::string nm, std::string col = Logger::BLACK): Logger("Grid", on, nm, col){};
 };
 
 void GridLogConfigure(std::vector<std::string> &logstreams);
@@ -49,6 +94,39 @@ extern GridLogger GridLogMessage;
 extern GridLogger GridLogDebug  ;
 extern GridLogger GridLogPerformance;
 extern GridLogger GridLogIterative  ;
+extern GridLogger GridLogIntegrator  ;
+
+
+#define _NBACKTRACE (256)
+extern void * Grid_backtrace_buffer[_NBACKTRACE];
+
+#define BACKTRACEFILE() {\
+    char string[20];					\
+    std::sprintf(string,"backtrace.%d",Rank());				\
+    std::FILE * fp = std::fopen(string,"w");				\
+    BACKTRACEFP(fp)\
+    std::fclose(fp);	    \
+}
+
+
+#ifdef HAVE_EXECINFO_H
+#define BACKTRACEFP(fp) { \
+  int symbols    = backtrace        (Grid_backtrace_buffer,_NBACKTRACE);\
+  char **strings = backtrace_symbols(Grid_backtrace_buffer,symbols);\
+  for (int i = 0; i < symbols; i++){\
+    std::fprintf (fp,"BackTrace Strings: %d %s\n",i, strings[i]); std::fflush(fp); \
+  }\
+}
+#else 
+#define BACKTRACEFP(fp) { \
+    std::fprintf (fp,"BT %d %lx\n",0, __builtin_return_address(0)); std::fflush(fp); \
+    std::fprintf (fp,"BT %d %lx\n",1, __builtin_return_address(1)); std::fflush(fp); \
+    std::fprintf (fp,"BT %d %lx\n",2, __builtin_return_address(2)); std::fflush(fp); \
+    std::fprintf (fp,"BT %d %lx\n",3, __builtin_return_address(3)); std::fflush(fp); \
+}
+#endif
+
+#define BACKTRACE() BACKTRACEFP(stdout) 
 
 }
 #endif

lib/Makefile.am

@@ -6,6 +6,10 @@ if BUILD_COMMS_MPI
   extra_sources+=communicator/Communicator_mpi.cc
 endif
 
+if BUILD_COMMS_SHMEM
+  extra_sources+=communicator/Communicator_shmem.cc
+endif
+
 if BUILD_COMMS_NONE
   extra_sources+=communicator/Communicator_none.cc
 endif

lib/Old/Tensor_peek.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Old/Tensor_peek.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_MATH_PEEK_H
 #define GRID_MATH_PEEK_H
 namespace Grid {

lib/Old/Tensor_poke.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Old/Tensor_poke.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_MATH_POKE_H
 #define GRID_MATH_POKE_H
 namespace Grid {

lib/PerfCount.cc

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/PerfCount.cc
+
+    Copyright (C) 2015
+
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 
 #include <Grid.h>
 #include <PerfCount.h>

lib/PerfCount.h

@@ -1,3 +1,32 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/PerfCount.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <peterboyle@MacBook-Pro.local>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_PERFCOUNT_H
 #define GRID_PERFCOUNT_H
 
@@ -5,7 +34,7 @@
 #include <ctime>
 #include <chrono>
 #include <string.h>
-
+#include <unistd.h>
 #include <sys/ioctl.h>
 
 #ifdef __linux__
@@ -14,8 +43,8 @@
 #else
 #include <sys/syscall.h>
 #endif
-namespace Grid {
 
+namespace Grid {
 
 #ifdef __linux__
 static long perf_event_open(struct perf_event_attr *hw_event, pid_t pid,
@@ -29,6 +58,26 @@ static long perf_event_open(struct perf_event_attr *hw_event, pid_t pid,
 }
 #endif
 
+#ifdef __bgq__
+inline uint64_t cyclecount(void){ 
+   uint64_t tmp;
+   asm volatile ("mfspr %0,0x10C" : "=&r" (tmp)  );
+   return tmp;
+}
+#elif defined __x86_64__
+#include <immintrin.h>
+#ifndef __INTEL_COMPILER
+#include <x86intrin.h>
+#endif
+inline uint64_t cyclecount(void){
+   return __rdtsc();
+}
+#else
+#warning No cycle counter implemented for this architecture
+inline uint64_t cyclecount(void){ 
+   return 0;
+}
+#endif
 
 class PerformanceCounter {
 private:
@@ -73,7 +122,7 @@ public:
 
   long long count;
   int fd;
-  uint64_t elapsed;
+  unsigned long long elapsed;
   uint64_t begin;
 
   static int NumTypes(void){ 
@@ -117,10 +166,10 @@ public:
   {
 #ifdef __linux__
     if ( fd!= -1) {
-      ioctl(fd, PERF_EVENT_IOC_RESET, 0);
-      ioctl(fd, PERF_EVENT_IOC_ENABLE, 0);
+      ::ioctl(fd, PERF_EVENT_IOC_RESET, 0);
+      ::ioctl(fd, PERF_EVENT_IOC_ENABLE, 0);
     }
-    begin  =__rdtsc();
+    begin  =cyclecount();
 #else
     begin = 0;
 #endif
@@ -130,10 +179,10 @@ public:
     count=0;
 #ifdef __linux__
     if ( fd!= -1) {
-      ioctl(fd, PERF_EVENT_IOC_DISABLE, 0);
+      ::ioctl(fd, PERF_EVENT_IOC_DISABLE, 0);
       ::read(fd, &count, sizeof(long long));
     }
-    elapsed = __rdtsc() - begin;
+    elapsed = cyclecount() - begin;
 #else
     elapsed = 0;
 #endif
@@ -141,16 +190,16 @@ public:
   }
   void Report(void) {
 #ifdef __linux__
-    printf("%llu cycles %s = %20llu\n", elapsed , PerformanceCounterConfigs[PCT].name, count);
+    std::printf("%llu cycles %s = %20llu\n", elapsed , PerformanceCounterConfigs[PCT].name, count);
 #else
-    printf("%llu cycles \n", elapsed );
+    std::printf("%llu cycles \n", elapsed );
 #endif
   }
 
   ~PerformanceCounter()
   {
 #ifdef __linux__
-    close(fd);
+    ::close(fd);
 #endif
   }
 

lib/Simd.h

@@ -1,3 +1,32 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Simd.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: neo <cossu@post.kek.jp>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_SIMD_H
 #define GRID_SIMD_H
 
@@ -13,10 +42,13 @@
 
 
 #define _MM_SELECT_FOUR_FOUR(A,B,C,D) ((A<<6)|(B<<4)|(C<<2)|(D))
+#define _MM_SELECT_FOUR_FOUR_STRING(A,B,C,D) "((" #A "<<6)|(" #B "<<4)|(" #C "<<2)|(" #D "))"
 #define _MM_SELECT_EIGHT_TWO(A,B,C,D,E,F,G,H) ((A<<7)|(B<<6)|(C<<5)|(D<<4)|(E<<3)|(F<<2)|(G<<4)|(H))
 #define _MM_SELECT_FOUR_TWO (A,B,C,D) _MM_SELECT_EIGHT_TWO(0,0,0,0,A,B,C,D)
 #define _MM_SELECT_TWO_TWO  (A,B)     _MM_SELECT_FOUR_TWO(0,0,A,B)
 
+#define RotateBit (0x100)
+
 namespace Grid {
 
   typedef uint32_t Integer;

lib/Stencil.h

@@ -1,6 +1,35 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Stencil.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_STENCIL_H
 #define GRID_STENCIL_H
 
+#include <thread>
+
 #include <stencil/Lebesgue.h>   // subdir aggregate
 
 //////////////////////////////////////////////////////////////////////////////////////////
@@ -48,7 +77,7 @@ namespace Grid {
     int _around_the_world;
   };
 
-  template<class vobj,class cobj, class compressor>
+  template<class vobj,class cobj>
   class CartesianStencil { // Stencil runs along coordinate axes only; NO diagonal fill in.
   public:
 
@@ -57,10 +86,158 @@ namespace Grid {
       typedef typename cobj::scalar_type scalar_type;
       typedef typename cobj::scalar_object scalar_object;
 
+      //////////////////////////////////////////
+      // Comms packet queue for asynch thread
+      //////////////////////////////////////////
+
+      struct Packet {
+	void * send_buf;
+	void * recv_buf;
+	Integer to_rank;
+	Integer from_rank;
+	Integer bytes;
+	volatile Integer done;
+      };
+
+      std::vector<Packet> Packets;
+
+#define SEND_IMMEDIATE
+#define SERIAL_SENDS
+
+      void AddPacket(void *xmit,void * rcv, Integer to,Integer from,Integer bytes){
+	comms_bytes+=2.0*bytes;
+#ifdef SEND_IMMEDIATE
+	commtime-=usecond();
+	_grid->SendToRecvFrom(xmit,to,rcv,from,bytes);
+	commtime+=usecond();
+#endif
+	Packet p;
+	p.send_buf = xmit;
+	p.recv_buf = rcv;
+	p.to_rank  = to;
+	p.from_rank= from;
+	p.bytes    = bytes;
+	p.done     = 0;
+	comms_bytes+=2.0*bytes;
+	Packets.push_back(p);
+
+      }
+
+#ifdef SERIAL_SENDS
+      void Communicate(void ) { 
+	commtime-=usecond();
+	for(int i=0;i<Packets.size();i++){
+#ifndef SEND_IMMEDIATE
+	  _grid->SendToRecvFrom(
+				Packets[i].send_buf,
+				Packets[i].to_rank,
+				Packets[i].recv_buf,
+				Packets[i].from_rank,
+				Packets[i].bytes);
+#endif
+	  Packets[i].done = 1;
+	}
+	commtime+=usecond();
+      }
+#else
+      void Communicate(void ) { 
+	typedef CartesianCommunicator::CommsRequest_t CommsRequest_t;
+	std::vector<std::vector<CommsRequest_t> > reqs(Packets.size());
+	commtime-=usecond();
+	const int concurrency=2;
+	for(int i=0;i<Packets.size();i+=concurrency){
+	  for(int ii=0;ii<concurrency;ii++){
+	    int j = i+ii;
+	    if ( j<Packets.size() ) {
+#ifndef SEND_IMMEDIATE
+	      _grid->SendToRecvFromBegin(reqs[j],
+					 Packets[j].send_buf,
+					 Packets[j].to_rank,
+					 Packets[j].recv_buf,
+					 Packets[j].from_rank,
+					 Packets[j].bytes);
+#endif
+	    }
+	  }
+	  for(int ii=0;ii<concurrency;ii++){
+	    int j = i+ii;
+	    if ( j<Packets.size() ) {
+#ifndef SEND_IMMEDIATE
+	      _grid->SendToRecvFromComplete(reqs[i]);
+#endif
+	    }
+	  }
+	  for(int ii=0;ii<concurrency;ii++){
+	    int j = i+ii;
+	    if ( j<Packets.size() ) {
+	      Packets[j].done = 1;
+	    }
+	  }
+	}
+	commtime+=usecond();
+      }
+#endif
+
+      ///////////////////////////////////////////
+      // Simd merge queue for asynch comms
+      ///////////////////////////////////////////
+      struct Merge {
+        cobj * mpointer;
+	std::vector<scalar_object *> rpointers;
+	Integer buffer_size;
+	Integer packet_id;
+      };
+
+      std::vector<Merge> Mergers;
+
+      void AddMerge(cobj *merge_p,std::vector<scalar_object *> &rpointers,Integer buffer_size,Integer packet_id) {
+	Merge m;
+	m.mpointer = merge_p;
+	m.rpointers= rpointers;
+	m.buffer_size = buffer_size;
+	m.packet_id   = packet_id;
+#ifdef SEND_IMMEDIATE
+	mergetime-=usecond();
+PARALLEL_FOR_LOOP
+        for(int o=0;o<m.buffer_size;o++){
+	  merge1(m.mpointer[o],m.rpointers,o);
+	}
+	mergetime+=usecond();
+#else
+	Mergers.push_back(m);
+#endif
+
+      }
+
+      void CommsMerge(void ) { 
+	//PARALLEL_NESTED_LOOP2 
+	for(int i=0;i<Mergers.size();i++){	
+	  
+	spintime-=usecond();
+	int packet_id = Mergers[i].packet_id;
+	while(! Packets[packet_id].done ); // spin for completion
+	spintime+=usecond();
+
+#ifndef SEND_IMMEDIATE
+	mergetime-=usecond();
+PARALLEL_FOR_LOOP
+	  for(int o=0;o<Mergers[i].buffer_size;o++){
+	    merge1(Mergers[i].mpointer[o],Mergers[i].rpointers,o);
+	  }
+	mergetime+=usecond();
+#endif
+
+	}
+      }
+
+      ////////////////////////////////////////
+      // Basic Grid and stencil info
+      ////////////////////////////////////////
+
       int                               _checkerboard;
       int                               _npoints; // Move to template param?
       GridBase *                        _grid;
-      
+
       // npoints of these
       std::vector<int>                  _directions;
       std::vector<int>                  _distances;
@@ -69,32 +246,32 @@ namespace Grid {
 
       // npoints x Osites() of these
       std::vector<std::vector<StencilEntry> > _entries;
-
-      // Comms buffers
-      std::vector<std::vector<scalar_object> > send_buf_extract;
-      std::vector<std::vector<scalar_object> > recv_buf_extract;
-      std::vector<scalar_object *> pointers;
-      std::vector<scalar_object *> rpointers;
-      Vector<cobj> send_buf;
-
       inline StencilEntry * GetEntry(int &ptype,int point,int osite) { ptype = _permute_type[point]; return & _entries[point][osite]; }
 
+      // Comms buffers
+      std::vector<Vector<scalar_object> > u_simd_send_buf;
+      std::vector<Vector<scalar_object> > u_simd_recv_buf;
+      Vector<cobj>          u_send_buf;
+      Vector<cobj>          comm_buf;
+      int u_comm_offset;
       int _unified_buffer_size;
-      int _request_count;
 
-      double buftime;
+      /////////////////////////////////////////
+      // Timing info; ugly; possibly temporary
+      /////////////////////////////////////////
+#define TIMING_HACK
+#ifdef TIMING_HACK
+      double jointime;
       double gathertime;
       double commtime;
-      double commstime;
-      double halotime;
-      double scattertime;
+      double halogtime;
       double mergetime;
+      double spintime;
+      double comms_bytes;
       double gathermtime;
       double splicetime;
       double nosplicetime;
-
-
-
+#endif
 
   CartesianStencil(GridBase *grid,
 				     int npoints,
@@ -103,28 +280,29 @@ namespace Grid {
 				     const std::vector<int> &distances) 
     :   _entries(npoints), _permute_type(npoints), _comm_buf_size(npoints)
     {
+#ifdef TIMING_HACK
       gathertime=0;
+      jointime=0;
       commtime=0;
-      commstime=0;
-      halotime=0;
-      scattertime=0;
+      halogtime=0;
       mergetime=0;
+      spintime=0;
       gathermtime=0;
-      buftime=0;
       splicetime=0;
       nosplicetime=0;
-
+      comms_bytes=0;
+#endif
       _npoints = npoints;
       _grid    = grid;
       _directions = directions;
       _distances  = distances;
       _unified_buffer_size=0;
-      _request_count =0;
 
       int osites  = _grid->oSites();
 
-      for(int i=0;i<npoints;i++){
+      for(int ii=0;ii<npoints;ii++){
 
+	int i = ii; // reverse direction to get SIMD comms done first
 	int point = i;
 
 	_entries[i].resize( osites);
@@ -143,6 +321,9 @@ namespace Grid {
 	int simd_layout     = _grid->_simd_layout[dimension];
 	int comm_dim        = _grid->_processors[dimension] >1 ;
 	int splice_dim      = _grid->_simd_layout[dimension]>1 && (comm_dim);
+	int rotate_dim      = _grid->_simd_layout[dimension]>2;
+
+	assert ( (rotate_dim && comm_dim) == false) ; // Do not think spread out is supported
 
 	int sshift[2];
 	
@@ -163,30 +344,35 @@ namespace Grid {
 	  //        So tables are the same whether comm_dim or splice_dim
 	  sshift[0] = _grid->CheckerBoardShiftForCB(_checkerboard,dimension,shift,Even);
 	  sshift[1] = _grid->CheckerBoardShiftForCB(_checkerboard,dimension,shift,Odd);
+
 	  if ( sshift[0] == sshift[1] ) {
 	    Comms(point,dimension,shift,0x3);
-	    //	    std::cout<<"Comms 0x3"<<std::endl;
 	  } else {
 	    Comms(point,dimension,shift,0x1);// if checkerboard is unfavourable take two passes
 	    Comms(point,dimension,shift,0x2);// both with block stride loop iteration
-	    //	    std::cout<<"Comms 0x1 ; 0x2"<<std::endl;
 	  }
 	}
-	//	for(int ss=0;ss<osites;ss++){
-	  //	  std::cout << "point["<<i<<"] "<<ss<<"-> o"<<_entries[i][ss]._offset<<"; l"<<
-	  //	    _entries[i][ss]._is_local<<"; p"<<_entries[i][ss]._permute<<std::endl;
-	//	}
+      }
+      u_send_buf.resize(_unified_buffer_size);
+      comm_buf.resize(_unified_buffer_size);
+      
+      const int Nsimd = grid->Nsimd();
+      u_simd_send_buf.resize(Nsimd);
+      u_simd_recv_buf.resize(Nsimd);
+      for(int l=0;l<Nsimd;l++){
+	u_simd_send_buf[l].resize(_unified_buffer_size);
+	u_simd_recv_buf[l].resize(_unified_buffer_size);
       }
     }
 
-
     void Local     (int point, int dimension,int shiftpm,int cbmask)
     {
       int fd = _grid->_fdimensions[dimension];
       int rd = _grid->_rdimensions[dimension];
       int ld = _grid->_ldimensions[dimension];
       int gd = _grid->_gdimensions[dimension];
-      
+      int ly = _grid->_simd_layout[dimension];
+
       // Map to always positive shift modulo global full dimension.
       int shift = (shiftpm+fd)%fd;
       
@@ -210,14 +396,13 @@ namespace Grid {
 	if ( (shiftpm== 1) && (sx<x)  ) {
 	  wraparound = 1;
 	}
-
 	  
 	int permute_slice=0;
 	if(permute_dim){
 	  int wrap = sshift/rd;
 	  int  num = sshift%rd;
 	  if ( x< rd-num ) permute_slice=wrap;
-	  else permute_slice = 1-wrap;
+	  else permute_slice = (wrap+1)%ly;
 	}
 
   	CopyPlane(point,dimension,x,sx,cbmask,permute_slice,wraparound);
@@ -228,6 +413,7 @@ namespace Grid {
     void Comms     (int point,int dimension,int shiftpm,int cbmask)
     {
       GridBase *grid=_grid;
+      const int Nsimd = grid->Nsimd();
       
       int fd              = _grid->_fdimensions[dimension];
       int ld              = _grid->_ldimensions[dimension];
@@ -236,27 +422,47 @@ namespace Grid {
       int simd_layout     = _grid->_simd_layout[dimension];
       int comm_dim        = _grid->_processors[dimension] >1 ;
       
-      //      assert(simd_layout==1); // Why?
       assert(comm_dim==1);
       int shift = (shiftpm + fd) %fd;
       assert(shift>=0);
       assert(shift<fd);
-      
-      int buffer_size = _grid->_slice_nblock[dimension]*_grid->_slice_block[dimension];
+
+      int buffer_size = _grid->_slice_nblock[dimension]*_grid->_slice_block[dimension]; // done in reduced dims, so SIMD factored
+
       _comm_buf_size[point] = buffer_size; // Size of _one_ plane. Multiple planes may be gathered and
                                            // send to one or more remote nodes.
 
       int cb= (cbmask==0x2)? Odd : Even;
       int sshift= _grid->CheckerBoardShiftForCB(_checkerboard,dimension,shift,cb);
-      
 
       for(int x=0;x<rd;x++){       
 
-	int sx        =  (x+sshift)%rd;
-	int comm_proc = ((x+sshift)/rd)%pd;
-    	int offnode = (comm_proc!= 0);
+	int permute_type=grid->PermuteType(dimension);
+
+	int sx        =  (x+sshift)%rd;
+	
+    	int offnode = 0;
+	if ( simd_layout > 1 ) {
+
+	  for(int i=0;i<Nsimd;i++){
+
+	    int inner_bit = (Nsimd>>(permute_type+1));
+	    int ic= (i&inner_bit)? 1:0;
+	    int my_coor          = rd*ic + x;
+	    int nbr_coor         = my_coor+sshift;
+	    int nbr_proc = ((nbr_coor)/ld) % pd;// relative shift in processors
+
+	    if ( nbr_proc ) { 
+	      offnode =1;
+	    }
+	  }
+	  
+	} else { 
+	  int comm_proc = ((x+sshift)/rd)%pd;
+	  offnode = (comm_proc!= 0);
+	}
+
 
-	//	std::cout << "Stencil shift "<<shift<<" sshift "<<sshift<<" fd "<<fd<<" rd " <<rd<<" offnode "<<offnode<<" sx "<<sx<<std::endl;
 	int wraparound=0;
 	if ( (shiftpm==-1) && (sx>x) && (grid->_processor_coor[dimension]==0) ) {
 	  wraparound = 1;
@@ -274,14 +480,13 @@ namespace Grid {
 	  int words = buffer_size;
 	  if (cbmask != 0x3) words=words>>1;
 	  
-	  //	  GatherPlaneSimple (point,dimension,sx,cbmask);
-	  
 	  int rank           = grid->_processor;
 	  int recv_from_rank;
 	  int xmit_to_rank;
 
 	  int unified_buffer_offset = _unified_buffer_size;
 	  _unified_buffer_size    += words;
+
 	  ScatterPlane(point,dimension,x,cbmask,unified_buffer_offset,wraparound); // permute/extract/merge is done in comms phase
 	  
 	}
@@ -327,7 +532,7 @@ namespace Grid {
 	      _entries[point][lo+o+b]._around_the_world=wrap;
 	    }
 	    
-	    }
+	  }
 	  o +=_grid->_slice_stride[dimension];
 	}
 	
@@ -377,35 +582,36 @@ namespace Grid {
       }
     }
 
-//      CartesianStencil(GridBase *grid,
-//		       int npoints,
-//		       int checkerboard,
-//		       const std::vector<int> &directions,
-//		       const std::vector<int> &distances);
 
+      template<class compressor>
+      std::thread HaloExchangeBegin(const Lattice<vobj> &source,compressor &compress) {
+	Mergers.resize(0); 
+	Packets.resize(0);
+	HaloGather(source,compress);
+        return std::thread([&] { this->Communicate(); });
+      }
 
-      // Add to tables for various cases;  is this mistaken. only local if 1 proc in dim
-      // Can this be avoided with simpler coding of comms?
-   //      void Local     (int point, int dimension,int shift,int cbmask);
-   //      void Comms     (int point, int dimension,int shift,int cbmask);
-   //      void CopyPlane(int point, int dimension,int lplane,int rplane,int cbmask,int permute,int wrap);
-   //      void ScatterPlane (int point,int dimension,int plane,int cbmask,int offset,int wrap);
-
-      // Could allow a functional munging of the halo to another type during the comms.
-      // this could implement the 16bit/32bit/64bit compression.
-      void HaloExchange(const Lattice<vobj> &source,std::vector<cobj,alignedAllocator<cobj> > &u_comm_buf,compressor &compress)
+      template<class compressor>
+      void HaloExchange(const Lattice<vobj> &source,compressor &compress) 
       {
-	// conformable(source._grid,_grid);
-	assert(source._grid==_grid);
-	halotime-=usecond();
-	if (u_comm_buf.size() != _unified_buffer_size ) u_comm_buf.resize(_unified_buffer_size);
-	int u_comm_offset=0;
+	Mergers.resize(0); 
+	Packets.resize(0);
+	HaloGather(source,compress);
+	Communicate();
+	CommsMerge();
+      }
 
-	// Gather all comms buffers
-	for(int point = 0 ; point < _npoints; point++) {
-
-	  compress.Point(point);
+      void HaloExchangeComplete(std::thread &thr) 
+      {
+	CommsMerge(); // spins
+	jointime-=usecond();
+	thr.join();
+	jointime+=usecond();
+      }
 
+      template<class compressor>
+      void HaloGatherDir(const Lattice<vobj> &source,compressor &compress,int point)
+      {
 	  int dimension    = _directions[point];
 	  int displacement = _distances[point];
 	  
@@ -432,35 +638,51 @@ namespace Grid {
 	    if ( sshift[0] == sshift[1] ) {
 	      if (splice_dim) {
 		splicetime-=usecond();
-		GatherStartCommsSimd(source,dimension,shift,0x3,u_comm_buf,u_comm_offset,compress);
+		GatherSimd(source,dimension,shift,0x3,compress);
 		splicetime+=usecond();
 	      } else { 
 		nosplicetime-=usecond();
-		GatherStartComms(source,dimension,shift,0x3,u_comm_buf,u_comm_offset,compress);
+		Gather(source,dimension,shift,0x3,compress);
 		nosplicetime+=usecond();
 	      }
 	    } else {
-	      std::cout << "dim "<<dimension<<"cb "<<_checkerboard<<"shift "<<shift<<" sshift " << sshift[0]<<" "<<sshift[1]<<std::endl;
 	      if(splice_dim){
 		splicetime-=usecond();
-		GatherStartCommsSimd(source,dimension,shift,0x1,u_comm_buf,u_comm_offset,compress);// if checkerboard is unfavourable take two passes
-		GatherStartCommsSimd(source,dimension,shift,0x2,u_comm_buf,u_comm_offset,compress);// both with block stride loop iteration
+		GatherSimd(source,dimension,shift,0x1,compress);// if checkerboard is unfavourable take two passes
+		GatherSimd(source,dimension,shift,0x2,compress);// both with block stride loop iteration
 		splicetime+=usecond();
 	      } else {
 		nosplicetime-=usecond();
-		GatherStartComms(source,dimension,shift,0x1,u_comm_buf,u_comm_offset,compress);
-		GatherStartComms(source,dimension,shift,0x2,u_comm_buf,u_comm_offset,compress);
+		Gather(source,dimension,shift,0x1,compress);
+		Gather(source,dimension,shift,0x2,compress);
 		nosplicetime+=usecond();
 	      }
 	    }
 	  }
-	}
-	halotime+=usecond();
       }
 
-        void GatherStartComms(const Lattice<vobj> &rhs,int dimension,int shift,int cbmask,
-			      std::vector<cobj,alignedAllocator<cobj> > &u_comm_buf,
-			      int &u_comm_offset,compressor & compress)
+      template<class compressor>
+      void HaloGather(const Lattice<vobj> &source,compressor &compress)
+      {
+	// conformable(source._grid,_grid);
+	assert(source._grid==_grid);
+	halogtime-=usecond();
+
+	assert (comm_buf.size() == _unified_buffer_size );
+	u_comm_offset=0;
+
+	// Gather all comms buffers
+	for(int point = 0 ; point < _npoints; point++) {
+	  compress.Point(point);
+	  HaloGatherDir(source,compress,point);
+	}
+
+	assert(u_comm_offset==_unified_buffer_size);
+	halogtime+=usecond();
+      }
+
+      template<class compressor>
+        void Gather(const Lattice<vobj> &rhs,int dimension,int shift,int cbmask,compressor & compress)
 	{
 	  typedef typename cobj::vector_type vector_type;
 	  typedef typename cobj::scalar_type scalar_type;
@@ -481,8 +703,6 @@ namespace Grid {
 
 	  int buffer_size = _grid->_slice_nblock[dimension]*_grid->_slice_block[dimension];
 
-	  if(send_buf.size()<buffer_size) send_buf.resize(buffer_size);
-
 	  int cb= (cbmask==0x2)? Odd : Even;
 	  int sshift= _grid->CheckerBoardShiftForCB(rhs.checkerboard,dimension,shift,cb);
 
@@ -499,38 +719,33 @@ namespace Grid {
 	      int bytes = words * sizeof(cobj);
 
 	      gathertime-=usecond();
-	      Gather_plane_simple (rhs,send_buf,dimension,sx,cbmask,compress);
+	      Gather_plane_simple (rhs,u_send_buf,dimension,sx,cbmask,compress,u_comm_offset);
 	      gathertime+=usecond();
 
 	      int rank           = _grid->_processor;
 	      int recv_from_rank;
 	      int xmit_to_rank;
 	      _grid->ShiftedRanks(dimension,comm_proc,xmit_to_rank,recv_from_rank);
-	      assert (xmit_to_rank != _grid->ThisRank());
+	      assert (xmit_to_rank   != _grid->ThisRank());
 	      assert (recv_from_rank != _grid->ThisRank());
 
 	      //      FIXME Implement asynchronous send & also avoid buffer copy
-	      commtime-=usecond();
-	      _grid->SendToRecvFrom((void *)&send_buf[0],
-				   xmit_to_rank,
-				    (void *)&u_comm_buf[u_comm_offset],
-				   recv_from_rank,
-				   bytes);
-	      commtime+=usecond();
-
+	      AddPacket((void *)&u_send_buf[u_comm_offset],
+			(void *)  &comm_buf[u_comm_offset],
+			xmit_to_rank,
+			recv_from_rank,
+			bytes);
+			
 	      u_comm_offset+=words;
 	    }
 	  }
 	}
 
 
-	void  GatherStartCommsSimd(const Lattice<vobj> &rhs,int dimension,int shift,int cbmask,
-				   std::vector<cobj,alignedAllocator<cobj> > &u_comm_buf,
-				   int &u_comm_offset,compressor &compress)
+      template<class compressor>
+	void  GatherSimd(const Lattice<vobj> &rhs,int dimension,int shift,int cbmask,compressor &compress)
 	{
-	  buftime-=usecond();
 	  const int Nsimd = _grid->Nsimd();
-
 	  
 	  int fd = _grid->_fdimensions[dimension];
 	  int rd = _grid->_rdimensions[dimension];
@@ -540,6 +755,7 @@ namespace Grid {
 	  int comm_dim        = _grid->_processors[dimension] >1 ;
 
 	  assert(comm_dim==1);
+	  // This will not work with a rotate dim
 	  assert(simd_layout==2);
 	  assert(shift>=0);
 	  assert(shift<fd);
@@ -554,22 +770,11 @@ namespace Grid {
 
 	  assert(cbmask==0x3); // Fixme think there is a latent bug if not true
 
-	  //	Should grow to max size and then cost very little thereafter
-	  send_buf_extract.resize(Nsimd);
-	  recv_buf_extract.resize(Nsimd);
-	  for(int l=0;l<Nsimd;l++){
-	    if( send_buf_extract[l].size() < buffer_size) {
-	      send_buf_extract[l].resize(buffer_size);
-	      recv_buf_extract[l].resize(buffer_size);
-	    }
-	  }
-	  pointers.resize(Nsimd);
-	  rpointers.resize(Nsimd);
-
 	  int bytes = buffer_size*sizeof(scalar_object);
-	  
-	  buftime+=usecond();
-	  
+
+	  std::vector<scalar_object *> rpointers(Nsimd);
+	  std::vector<scalar_object *> spointers(Nsimd);
+
 	  ///////////////////////////////////////////
 	  // Work out what to send where
 	  ///////////////////////////////////////////
@@ -585,16 +790,20 @@ namespace Grid {
 	    if ( any_offnode ) {
 
 	      for(int i=0;i<Nsimd;i++){       
-		pointers[i] = &send_buf_extract[i][0];
+		spointers[i] = &u_simd_send_buf[i][u_comm_offset];
 	      }
+
 	      int sx   = (x+sshift)%rd;
 	      
 	      gathermtime-=usecond();
-	      Gather_plane_extract<cobj>(rhs,pointers,dimension,sx,cbmask,compress);
+	      Gather_plane_extract<cobj>(rhs,spointers,dimension,sx,cbmask,compress);
 	      gathermtime+=usecond();
 
 	      for(int i=0;i<Nsimd;i++){
 		
+		// FIXME 
+		// This logic is hard coded to simd_layout ==2 and not allowing >2
+		//		std::cout << "GatherSimd : lane 1st elem " << i << u_simd_send_buf[i ][u_comm_offset]<<std::endl;
 
 		int inner_bit = (Nsimd>>(permute_type+1));
 		int ic= (i&inner_bit)? 1:0;
@@ -607,44 +816,41 @@ namespace Grid {
 		int nbr_ox   = (nbr_lcoor%rd);    // outer coord of peer
 		int nbr_lane = (i&(~inner_bit));
 		
-		int recv_from_rank;
-		int xmit_to_rank;
-		
 		if (nbr_ic) nbr_lane|=inner_bit;
 		assert (sx == nbr_ox);
 
-		
+		auto rp = &u_simd_recv_buf[i       ][u_comm_offset];
+		auto sp = &u_simd_send_buf[nbr_lane][u_comm_offset];
+
+		void *vrp = (void *)rp;
+		void *vsp = (void *)sp;
+
+
 		if(nbr_proc){
 		  
+		  int recv_from_rank;
+		  int xmit_to_rank;
+
 		  _grid->ShiftedRanks(dimension,nbr_proc,xmit_to_rank,recv_from_rank); 
 		  
-		  commstime-=usecond();
-		  _grid->SendToRecvFrom((void *)&send_buf_extract[nbr_lane][0],
-					xmit_to_rank,
-					(void *)&recv_buf_extract[i][0],
-					recv_from_rank,
-					bytes);
-		  commstime+=usecond();
+		  AddPacket( vsp,vrp,xmit_to_rank,recv_from_rank,bytes);
 		  
-		  rpointers[i] = &recv_buf_extract[i][0];
+		  rpointers[i] = rp;
 
 		} else { 
-		  rpointers[i] = &send_buf_extract[nbr_lane][0];
+
+		  rpointers[i] = sp;
+
 		}
 	      }
 
-	      // Here we don't want to scatter, just place into a buffer.
-	      mergetime-=usecond();
-PARALLEL_FOR_LOOP
-	      for(int i=0;i<buffer_size;i++){
-		//		assert(u_comm_offset+i<_unified_buffer_size);
-		merge(u_comm_buf[u_comm_offset+i],rpointers,i);
-	      }
-	      mergetime+=usecond();
-	      u_comm_offset+=buffer_size;
+	      AddMerge(&comm_buf[u_comm_offset],rpointers,buffer_size,Packets.size()-1);
+
+	      u_comm_offset     +=buffer_size;
 	    }
 	  }
 	}
+
   };
 }
 #endif

lib/Tensors.h

@@ -1,3 +1,32 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Tensors.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: neo <cossu@post.kek.jp>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_MATH_H
 #define GRID_MATH_H
 

lib/Threads.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Threads.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_THREADS_H
 #define GRID_THREADS_H
 

lib/Timer.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/Timer.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_TIME_H
 #define GRID_TIME_H
 
@@ -16,13 +44,19 @@ double usecond(void);
 typedef  std::chrono::system_clock          GridClock;
 typedef  std::chrono::time_point<GridClock> GridTimePoint;
 typedef  std::chrono::milliseconds          GridTime;
+typedef  std::chrono::microseconds          GridUsecs;
 
+inline std::ostream& operator<< (std::ostream & stream, const std::chrono::milliseconds & time)
+{
+  stream << time.count()<<" ms";
+  return stream;
+}
  
 class GridStopWatch {
 private:
   bool running;
   GridTimePoint start;
-  GridTime accumulator;
+  GridUsecs accumulator;
 public:
   GridStopWatch () { 
     Reset();
@@ -34,17 +68,21 @@ public:
   }
   void     Stop(void)  { 
     assert(running == true);
-    accumulator+= std::chrono::duration_cast<GridTime>(GridClock::now()-start); 
+    accumulator+= std::chrono::duration_cast<GridUsecs>(GridClock::now()-start); 
     running = false; 
   };
   void     Reset(void){
     running = false;
     start = GridClock::now();
-    accumulator = std::chrono::duration_cast<GridTime>(start-start); 
+    accumulator = std::chrono::duration_cast<GridUsecs>(start-start); 
   }
   GridTime Elapsed(void) {
     assert(running == false);
-    return accumulator;
+    return std::chrono::duration_cast<GridTime>( accumulator );
+  }
+  uint64_t useconds(void){
+    assert(running == false);
+    return (uint64_t) accumulator.count();
   }
 };
 

lib/algorithms/CoarsenedMatrix.h

@@ -1,3 +1,33 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/CoarsenedMatrix.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: Peter Boyle <peterboyle@Peters-MacBook-Pro-2.local>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef  GRID_ALGORITHM_COARSENED_MATRIX_H
 #define  GRID_ALGORITHM_COARSENED_MATRIX_H
 
@@ -117,6 +147,56 @@ namespace Grid {
       }
       Orthogonalise();
     }
+
+    virtual void CreateSubspaceLanczos(GridParallelRNG  &RNG,LinearOperatorBase<FineField> &hermop,int nn=nbasis) 
+    {
+      // Run a Lanczos with sloppy convergence
+	const int Nstop = nn;
+	const int Nk = nn+20;
+	const int Np = nn+20;
+	const int Nm = Nk+Np;
+	const int MaxIt= 10000;
+	RealD resid = 1.0e-3;
+
+	Chebyshev<FineField> Cheb(0.5,64.0,21);
+	ImplicitlyRestartedLanczos<FineField> IRL(hermop,Cheb,Nstop,Nk,Nm,resid,MaxIt);
+	//	IRL.lock = 1;
+
+	FineField noise(FineGrid); gaussian(RNG,noise);
+	FineField tmp(FineGrid); 
+	std::vector<RealD>     eval(Nm);
+	std::vector<FineField> evec(Nm,FineGrid);
+
+	int Nconv;
+	IRL.calc(eval,evec,
+		 noise,
+		 Nconv);
+
+    	// pull back nn vectors
+	for(int b=0;b<nn;b++){
+
+	  subspace[b]   = evec[b];
+
+	  std::cout << GridLogMessage <<"subspace["<<b<<"] = "<<norm2(subspace[b])<<std::endl;
+
+	  hermop.Op(subspace[b],tmp); 
+	  std::cout<<GridLogMessage << "filtered["<<b<<"] <f|MdagM|f> "<<norm2(tmp)<<std::endl;
+
+	  noise = tmp -  sqrt(eval[b])*subspace[b] ;
+
+	  std::cout<<GridLogMessage << " lambda_"<<b<<" = "<< eval[b] <<"  ;  [ M - Lambda ]_"<<b<<" vec_"<<b<<"  = " <<norm2(noise)<<std::endl;
+
+	  noise = tmp +  eval[b]*subspace[b] ;
+
+	  std::cout<<GridLogMessage << " lambda_"<<b<<" = "<< eval[b] <<"  ;  [ M - Lambda ]_"<<b<<" vec_"<<b<<"  = " <<norm2(noise)<<std::endl;
+
+	}
+	Orthogonalise();
+	for(int b=0;b<nn;b++){
+	  std::cout << GridLogMessage <<"subspace["<<b<<"] = "<<norm2(subspace[b])<<std::endl;
+	}
+    }
+
     virtual void CreateSubspace(GridParallelRNG  &RNG,LinearOperatorBase<FineField> &hermop,int nn=nbasis) {
 
       RealD scale;
@@ -170,11 +250,10 @@ namespace Grid {
     ////////////////////
     Geometry         geom;
     GridBase *       _grid; 
-    CartesianStencil<siteVector,siteVector,SimpleCompressor<siteVector> > Stencil; 
+    CartesianStencil<siteVector,siteVector> Stencil; 
 
     std::vector<CoarseMatrix> A;
 
-    std::vector<siteVector,alignedAllocator<siteVector> >   comm_buf;
       
     ///////////////////////
     // Interface
@@ -187,7 +266,7 @@ namespace Grid {
       conformable(in._grid,out._grid);
 
       SimpleCompressor<siteVector> compressor;
-      Stencil.HaloExchange(in,comm_buf,compressor);
+      Stencil.HaloExchange(in,compressor);
 
 PARALLEL_FOR_LOOP
       for(int ss=0;ss<Grid()->oSites();ss++){
@@ -204,7 +283,7 @@ PARALLEL_FOR_LOOP
 	  } else if(SE->_is_local) { 
 	    nbr = in._odata[SE->_offset];
 	  } else {
-	    nbr = comm_buf[SE->_offset];
+	    nbr = Stencil.comm_buf[SE->_offset];
 	  }
 	  res = res + A[point]._odata[ss]*nbr;
 	}
@@ -228,7 +307,6 @@ PARALLEL_FOR_LOOP
       Stencil(&CoarseGrid,geom.npoint,Even,geom.directions,geom.displacements),
       A(geom.npoint,&CoarseGrid)
     {
-      comm_buf.resize(Stencil._unified_buffer_size);
     };
 
     void CoarsenOperator(GridBase *FineGrid,LinearOperatorBase<Lattice<Fobj> > &linop,

lib/algorithms/LinearOperator.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/LinearOperator.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef  GRID_ALGORITHM_LINEAR_OP_H
 #define  GRID_ALGORITHM_LINEAR_OP_H
 
@@ -194,6 +222,7 @@ namespace Grid {
       SchurDiagMooeeOperator (Matrix &Mat): _Mat(Mat){};
       virtual  RealD Mpc      (const Field &in, Field &out) {
 	Field tmp(in._grid);
+//	std::cout <<"grid pointers: in._grid="<< in._grid << " out._grid=" << out._grid << "  _Mat.Grid=" << _Mat.Grid() << " _Mat.RedBlackGrid=" << _Mat.RedBlackGrid() << std::endl;
 
 	_Mat.Meooe(in,tmp);
 	_Mat.MooeeInv(tmp,out);
@@ -223,10 +252,10 @@ namespace Grid {
       virtual  RealD Mpc      (const Field &in, Field &out) {
 	Field tmp(in._grid);
 
-	_Mat.Meooe(in,tmp);
-	_Mat.MooeeInv(tmp,out);
-	_Mat.Meooe(out,tmp);
-	_Mat.MooeeInv(tmp,out);
+	_Mat.Meooe(in,out);
+	_Mat.MooeeInv(out,tmp);
+	_Mat.Meooe(tmp,out);
+	_Mat.MooeeInv(out,tmp);
 
 	return axpy_norm(out,-1.0,tmp,in);
       }
@@ -242,6 +271,35 @@ namespace Grid {
       }
     };
 
+    template<class Matrix,class Field>
+      class SchurDiagTwoOperator :  public SchurOperatorBase<Field> {
+    protected:
+      Matrix &_Mat;
+    public:
+      SchurDiagTwoOperator (Matrix &Mat): _Mat(Mat){};
+
+      virtual  RealD Mpc      (const Field &in, Field &out) {
+	Field tmp(in._grid);
+
+	_Mat.MooeeInv(in,out);
+	_Mat.Meooe(out,tmp);
+	_Mat.MooeeInv(tmp,out);
+	_Mat.Meooe(out,tmp);
+
+	return axpy_norm(out,-1.0,tmp,in);
+      }
+      virtual  RealD MpcDag   (const Field &in, Field &out){
+	Field tmp(in._grid);
+
+	_Mat.MeooeDag(in,out);
+	_Mat.MooeeInvDag(out,tmp);
+	_Mat.MeooeDag(tmp,out);
+	_Mat.MooeeInvDag(out,tmp);
+
+	return axpy_norm(out,-1.0,tmp,in);
+      }
+    };
+
 
     /////////////////////////////////////////////////////////////
     // Base classes for functions of operators

lib/algorithms/Preconditioner.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/Preconditioner.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_PRECONDITIONER_H
 #define GRID_PRECONDITIONER_H
 

lib/algorithms/SparseMatrix.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/SparseMatrix.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef  GRID_ALGORITHM_SPARSE_MATRIX_H
 #define  GRID_ALGORITHM_SPARSE_MATRIX_H
 

lib/algorithms/approx/Chebyshev.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/approx/Chebyshev.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_CHEBYSHEV_H
 #define GRID_CHEBYSHEV_H
 
@@ -30,13 +58,14 @@ namespace Grid {
       Field Mtmp(in._grid);
       AtoN = in;
       out = AtoN*Coeffs[0];
-      //      std::cout <<"Poly in " <<norm2(in)<<std::endl;
-      //      std::cout <<"0 " <<norm2(out)<<std::endl;
+//            std::cout <<"Poly in " <<norm2(in)<<" size "<< Coeffs.size()<<std::endl;
+//            std::cout <<"Coeffs[0]= "<<Coeffs[0]<< " 0 " <<norm2(out)<<std::endl;
       for(int n=1;n<Coeffs.size();n++){
 	Mtmp = AtoN;
 	Linop.HermOp(Mtmp,AtoN);
 	out=out+AtoN*Coeffs[n];
-	//	std::cout << n<<" " <<norm2(out)<<std::endl;
+//            std::cout <<"Coeffs "<<n<<"= "<< Coeffs[n]<< " 0 " <<std::endl;
+//		std::cout << n<<" " <<norm2(out)<<std::endl;
       }
     };
   };
@@ -54,7 +83,8 @@ namespace Grid {
 
   public:
     void csv(std::ostream &out){
-      for (RealD x=lo; x<hi; x+=(hi-lo)/1000) {
+	RealD diff = hi-lo;
+      for (RealD x=lo-0.2*diff; x<hi+0.2*diff; x+=(hi-lo)/1000) {
 	RealD f = approx(x);
 	out<< x<<" "<<f<<std::endl;
       }
@@ -71,10 +101,24 @@ namespace Grid {
 
     Chebyshev(){};
     Chebyshev(RealD _lo,RealD _hi,int _order, RealD (* func)(RealD) ) {Init(_lo,_hi,_order,func);};
-    
+    Chebyshev(RealD _lo,RealD _hi,int _order) {Init(_lo,_hi,_order);};
+
     ////////////////////////////////////////////////////////////////////////////////////////////////////
     // c.f. numerical recipes "chebft"/"chebev". This is sec 5.8 "Chebyshev approximation".
     ////////////////////////////////////////////////////////////////////////////////////////////////////
+// CJ: the one we need for Lanczos
+    void Init(RealD _lo,RealD _hi,int _order)
+    {
+      lo=_lo;
+      hi=_hi;
+      order=_order;
+      
+      if(order < 2) exit(-1);
+      Coeffs.resize(order);
+      Coeffs.assign(0.,order);
+      Coeffs[order-1] = 1.;
+    };
+
     void Init(RealD _lo,RealD _hi,int _order, RealD (* func)(RealD))
     {
       lo=_lo;
@@ -154,6 +198,8 @@ namespace Grid {
     void operator() (LinearOperatorBase<Field> &Linop, const Field &in, Field &out) {
 
       GridBase *grid=in._grid;
+//std::cout << "Chevyshef(): in._grid="<<in._grid<<std::endl;
+//<<" Linop.Grid()="<<Linop.Grid()<<"Linop.RedBlackGrid()="<<Linop.RedBlackGrid()<<std::endl;
 
       int vol=grid->gSites();
 

lib/algorithms/approx/MultiShiftFunction.cc

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/approx/MultiShiftFunction.cc
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include <Grid.h>
 
 namespace Grid {

lib/algorithms/approx/MultiShiftFunction.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/approx/MultiShiftFunction.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef MULTI_SHIFT_FUNCTION
 #define MULTI_SHIFT_FUNCTION
 

lib/algorithms/approx/Remez.h

@@ -16,9 +16,13 @@
 #define INCLUDED_ALG_REMEZ_H
 
 #include <stddef.h>
+#include <Config.h>
 
-//#include <algorithms/approx/bigfloat.h>
+#ifdef HAVE_GMP_H
+#include <algorithms/approx/bigfloat.h>
+#else
 #include <algorithms/approx/bigfloat_double.h>
+#endif
 
 #define JMAX 10000 //Maximum number of iterations of Newton's approximation
 #define SUM_MAX 10 // Maximum number of terms in exponential

lib/algorithms/approx/bigfloat_double.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/approx/bigfloat_double.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include <math.h>
 
 typedef double mfloat; 

lib/algorithms/iterative/AdefGeneric.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/AdefGeneric.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_ALGORITHMS_ITERATIVE_GENERIC_PCG
 #define GRID_ALGORITHMS_ITERATIVE_GENERIC_PCG
 

lib/algorithms/iterative/ConjugateGradient.h

@@ -1,3 +1,32 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/ConjugateGradient.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_CONJUGATE_GRADIENT_H
 #define GRID_CONJUGATE_GRADIENT_H
 
@@ -30,7 +59,8 @@ public:
       
       //Initial residual computation & set up
       RealD guess = norm2(psi);
-      
+      assert(std::isnan(guess)==0);
+
       Linop.HermOpAndNorm(psi,mmp,d,b);
       
       r= src-mmp;
@@ -54,17 +84,25 @@ public:
 	return;
       }
       
-      std::cout<<GridLogIterative << std::setprecision(4)<< "ConjugateGradient: k=0 residual "<<cp<<" rsq"<<rsq<<std::endl;
-      
+      std::cout<<GridLogIterative << std::setprecision(4)<< "ConjugateGradient: k=0 residual "<<cp<<" target "<<rsq<<std::endl;
+
+      GridStopWatch LinalgTimer;
+      GridStopWatch MatrixTimer;
+      GridStopWatch SolverTimer;
+
+      SolverTimer.Start();
       int k;
       for (k=1;k<=MaxIterations;k++){
 	
 	c=cp;
-	
-	Linop.HermOpAndNorm(p,mmp,d,qq);
 
-	RealD    qqck = norm2(mmp);
-	ComplexD dck  = innerProduct(p,mmp);
+	MatrixTimer.Start();
+	Linop.HermOpAndNorm(p,mmp,d,qq);
+	MatrixTimer.Stop();
+
+	LinalgTimer.Start();
+	//	RealD    qqck = norm2(mmp);
+	//	ComplexD dck  = innerProduct(p,mmp);
       
 	a      = c/d;
 	b_pred = a*(a*qq-d)/c;
@@ -76,11 +114,13 @@ public:
 	psi= a*p+psi;
 	p  = p*b+r;
 	  
-	std::cout<<GridLogIterative<<"ConjugateGradient: Iteration " <<k<<" residual "<<cp<< " target"<< rsq<<std::endl;
+	LinalgTimer.Stop();
+	std::cout<<GridLogIterative<<"ConjugateGradient: Iteration " <<k<<" residual "<<cp<< " target "<< rsq<<std::endl;
 	
 	// Stopping condition
 	if ( cp <= rsq ) { 
 	  
+	  SolverTimer.Stop();
 	  Linop.HermOpAndNorm(psi,mmp,d,qq);
 	  p=mmp-src;
 	  
@@ -92,8 +132,13 @@ public:
 
 	  std::cout<<GridLogMessage<<"ConjugateGradient: Converged on iteration " <<k
 		   <<" computed residual "<<sqrt(cp/ssq)
-		   <<" true residual     "<<true_residual
+		   <<" true residual "    <<true_residual
 		   <<" target "<<Tolerance<<std::endl;
+	  std::cout<<GridLogMessage<<"Time elapsed: Total "<< SolverTimer.Elapsed() << " Matrix  "<<MatrixTimer.Elapsed() << " Linalg "<<LinalgTimer.Elapsed();
+	  std::cout<<std::endl;
+	  
+	  assert(true_residual/Tolerance < 1000.0);
+
 	  return;
 	}
       }

lib/algorithms/iterative/ConjugateGradientMultiShift.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/ConjugateGradientMultiShift.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_CONJUGATE_MULTI_SHIFT_GRADIENT_H
 #define GRID_CONJUGATE_MULTI_SHIFT_GRADIENT_H
 
@@ -215,6 +243,8 @@ void operator() (LinearOperatorBase<Field> &Linop, const Field &src, std::vector
       if ( (!converged[s]) ){
 	
 	RealD css  = c * z[s][iz]* z[s][iz];
+    if((k%100)==0 && (s==0) )
+	    std::cout<<GridLogMessage<<"ConjugateGradientMultiShift k="<<k<<" css " <<css <<std::endl;
 	
 	if(css<rsq[s]){
 	  if ( ! converged[s] )
@@ -246,7 +276,7 @@ void operator() (LinearOperatorBase<Field> &Linop, const Field &src, std::vector
   }
   // ugly hack
   std::cout<<GridLogMessage<<"CG multi shift did not converge"<<std::endl;
-  assert(0);
+//  assert(0);
 }
 
   };

lib/algorithms/iterative/ConjugateResidual.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/ConjugateResidual.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_CONJUGATE_RESIDUAL_H
 #define GRID_CONJUGATE_RESIDUAL_H
 

lib/algorithms/iterative/DenseMatrix.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/DenseMatrix.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_DENSE_MATRIX_H
 #define GRID_DENSE_MATRIX_H
 

lib/algorithms/iterative/EigenSort.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/EigenSort.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_EIGENSORT_H
 #define GRID_EIGENSORT_H
 
@@ -11,32 +38,34 @@ template<class Field>
 class SortEigen {
  private:
   
+//hacking for testing for now
+ private:
   static bool less_lmd(RealD left,RealD right){
-    return fabs(left) < fabs(right);
+    return left > right;
   }  
-  static bool less_pair(std::pair<RealD,Field>& left,
-		 std::pair<RealD,Field>& right){
-    return fabs(left.first) < fabs(right.first);
+  static bool less_pair(std::pair<RealD,Field const*>& left,
+                        std::pair<RealD,Field const*>& right){
+    return left.first > (right.first);
   }  
   
+  
  public:
 
   void push(DenseVector<RealD>& lmd,
-	    DenseVector<Field>& evec,int N) {
-
-    DenseVector<std::pair<RealD, Field> > emod;
-    typename DenseVector<std::pair<RealD, Field> >::iterator it;
+            DenseVector<Field>& evec,int N) {
+    DenseVector<Field> cpy(lmd.size(),evec[0]._grid);
+    for(int i=0;i<lmd.size();i++) cpy[i] = evec[i];
     
-    for(int i=0;i<lmd.size();++i){
-      emod.push_back(std::pair<RealD,Field>(lmd[i],evec[i]));
-    }
+    DenseVector<std::pair<RealD, Field const*> > emod(lmd.size());    
+    for(int i=0;i<lmd.size();++i)
+      emod[i] = std::pair<RealD,Field const*>(lmd[i],&cpy[i]);
 
     partial_sort(emod.begin(),emod.begin()+N,emod.end(),less_pair);
 
-    it=emod.begin();
+    typename DenseVector<std::pair<RealD, Field const*> >::iterator it = emod.begin();
     for(int i=0;i<N;++i){
       lmd[i]=it->first;
-      evec[i]=it->second;
+      evec[i]=*(it->second);
       ++it;
     }
   }

lib/algorithms/iterative/Francis.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/Francis.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef FRANCIS_H
 #define FRANCIS_H
 

lib/algorithms/iterative/Householder.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/Householder.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef HOUSEHOLDER_H
 #define HOUSEHOLDER_H
 

lib/algorithms/iterative/ImplicitlyRestartedLanczos.h

@@ -1,6 +1,38 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/ImplicitlyRestartedLanczos.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_IRL_H
 #define GRID_IRL_H
 
+#include <string.h> //memset
+#ifdef USE_LAPACK
+#include <lapacke.h>
+#endif
 #include <algorithms/iterative/DenseMatrix.h>
 #include <algorithms/iterative/EigenSort.h>
 
@@ -21,6 +53,7 @@ public:
     int Niter;
     int converged;
 
+    int Nstop;   // Number of evecs checked for convergence
     int Nk;      // Number of converged sought
     int Np;      // Np -- Number of spare vecs in kryloc space
     int Nm;      // Nm -- total number of vectors
@@ -29,6 +62,8 @@ public:
 
     SortEigen<Field> _sort;
 
+//    GridCartesian &_fgrid;
+
     LinearOperatorBase<Field> &_Linop;
 
     OperatorFunction<Field>   &_poly;
@@ -39,7 +74,27 @@ public:
     void init(void){};
     void Abort(int ff, DenseVector<RealD> &evals,  DenseVector<DenseVector<RealD> > &evecs);
 
-    ImplicitlyRestartedLanczos(LinearOperatorBase<Field> &Linop, // op
+    ImplicitlyRestartedLanczos(
+				LinearOperatorBase<Field> &Linop, // op
+			       OperatorFunction<Field> & poly,   // polynmial
+			       int _Nstop, // sought vecs
+			       int _Nk, // sought vecs
+			       int _Nm, // spare vecs
+			       RealD _eresid, // resid in lmdue deficit 
+			       int _Niter) : // Max iterations
+      _Linop(Linop),
+      _poly(poly),
+      Nstop(_Nstop),
+      Nk(_Nk),
+      Nm(_Nm),
+      eresid(_eresid),
+      Niter(_Niter)
+    { 
+      Np = Nm-Nk; assert(Np>0);
+    };
+
+    ImplicitlyRestartedLanczos(
+				LinearOperatorBase<Field> &Linop, // op
 			       OperatorFunction<Field> & poly,   // polynmial
 			       int _Nk, // sought vecs
 			       int _Nm, // spare vecs
@@ -47,6 +102,7 @@ public:
 			       int _Niter) : // Max iterations
       _Linop(Linop),
       _poly(poly),
+      Nstop(_Nk),
       Nk(_Nk),
       Nm(_Nm),
       eresid(_eresid),
@@ -114,10 +170,11 @@ public:
       RealD beta = normalise(w); // 6. βk+1 := ∥wk∥2. If βk+1 = 0 then Stop
                                  // 7. vk+1 := wk/βk+1
 
+//	std::cout << "alpha = " << zalph << " beta "<<beta<<std::endl;
       const RealD tiny = 1.0e-20;
       if ( beta < tiny ) { 
 	std::cout << " beta is tiny "<<beta<<std::endl;
-      }
+     }
       lmd[k] = alph;
       lme[k]  = beta;
 
@@ -191,15 +248,122 @@ public:
       }
     }
 
+#ifdef USE_LAPACK
+    void diagonalize_lapack(DenseVector<RealD>& lmd,
+		     DenseVector<RealD>& lme, 
+		     int N1,
+		     int N2,
+		     DenseVector<RealD>& Qt,
+		     GridBase *grid){
+  const int size = Nm;
+//  tevals.resize(size);
+//  tevecs.resize(size);
+  int NN = N1;
+  double evals_tmp[NN];
+  double evec_tmp[NN][NN];
+  memset(evec_tmp[0],0,sizeof(double)*NN*NN);
+//  double AA[NN][NN];
+  double DD[NN];
+  double EE[NN];
+  for (int i = 0; i< NN; i++)
+    for (int j = i - 1; j <= i + 1; j++)
+      if ( j < NN && j >= 0 ) {
+        if (i==j) DD[i] = lmd[i];
+        if (i==j) evals_tmp[i] = lmd[i];
+        if (j==(i-1)) EE[j] = lme[j];
+      }
+  int evals_found;
+  int lwork = ( (18*NN) > (1+4*NN+NN*NN)? (18*NN):(1+4*NN+NN*NN)) ;
+  int liwork =  3+NN*10 ;
+  int iwork[liwork];
+  double work[lwork];
+  int isuppz[2*NN];
+  char jobz = 'V'; // calculate evals & evecs
+  char range = 'I'; // calculate all evals
+  //    char range = 'A'; // calculate all evals
+  char uplo = 'U'; // refer to upper half of original matrix
+  char compz = 'I'; // Compute eigenvectors of tridiagonal matrix
+  int ifail[NN];
+  int info;
+//  int total = QMP_get_number_of_nodes();
+//  int node = QMP_get_node_number();
+//  GridBase *grid = evec[0]._grid;
+  int total = grid->_Nprocessors;
+  int node = grid->_processor;
+  int interval = (NN/total)+1;
+  double vl = 0.0, vu = 0.0;
+  int il = interval*node+1 , iu = interval*(node+1);
+  if (iu > NN)  iu=NN;
+  double tol = 0.0;
+    if (1) {
+      memset(evals_tmp,0,sizeof(double)*NN);
+      if ( il <= NN){
+        printf("total=%d node=%d il=%d iu=%d\n",total,node,il,iu);
+        LAPACK_dstegr(&jobz, &range, &NN,
+            (double*)DD, (double*)EE,
+            &vl, &vu, &il, &iu, // these four are ignored if second parameteris 'A'
+            &tol, // tolerance
+            &evals_found, evals_tmp, (double*)evec_tmp, &NN,
+            isuppz,
+            work, &lwork, iwork, &liwork,
+            &info);
+        for (int i = iu-1; i>= il-1; i--){
+          printf("node=%d evals_found=%d evals_tmp[%d] = %g\n",node,evals_found, i - (il-1),evals_tmp[i - (il-1)]);
+          evals_tmp[i] = evals_tmp[i - (il-1)];
+          if (il>1) evals_tmp[i-(il-1)]=0.;
+          for (int j = 0; j< NN; j++){
+            evec_tmp[i][j] = evec_tmp[i - (il-1)][j];
+            if (il>1) evec_tmp[i-(il-1)][j]=0.;
+          }
+        }
+      }
+      {
+//        QMP_sum_double_array(evals_tmp,NN);
+//        QMP_sum_double_array((double *)evec_tmp,NN*NN);
+         grid->GlobalSumVector(evals_tmp,NN);
+         grid->GlobalSumVector((double*)evec_tmp,NN*NN);
+      }
+    } 
+// cheating a bit. It is better to sort instead of just reversing it, but the document of the routine says evals are sorted in increasing order. qr gives evals in decreasing order.
+  for(int i=0;i<NN;i++){
+    for(int j=0;j<NN;j++)
+      Qt[(NN-1-i)*N2+j]=evec_tmp[i][j];
+      lmd [NN-1-i]=evals_tmp[i];
+  }
+}
+#endif
+
+
     void diagonalize(DenseVector<RealD>& lmd,
 		     DenseVector<RealD>& lme, 
-		     int Nm2,
-		     int Nm,
-		     DenseVector<RealD>& Qt)
+		     int N2,
+		     int N1,
+		     DenseVector<RealD>& Qt,
+		     GridBase *grid)
     {
-      int Niter = 100*Nm;
+
+#ifdef USE_LAPACK
+    const int check_lapack=0; // just use lapack if 0, check against lapack if 1
+
+    if(!check_lapack)
+	return diagonalize_lapack(lmd,lme,N2,N1,Qt,grid);
+
+	DenseVector <RealD> lmd2(N1);
+	DenseVector <RealD> lme2(N1);
+	DenseVector<RealD> Qt2(N1*N1);
+         for(int k=0; k<N1; ++k){
+	    lmd2[k] = lmd[k];
+	    lme2[k] = lme[k];
+	  }
+         for(int k=0; k<N1*N1; ++k)
+	Qt2[k] = Qt[k];
+
+//	diagonalize_lapack(lmd2,lme2,Nm2,Nm,Qt,grid);
+#endif
+
+      int Niter = 100*N1;
       int kmin = 1;
-      int kmax = Nk;
+      int kmax = N2;
       // (this should be more sophisticated)
 
       for(int iter=0; iter<Niter; ++iter){
@@ -211,7 +375,7 @@ public:
 	// (Dsh: shift)
 	
 	// transformation
-	qr_decomp(lmd,lme,Nk,Nm,Qt,Dsh,kmin,kmax);
+	qr_decomp(lmd,lme,N2,N1,Qt,Dsh,kmin,kmax);
 	
 	// Convergence criterion (redef of kmin and kamx)
 	for(int j=kmax-1; j>= kmin; --j){
@@ -222,6 +386,23 @@ public:
 	  }
 	}
 	Niter = iter;
+#ifdef USE_LAPACK
+    if(check_lapack){
+	const double SMALL=1e-8;
+	diagonalize_lapack(lmd2,lme2,N2,N1,Qt2,grid);
+	DenseVector <RealD> lmd3(N2);
+         for(int k=0; k<N2; ++k) lmd3[k]=lmd[k];
+        _sort.push(lmd3,N2);
+        _sort.push(lmd2,N2);
+         for(int k=0; k<N2; ++k){
+	    if (fabs(lmd2[k] - lmd3[k]) >SMALL)  std::cout <<"lmd(qr) lmd(lapack) "<< k << ": " << lmd2[k] <<" "<< lmd3[k] <<std::endl;
+//	    if (fabs(lme2[k] - lme[k]) >SMALL)  std::cout <<"lme(qr)-lme(lapack) "<< k << ": " << lme2[k] - lme[k] <<std::endl;
+	  }
+         for(int k=0; k<N1*N1; ++k){
+//	    if (fabs(Qt2[k] - Qt[k]) >SMALL)  std::cout <<"Qt(qr)-Qt(lapack) "<< k << ": " << Qt2[k] - Qt[k] <<std::endl;
+	}
+    }
+#endif
 	return;
 
       continued:
@@ -237,6 +418,7 @@ public:
       abort();
     }
 
+#if 1
     static RealD normalise(Field& v) 
     {
       RealD nn = norm2(v);
@@ -298,6 +480,7 @@ until convergence
       {
 
 	GridBase *grid = evec[0]._grid;
+	assert(grid == src._grid);
 
 	std::cout << " -- Nk = " << Nk << " Np = "<< Np << std::endl;
 	std::cout << " -- Nm = " << Nm << std::endl;
@@ -328,11 +511,21 @@ until convergence
 	// (uniform vector) Why not src??
 	//	evec[0] = 1.0;
 	evec[0] = src;
+	std:: cout <<"norm2(src)= " << norm2(src)<<std::endl;
+// << src._grid  << std::endl;
 	normalise(evec[0]);
+	std:: cout <<"norm2(evec[0])= " << norm2(evec[0]) <<std::endl;
+// << evec[0]._grid << std::endl;
 	
 	// Initial Nk steps
 	for(int k=0; k<Nk; ++k) step(eval,lme,evec,f,Nm,k);
+//	std:: cout <<"norm2(evec[1])= " << norm2(evec[1]) << std::endl;
+//	std:: cout <<"norm2(evec[2])= " << norm2(evec[2]) << std::endl;
 	RitzMatrix(evec,Nk);
+	for(int k=0; k<Nk; ++k){
+//	std:: cout <<"eval " << k << " " <<eval[k] << std::endl;
+//	std:: cout <<"lme " << k << " " << lme[k] << std::endl;
+	}
 
 	// Restarting loop begins
 	for(int iter = 0; iter<Niter; ++iter){
@@ -354,20 +547,24 @@ until convergence
 	    lme2[k] = lme[k+k1-1];
 	  }
 	  setUnit_Qt(Nm,Qt);
-	  diagonalize(eval2,lme2,Nm,Nm,Qt);
+	  diagonalize(eval2,lme2,Nm,Nm,Qt,grid);
 
 	  // sorting
 	  _sort.push(eval2,Nm);
 	  
 	  // Implicitly shifted QR transformations
 	  setUnit_Qt(Nm,Qt);
-	  for(int ip=k2; ip<Nm; ++ip) 
+	  for(int ip=k2; ip<Nm; ++ip){ 
+	std::cout << "qr_decomp "<< ip << " "<< eval2[ip] << std::endl;
 	    qr_decomp(eval,lme,Nm,Nm,Qt,eval2[ip],k1,Nm);
+		
+	}
     
 	  for(int i=0; i<(Nk+1); ++i) B[i] = 0.0;
 	  
 	  for(int j=k1-1; j<k2+1; ++j){
 	    for(int k=0; k<Nm; ++k){
+	    B[j].checkerboard = evec[k].checkerboard;
 	      B[j] += Qt[k+Nm*j] * evec[k];
 	    }
 	  }
@@ -390,21 +587,25 @@ until convergence
 	    lme2[k] = lme[k];
 	  }
 	  setUnit_Qt(Nm,Qt);
-	  diagonalize(eval2,lme2,Nk,Nm,Qt);
+	  diagonalize(eval2,lme2,Nk,Nm,Qt,grid);
 	  
 	  for(int k = 0; k<Nk; ++k) B[k]=0.0;
 	  
 	  for(int j = 0; j<Nk; ++j){
 	    for(int k = 0; k<Nk; ++k){
+	    B[j].checkerboard = evec[k].checkerboard;
 	      B[j] += Qt[k+j*Nm] * evec[k];
 	    }
+//	    std::cout << "norm(B["<<j<<"])="<<norm2(B[j])<<std::endl;
 	  }
+//	_sort.push(eval2,B,Nk);
 
 	  Nconv = 0;
 	  //	  std::cout << std::setiosflags(std::ios_base::scientific);
 	  for(int i=0; i<Nk; ++i){
 
-	    _poly(_Linop,B[i],v);
+//	    _poly(_Linop,B[i],v);
+	    _Linop.HermOp(B[i],v);
 	    
 	    RealD vnum = real(innerProduct(B[i],v)); // HermOp.
 	    RealD vden = norm2(B[i]);
@@ -412,11 +613,13 @@ until convergence
 	    v -= eval2[i]*B[i];
 	    RealD vv = norm2(v);
 	    
+	    std::cout.precision(13);
 	    std::cout << "[" << std::setw(3)<< std::setiosflags(std::ios_base::right) <<i<<"] ";
 	    std::cout << "eval = "<<std::setw(25)<< std::setiosflags(std::ios_base::left)<< eval2[i];
 	    std::cout <<" |H B[i] - eval[i]B[i]|^2 "<< std::setw(25)<< std::setiosflags(std::ios_base::right)<< vv<< std::endl;
 	    
-	    if(vv<eresid*eresid){
+	// change the criteria as evals are supposed to be sorted, all evals smaller(larger) than Nstop should have converged
+	    if((vv<eresid*eresid) && (i == Nconv) ){
 	      Iconv[Nconv] = i;
 	      ++Nconv;
 	    }
@@ -427,7 +630,7 @@ until convergence
 
 	  std::cout<<" #modes converged: "<<Nconv<<std::endl;
 
-	  if( Nconv>=Nk ){
+	  if( Nconv>=Nstop ){
 	    goto converged;
 	  }
 	} // end of iter loop
@@ -436,21 +639,20 @@ until convergence
 	abort();
 	
       converged:
-	// Sorting
-	
-	eval.clear();
-	evec.clear();
-	for(int i=0; i<Nconv; ++i){
-	  eval.push_back(eval2[Iconv[i]]);
-	  evec.push_back(B[Iconv[i]]);
-	}
-	_sort.push(eval,evec,Nconv);
-	
-	std::cout << "\n Converged\n Summary :\n";
-	std::cout << " -- Iterations  = "<< Nconv  << "\n";
-	std::cout << " -- beta(k)     = "<< beta_k << "\n";
-	std::cout << " -- Nconv       = "<< Nconv  << "\n";
-      }
+       // Sorting
+       eval.resize(Nconv);
+       evec.resize(Nconv,grid);
+       for(int i=0; i<Nconv; ++i){
+         eval[i] = eval2[Iconv[i]];
+         evec[i] = B[Iconv[i]];
+       }
+      _sort.push(eval,evec,Nconv);
+
+      std::cout << "\n Converged\n Summary :\n";
+      std::cout << " -- Iterations  = "<< Nconv  << "\n";
+      std::cout << " -- beta(k)     = "<< beta_k << "\n";
+      std::cout << " -- Nconv       = "<< Nconv  << "\n";
+     }
 
     /////////////////////////////////////////////////
     // Adapted from Rudy's lanczos factor routine
@@ -997,6 +1199,7 @@ static void Lock(DenseMatrix<T> &H, 	///Hess mtx
 
   }
 }
+#endif
 
 
  };

lib/algorithms/iterative/Matrix.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/Matrix.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef MATRIX_H
 #define MATRIX_H
 

lib/algorithms/iterative/MatrixUtils.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/MatrixUtils.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_MATRIX_UTILS_H
 #define GRID_MATRIX_UTILS_H
 

lib/algorithms/iterative/NormalEquations.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/NormalEquations.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_NORMAL_EQUATIONS_H
 #define GRID_NORMAL_EQUATIONS_H
 

lib/algorithms/iterative/PrecConjugateResidual.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/PrecConjugateResidual.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_PREC_CONJUGATE_RESIDUAL_H
 #define GRID_PREC_CONJUGATE_RESIDUAL_H
 

lib/algorithms/iterative/PrecGeneralisedConjugateResidual.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/PrecGeneralisedConjugateResidual.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_PREC_GCR_H
 #define GRID_PREC_GCR_H
 
@@ -19,6 +47,10 @@ namespace Grid {
     int mmax;
     int nstep;
     int steps;
+    GridStopWatch PrecTimer;
+    GridStopWatch MatTimer;
+    GridStopWatch LinalgTimer;
+
     LinearFunction<Field> &Preconditioner;
 
    PrecGeneralisedConjugateResidual(RealD tol,Integer maxit,LinearFunction<Field> &Prec,int _mmax,int _nstep) : 
@@ -40,14 +72,24 @@ namespace Grid {
       
       Field r(src._grid);
 
+        PrecTimer.Reset();
+         MatTimer.Reset();
+      LinalgTimer.Reset();
+
+      GridStopWatch SolverTimer;
+      SolverTimer.Start();
+
       steps=0;
       for(int k=0;k<MaxIterations;k++){
 
 	cp=GCRnStep(Linop,src,psi,rsq);
 
-	if ( verbose ) std::cout<<GridLogMessage<<"VPGCR("<<mmax<<","<<nstep<<") "<< steps <<" steps cp = "<<cp<<std::endl;
+	std::cout<<GridLogMessage<<"VPGCR("<<mmax<<","<<nstep<<") "<< steps <<" steps cp = "<<cp<<std::endl;
 
 	if(cp<rsq) {
+
+	  SolverTimer.Stop();
+
 	  Linop.HermOp(psi,r);
 	  axpy(r,-1.0,src,r);
 	  RealD tr = norm2(r);
@@ -55,6 +97,11 @@ namespace Grid {
 		   << " computed residual "<<sqrt(cp/ssq)
 	           << " true residual "    <<sqrt(tr/ssq)
 	           << " target "           <<Tolerance <<std::endl;
+
+	  std::cout<<GridLogMessage<<"VPGCR Time elapsed: Total  "<< SolverTimer.Elapsed() <<std::endl;
+	  std::cout<<GridLogMessage<<"VPGCR Time elapsed: Precon "<<   PrecTimer.Elapsed() <<std::endl;
+	  std::cout<<GridLogMessage<<"VPGCR Time elapsed: Matrix "<<    MatTimer.Elapsed() <<std::endl;
+	  std::cout<<GridLogMessage<<"VPGCR Time elapsed: Linalg "<< LinalgTimer.Elapsed() <<std::endl;
 	  return;
 	}
 
@@ -62,6 +109,7 @@ namespace Grid {
       std::cout<<GridLogMessage<<"Variable Preconditioned GCR did not converge"<<std::endl;
       assert(0);
     }
+
     RealD GCRnStep(LinearOperatorBase<Field> &Linop,const Field &src, Field &psi,RealD rsq){
 
       RealD cp;
@@ -88,24 +136,25 @@ namespace Grid {
       // initial guess x0 is taken as nonzero.
       // r0=src-A x0 = src
       //////////////////////////////////
+      MatTimer.Start();
       Linop.HermOpAndNorm(psi,Az,zAz,zAAz); 
+      MatTimer.Stop();
       r=src-Az;
       
       /////////////////////
       // p = Prec(r)
       /////////////////////
+      PrecTimer.Start();
       Preconditioner(r,z);
+      PrecTimer.Stop();
 
-      std::cout<<GridLogMessage<< " Preconditioner in " << norm2(r)<<std::endl; 
-      std::cout<<GridLogMessage<< " Preconditioner out " << norm2(z)<<std::endl; 
-      
+      MatTimer.Start();
       Linop.HermOp(z,tmp); 
+      MatTimer.Stop();
 
-      std::cout<<GridLogMessage<< " Preconditioner Aout " << norm2(tmp)<<std::endl; 
       ttmp=tmp;
       tmp=tmp-r;
 
-      std::cout<<GridLogMessage<< " Preconditioner resid " << std::sqrt(norm2(tmp)/norm2(r))<<std::endl; 
       /*
       std::cout<<GridLogMessage<<r<<std::endl;
       std::cout<<GridLogMessage<<z<<std::endl;
@@ -113,7 +162,9 @@ namespace Grid {
       std::cout<<GridLogMessage<<tmp<<std::endl;
       */
 
+      MatTimer.Start();
       Linop.HermOpAndNorm(z,Az,zAz,zAAz); 
+      MatTimer.Stop();
 
       //p[0],q[0],qq[0] 
       p[0]= z;
@@ -137,18 +188,22 @@ namespace Grid {
 
 	cp = axpy_norm(r,-a,q[peri_k],r);  
 
-	std::cout<<GridLogMessage<< " VPGCR_step resid" <<sqrt(cp/rsq)<<std::endl; 
 	if((k==nstep-1)||(cp<rsq)){
 	  return cp;
 	}
 
+	std::cout<<GridLogMessage<< " VPGCR_step["<<steps<<"]  resid " <<sqrt(cp/rsq)<<std::endl; 
+
+	PrecTimer.Start();
 	Preconditioner(r,z);// solve Az = r
+	PrecTimer.Stop();
+
+	MatTimer.Start();
 	Linop.HermOpAndNorm(z,Az,zAz,zAAz);
-
-
 	Linop.HermOp(z,tmp);
+	MatTimer.Stop();
         tmp=tmp-r;
-	std::cout<<GridLogMessage<< " Preconditioner resid" <<sqrt(norm2(tmp)/norm2(r))<<std::endl; 
+	std::cout<<GridLogMessage<< " Preconditioner resid " <<sqrt(norm2(tmp)/norm2(r))<<std::endl; 
 
 	q[peri_kp]=Az;
 	p[peri_kp]=z;

lib/algorithms/iterative/SchurRedBlack.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/algorithms/iterative/SchurRedBlack.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_SCHUR_RED_BLACK_H
 #define GRID_SCHUR_RED_BLACK_H
 
@@ -75,7 +102,9 @@ namespace Grid {
 
       pickCheckerboard(Even,src_e,in);
       pickCheckerboard(Odd ,src_o,in);
-
+      pickCheckerboard(Even,sol_e,out);
+      pickCheckerboard(Odd ,sol_o,out);
+    
       /////////////////////////////////////////////////////
       // src_o = Mdag * (source_o - Moe MeeInv source_e)
       /////////////////////////////////////////////////////

lib/cartesian/Cartesian_base.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/cartesian/Cartesian_base.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_CARTESIAN_BASE_H
 #define GRID_CARTESIAN_BASE_H
 
@@ -87,27 +115,11 @@ public:
       for(int d=0;d<_ndimension;d++) idx+=_ostride[d]*ocoor[d];
       return idx;
     }
-    static inline void CoorFromIndex (std::vector<int>& coor,int index,std::vector<int> &dims){
-      int nd= dims.size();
-      coor.resize(nd);
-      for(int d=0;d<nd;d++){
-	coor[d] = index % dims[d];
-	index   = index / dims[d];
-      }
-    }
     inline void oCoorFromOindex (std::vector<int>& coor,int Oindex){
-      CoorFromIndex(coor,Oindex,_rdimensions);
-    }
-    static inline void IndexFromCoor (std::vector<int>& coor,int &index,std::vector<int> &dims){
-      int nd=dims.size();
-      int stride=1;
-      index=0;
-      for(int d=0;d<nd;d++){
-	index = index+stride*coor[d];
-	stride=stride*dims[d];
-      }
+      Lexicographic::CoorFromIndex(coor,Oindex,_rdimensions);
     }
 
+
     //////////////////////////////////////////////////////////
     // SIMD lane addressing
     //////////////////////////////////////////////////////////
@@ -119,13 +131,32 @@ public:
     }
     inline void iCoorFromIindex(std::vector<int> &coor,int lane)
     {
-      CoorFromIndex(coor,lane,_simd_layout);
+      Lexicographic::CoorFromIndex(coor,lane,_simd_layout);
     }
     inline int PermuteDim(int dimension){
       return _simd_layout[dimension]>1;
     }
     inline int PermuteType(int dimension){
       int permute_type=0;
+      //
+      // FIXME:
+      //
+      // Best way to encode this would be to present a mask 
+      // for which simd dimensions are rotated, and the rotation
+      // size. If there is only one simd dimension rotated, this is just 
+      // a permute. 
+      //
+      // Cases: PermuteType == 1,2,4,8
+      // Distance should be either 0,1,2..
+      //
+      if ( _simd_layout[dimension] > 2 ) { 
+	for(int d=0;d<_ndimension;d++){
+	  if ( d != dimension ) assert ( (_simd_layout[d]==1)  );
+	}
+	permute_type = RotateBit; // How to specify distance; this is not just direction.
+	return permute_type;
+      }
+
       for(int d=_ndimension-1;d>dimension;d--){
 	if (_simd_layout[d]>1 ) permute_type++;
       }
@@ -135,12 +166,12 @@ public:
     // Array sizing queries
     ////////////////////////////////////////////////////////////////
 
-    inline int iSites(void) { return _isites; };
-    inline int Nsimd(void)  { return _isites; };// Synonymous with iSites
-    inline int oSites(void) { return _osites; };
-    inline int lSites(void) { return _isites*_osites; }; 
-    inline int gSites(void) { return _isites*_osites*_Nprocessors; }; 
-    inline int Nd    (void) { return _ndimension;};
+    inline int iSites(void) const { return _isites; };
+    inline int Nsimd(void)  const { return _isites; };// Synonymous with iSites
+    inline int oSites(void) const { return _osites; };
+    inline int lSites(void) const { return _isites*_osites; }; 
+    inline int gSites(void) const { return _isites*_osites*_Nprocessors; }; 
+    inline int Nd    (void) const { return _ndimension;};
 
     inline const std::vector<int> &FullDimensions(void)         { return _fdimensions;};
     inline const std::vector<int> &GlobalDimensions(void)       { return _gdimensions;};
@@ -151,7 +182,10 @@ public:
     // Global addressing
     ////////////////////////////////////////////////////////////////
     void GlobalIndexToGlobalCoor(int gidx,std::vector<int> &gcoor){
-      CoorFromIndex(gcoor,gidx,_gdimensions);
+      Lexicographic::CoorFromIndex(gcoor,gidx,_gdimensions);
+    }
+    void LocalIndexToLocalCoor(int lidx,std::vector<int> &lcoor){
+      Lexicographic::CoorFromIndex(lcoor,lidx,_ldimensions);
     }
     void GlobalCoorToGlobalIndex(const std::vector<int> & gcoor,int & gidx){
       gidx=0;

lib/cartesian/Cartesian_full.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/cartesian/Cartesian_full.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_CARTESIAN_FULL_H
 #define GRID_CARTESIAN_FULL_H
 

lib/cartesian/Cartesian_red_black.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/cartesian/Cartesian_red_black.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_CARTESIAN_RED_BLACK_H
 #define GRID_CARTESIAN_RED_BLACK_H
 

lib/communicator/Communicator_base.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/communicator/Communicator_base.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_COMMUNICATOR_BASE_H
 #define GRID_COMMUNICATOR_BASE_H
 
@@ -7,6 +34,9 @@
 #ifdef GRID_COMMS_MPI
 #include <mpi.h>
 #endif
+#ifdef GRID_COMMS_SHMEM
+#include <mpp/shmem.h>
+#endif
 namespace Grid {
 class CartesianCommunicator {
   public:    
@@ -26,6 +56,8 @@ class CartesianCommunicator {
     typedef int CommsRequest_t;
 #endif
 
+    static void Init(int *argc, char ***argv);
+
     // Constructor
     CartesianCommunicator(const std::vector<int> &pdimensions_in);
 
@@ -54,6 +86,7 @@ class CartesianCommunicator {
     void GlobalSumVector(RealD *,int N);
 
     void GlobalSum(uint32_t &);
+    void GlobalSum(uint64_t &);
 
     void GlobalSum(ComplexF &c)
     {
@@ -88,12 +121,11 @@ class CartesianCommunicator {
 			int recv_from_rank,
 			int bytes);
 
-    void RecvFrom(void *recv,
-		  int recv_from_rank,
-		  int bytes);
-    void SendTo(void *xmit,
-		int xmit_to_rank,
-		int bytes);
+    void SendRecvPacket(void *xmit,
+			void *recv,
+			int xmit_to_rank,
+			int recv_from_rank,
+			int bytes);
 
     void SendToRecvFromBegin(std::vector<CommsRequest_t> &list,
 			 void *xmit,

lib/communicator/Communicator_mpi.cc

@@ -1,9 +1,49 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/communicator/Communicator_mpi.cc
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include "Grid.h"
 #include <mpi.h>
 
 namespace Grid {
 
   // Should error check all MPI calls.
+void CartesianCommunicator::Init(int *argc, char ***argv) {
+  int flag;
+  MPI_Initialized(&flag); // needed to coexist with other libs apparently
+  if ( !flag ) {
+    MPI_Init(argc,argv);
+  }
+}
+
+  int Rank(void) {
+    int pe;
+    MPI_Comm_rank(MPI_COMM_WORLD,&pe);
+    return pe;
+  }
 
 CartesianCommunicator::CartesianCommunicator(const std::vector<int> &processors)
 {
@@ -13,6 +53,7 @@ CartesianCommunicator::CartesianCommunicator(const std::vector<int> &processors)
   _Nprocessors=1;
   _processors = processors;
   _processor_coor.resize(_ndimension);
+  std::cout << processors << std::endl;
   
   MPI_Cart_create(MPI_COMM_WORLD, _ndimension,&_processors[0],&periodic[0],1,&communicator);
   MPI_Comm_rank(communicator,&_processor);
@@ -32,6 +73,10 @@ void CartesianCommunicator::GlobalSum(uint32_t &u){
   int ierr=MPI_Allreduce(MPI_IN_PLACE,&u,1,MPI_UINT32_T,MPI_SUM,communicator);
   assert(ierr==0);
 }
+void CartesianCommunicator::GlobalSum(uint64_t &u){
+  int ierr=MPI_Allreduce(MPI_IN_PLACE,&u,1,MPI_UINT64_T,MPI_SUM,communicator);
+  assert(ierr==0);
+}
 void CartesianCommunicator::GlobalSum(float &f){
   int ierr=MPI_Allreduce(MPI_IN_PLACE,&f,1,MPI_FLOAT,MPI_SUM,communicator);
   assert(ierr==0);
@@ -81,21 +126,22 @@ void CartesianCommunicator::SendToRecvFrom(void *xmit,
   SendToRecvFromBegin(reqs,xmit,dest,recv,from,bytes);
   SendToRecvFromComplete(reqs);
 }
-void CartesianCommunicator::RecvFrom(void *recv,
-				     int from,
-				     int bytes) 
+
+void CartesianCommunicator::SendRecvPacket(void *xmit,
+					   void *recv,
+					   int sender,
+					   int receiver,
+					   int bytes)
 {
   MPI_Status stat;
-  int ierr=MPI_Recv(recv, bytes, MPI_CHAR,from,from,communicator,&stat);
-  assert(ierr==0);
-}
-void CartesianCommunicator::SendTo(void *xmit,
-				   int dest,
-				   int bytes)
-{
-  int rank = _processor; // used for tag; must know who it comes from
-  int ierr = MPI_Send(xmit, bytes, MPI_CHAR,dest,_processor,communicator);
-  assert(ierr==0);
+  assert(sender != receiver);
+  int tag = sender;
+  if ( _processor == sender ) {
+    MPI_Send(xmit, bytes, MPI_CHAR,receiver,tag,communicator);
+  }
+  if ( _processor == receiver ) { 
+    MPI_Recv(recv, bytes, MPI_CHAR,sender,tag,communicator,&stat);
+  }
 }
 
 // Basic Halo comms primitive

lib/communicator/Communicator_none.cc

@@ -1,6 +1,39 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/communicator/Communicator_none.cc
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #include "Grid.h"
 namespace Grid {
 
+void CartesianCommunicator::Init(int *argc, char *** arv)
+{
+}
+
+int Rank(void ){ return 0; };
+
 CartesianCommunicator::CartesianCommunicator(const std::vector<int> &processors)
 {
   _processors = processors;
@@ -20,17 +53,14 @@ void CartesianCommunicator::GlobalSum(float &){}
 void CartesianCommunicator::GlobalSumVector(float *,int N){}
 void CartesianCommunicator::GlobalSum(double &){}
 void CartesianCommunicator::GlobalSum(uint32_t &){}
+void CartesianCommunicator::GlobalSum(uint64_t &){}
 void CartesianCommunicator::GlobalSumVector(double *,int N){}
 
-void CartesianCommunicator::RecvFrom(void *recv,
-				     int recv_from_rank,
-				     int bytes) 
-{
-  assert(0);
-}
-void CartesianCommunicator::SendTo(void *xmit,
-				   int xmit_to_rank,
-				   int bytes)
+void CartesianCommunicator::SendRecvPacket(void *xmit,
+					   void *recv,
+					   int xmit_to_rank,
+					   int recv_from_rank,
+					   int bytes)
 {
   assert(0);
 }

lib/communicator/Communicator_shmem.cc

@@ -0,0 +1,334 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/communicator/Communicator_shmem.cc
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
+#include "Grid.h"
+#include <mpp/shmem.h>
+
+namespace Grid {
+
+  // Should error check all MPI calls.
+#define SHMEM_VET(addr) 
+
+#define SHMEM_VET_DEBUG(addr) {				\
+  if ( ! shmem_addr_accessible(addr,_processor) ) {\
+    std::fprintf(stderr,"%d Inaccessible shmem address %lx %s %s\n",_processor,addr,__FUNCTION__,#addr); \
+    BACKTRACEFILE();		   \
+  }\
+}
+int Rank(void) {
+  return shmem_my_pe();
+}
+typedef struct HandShake_t { 
+  uint64_t seq_local;
+  uint64_t seq_remote;
+} HandShake;
+
+static Vector< HandShake > XConnections;
+static Vector< HandShake > RConnections;
+
+void CartesianCommunicator::Init(int *argc, char ***argv) {
+  shmem_init();
+  XConnections.resize(shmem_n_pes());
+  RConnections.resize(shmem_n_pes());
+  for(int pe =0 ; pe<shmem_n_pes();pe++){
+    XConnections[pe].seq_local = 0;
+    XConnections[pe].seq_remote= 0;
+    RConnections[pe].seq_local = 0;
+    RConnections[pe].seq_remote= 0;
+  }
+  shmem_barrier_all();
+}
+CartesianCommunicator::CartesianCommunicator(const std::vector<int> &processors)
+{
+  _ndimension = processors.size();
+  std::vector<int> periodic(_ndimension,1);
+
+  _Nprocessors=1;
+  _processors = processors;
+  _processor_coor.resize(_ndimension);
+
+  _processor = shmem_my_pe();
+  
+  Lexicographic::CoorFromIndex(_processor_coor,_processor,_processors);
+
+  for(int i=0;i<_ndimension;i++){
+    _Nprocessors*=_processors[i];
+  }
+
+  int Size = shmem_n_pes(); 
+
+
+  assert(Size==_Nprocessors);
+}
+
+void CartesianCommunicator::GlobalSum(uint32_t &u){
+  static long long source ;
+  static long long dest   ;
+  static long long llwrk[_SHMEM_REDUCE_MIN_WRKDATA_SIZE];
+  static long      psync[_SHMEM_REDUCE_SYNC_SIZE];
+
+  //  int nreduce=1;
+  //  int pestart=0;
+  //  int logStride=0;
+
+  source = u;
+  dest   = 0;
+  shmem_longlong_sum_to_all(&dest,&source,1,0,0,_Nprocessors,llwrk,psync);
+  shmem_barrier_all(); // necessary?
+  u = dest;
+}
+void CartesianCommunicator::GlobalSum(uint64_t &u){
+  static long long source ;
+  static long long dest   ;
+  static long long llwrk[_SHMEM_REDUCE_MIN_WRKDATA_SIZE];
+  static long      psync[_SHMEM_REDUCE_SYNC_SIZE];
+
+  //  int nreduce=1;
+  //  int pestart=0;
+  //  int logStride=0;
+
+  source = u;
+  dest   = 0;
+  shmem_longlong_sum_to_all(&dest,&source,1,0,0,_Nprocessors,llwrk,psync);
+  shmem_barrier_all(); // necessary?
+  u = dest;
+}
+void CartesianCommunicator::GlobalSum(float &f){
+  static float source ;
+  static float dest   ;
+  static float llwrk[_SHMEM_REDUCE_MIN_WRKDATA_SIZE];
+  static long  psync[_SHMEM_REDUCE_SYNC_SIZE];
+
+  source = f;
+  dest   =0.0;
+  shmem_float_sum_to_all(&dest,&source,1,0,0,_Nprocessors,llwrk,psync);
+  f = dest;
+}
+void CartesianCommunicator::GlobalSumVector(float *f,int N)
+{
+  static float source ;
+  static float dest   = 0 ;
+  static float llwrk[_SHMEM_REDUCE_MIN_WRKDATA_SIZE];
+  static long  psync[_SHMEM_REDUCE_SYNC_SIZE];
+
+  if ( shmem_addr_accessible(f,_processor)  ){
+    shmem_float_sum_to_all(f,f,N,0,0,_Nprocessors,llwrk,psync);
+    return;
+  }
+
+  for(int i=0;i<N;i++){
+    dest   =0.0;
+    source = f[i];
+    shmem_float_sum_to_all(&dest,&source,1,0,0,_Nprocessors,llwrk,psync);
+    f[i] = dest;
+  }
+}
+void CartesianCommunicator::GlobalSum(double &d)
+{
+  static double source;
+  static double dest  ;
+  static double llwrk[_SHMEM_REDUCE_MIN_WRKDATA_SIZE];
+  static long  psync[_SHMEM_REDUCE_SYNC_SIZE];
+
+  source = d;
+  dest   = 0;
+  shmem_double_sum_to_all(&dest,&source,1,0,0,_Nprocessors,llwrk,psync);
+  d = dest;
+}
+void CartesianCommunicator::GlobalSumVector(double *d,int N)
+{
+  static double source ;
+  static double dest   ;
+  static double llwrk[_SHMEM_REDUCE_MIN_WRKDATA_SIZE];
+  static long  psync[_SHMEM_REDUCE_SYNC_SIZE];
+
+  if ( shmem_addr_accessible(d,_processor)  ){
+    shmem_double_sum_to_all(d,d,N,0,0,_Nprocessors,llwrk,psync);
+    return;
+  }
+
+  for(int i=0;i<N;i++){
+    source = d[i];
+    dest   =0.0;
+    shmem_double_sum_to_all(&dest,&source,1,0,0,_Nprocessors,llwrk,psync);
+    d[i] = dest;
+  }
+}
+void CartesianCommunicator::ShiftedRanks(int dim,int shift,int &source,int &dest)
+{
+  std::vector<int> coor = _processor_coor;
+
+  assert(std::abs(shift) <_processors[dim]);
+
+  coor[dim] = (_processor_coor[dim] + shift + _processors[dim])%_processors[dim];
+  Lexicographic::IndexFromCoor(coor,source,_processors);
+
+  coor[dim] = (_processor_coor[dim] - shift + _processors[dim])%_processors[dim];
+  Lexicographic::IndexFromCoor(coor,dest,_processors);
+
+}
+int CartesianCommunicator::RankFromProcessorCoor(std::vector<int> &coor)
+{
+  int rank;
+  Lexicographic::IndexFromCoor(coor,rank,_processors);
+  return rank;
+}
+void  CartesianCommunicator::ProcessorCoorFromRank(int rank, std::vector<int> &coor)
+{
+  Lexicographic::CoorFromIndex(coor,rank,_processors);
+}
+
+// Basic Halo comms primitive
+void CartesianCommunicator::SendToRecvFrom(void *xmit,
+					   int dest,
+					   void *recv,
+					   int from,
+					   int bytes)
+{
+  SHMEM_VET(xmit);
+  SHMEM_VET(recv);
+  std::vector<CommsRequest_t> reqs(0);
+  SendToRecvFromBegin(reqs,xmit,dest,recv,from,bytes);
+  SendToRecvFromComplete(reqs);
+}
+
+void CartesianCommunicator::SendRecvPacket(void *xmit,
+					   void *recv,
+					   int sender,
+					   int receiver,
+					   int bytes)
+{
+  static uint64_t seq;
+
+  assert(recv!=xmit);
+  volatile HandShake *RecvSeq = (volatile HandShake *) & RConnections[sender];
+  volatile HandShake *SendSeq = (volatile HandShake *) & XConnections[receiver];
+
+  if ( _processor == sender ) {
+
+    printf("Sender SHMEM pt2pt %d -> %d\n",sender,receiver);
+    // Check he has posted a receive
+    while(SendSeq->seq_remote == SendSeq->seq_local);
+
+    printf("Sender receive %d posted\n",sender,receiver);
+
+    // Advance our send count
+    seq = ++(SendSeq->seq_local);
+    
+    // Send this packet 
+    SHMEM_VET(recv);
+    shmem_putmem(recv,xmit,bytes,receiver);
+    shmem_fence();
+
+    printf("Sender sent payload %d\n",seq);
+    //Notify him we're done
+    shmem_putmem((void *)&(RecvSeq->seq_remote),&seq,sizeof(seq),receiver);
+    shmem_fence();
+    printf("Sender ringing door bell  %d\n",seq);
+  }
+  if ( _processor == receiver ) {
+
+    printf("Receiver SHMEM pt2pt %d->%d\n",sender,receiver);
+    // Post a receive
+    seq = ++(RecvSeq->seq_local);
+    shmem_putmem((void *)&(SendSeq->seq_remote),&seq,sizeof(seq),sender);
+
+    printf("Receiver Opening letter box %d\n",seq);
+
+    
+    // Now wait until he has advanced our reception counter
+    while(RecvSeq->seq_remote != RecvSeq->seq_local);
+
+    printf("Receiver Got the mail %d\n",seq);
+  }
+}
+
+// Basic Halo comms primitive
+void CartesianCommunicator::SendToRecvFromBegin(std::vector<CommsRequest_t> &list,
+						void *xmit,
+						int dest,
+						void *recv,
+						int from,
+						int bytes)
+{
+  SHMEM_VET(xmit);
+  SHMEM_VET(recv);
+  //  shmem_putmem_nb(recv,xmit,bytes,dest,NULL);
+  shmem_putmem(recv,xmit,bytes,dest);
+}
+void CartesianCommunicator::SendToRecvFromComplete(std::vector<CommsRequest_t> &list)
+{
+  //  shmem_quiet();      // I'm done
+  shmem_barrier_all();// He's done too
+}
+void CartesianCommunicator::Barrier(void)
+{
+  shmem_barrier_all();
+}
+void CartesianCommunicator::Broadcast(int root,void* data, int bytes)
+{
+  static long  psync[_SHMEM_REDUCE_SYNC_SIZE];
+  static uint32_t word;
+  uint32_t *array = (uint32_t *) data;
+  assert( (bytes % 4)==0);
+  int words = bytes/4;
+
+  if ( shmem_addr_accessible(data,_processor)  ){
+    shmem_broadcast32(data,data,words,root,0,0,shmem_n_pes(),psync);
+    return;
+  }
+
+  for(int w=0;w<words;w++){
+    word = array[w];
+    shmem_broadcast32((void *)&word,(void *)&word,1,root,0,0,shmem_n_pes(),psync);
+    if ( shmem_my_pe() != root ) {
+      array[w] = word;
+    }
+    shmem_barrier_all();
+  }
+
+}
+void CartesianCommunicator::BroadcastWorld(int root,void* data, int bytes)
+{
+  static long  psync[_SHMEM_REDUCE_SYNC_SIZE];
+  static uint32_t word;
+  uint32_t *array = (uint32_t *) data;
+  assert( (bytes % 4)==0);
+  int words = bytes/4;
+
+  for(int w=0;w<words;w++){
+    word = array[w];
+    shmem_broadcast32((void *)&word,(void *)&word,1,root,0,0,shmem_n_pes(),psync);
+    if ( shmem_my_pe() != root ) {
+      array[w]= word;
+    }
+    shmem_barrier_all();
+  }
+}
+
+}
+

lib/cshift/Cshift_common.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/cshift/Cshift_common.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef _GRID_CSHIFT_COMMON_H_
 #define _GRID_CSHIFT_COMMON_H_
 
@@ -8,7 +35,7 @@ class SimpleCompressor {
 public:
   void Point(int) {};
 
-  vobj operator() (const vobj &arg,int dimension,int plane,int osite,GridBase *grid) {
+  vobj operator() (const vobj &arg) {
     return arg;
   }
 };
@@ -17,7 +44,7 @@ public:
 // Gather for when there is no need to SIMD split with compression
 ///////////////////////////////////////////////////////////////////
 template<class vobj,class cobj,class compressor> void 
-Gather_plane_simple (const Lattice<vobj> &rhs,std::vector<cobj,alignedAllocator<cobj> > &buffer,int dimension,int plane,int cbmask,compressor &compress)
+Gather_plane_simple (const Lattice<vobj> &rhs,std::vector<cobj,alignedAllocator<cobj> > &buffer,int dimension,int plane,int cbmask,compressor &compress, int off=0)
 {
   int rd = rhs._grid->_rdimensions[dimension];
 
@@ -29,24 +56,24 @@ Gather_plane_simple (const Lattice<vobj> &rhs,std::vector<cobj,alignedAllocator<
   
   int e1=rhs._grid->_slice_nblock[dimension];
   int e2=rhs._grid->_slice_block[dimension];
-
+  int stride=rhs._grid->_slice_stride[dimension];
   if ( cbmask == 0x3 ) { 
 PARALLEL_NESTED_LOOP2
     for(int n=0;n<e1;n++){
       for(int b=0;b<e2;b++){
-	int o  = n*rhs._grid->_slice_stride[dimension];
-	int bo = n*rhs._grid->_slice_block[dimension];
-	buffer[bo+b]=compress(rhs._odata[so+o+b],dimension,plane,so+o+b,rhs._grid);
+	int o  = n*stride;
+	int bo = n*e2;
+	buffer[off+bo+b]=compress(rhs._odata[so+o+b]);
       }
     }
   } else { 
      int bo=0;
      for(int n=0;n<e1;n++){
        for(int b=0;b<e2;b++){
-	 int o  = n*rhs._grid->_slice_stride[dimension];
+	 int o  = n*stride;
 	 int ocb=1<<rhs._grid->CheckerBoardFromOindex(o+b);// Could easily be a table lookup
 	 if ( ocb &cbmask ) {
-	   buffer[bo++]=compress(rhs._odata[so+o+b],dimension,plane,so+o+b,rhs._grid);
+	   buffer[off+bo++]=compress(rhs._odata[so+o+b]);
 	 }
        }
      }
@@ -70,16 +97,16 @@ Gather_plane_extract(const Lattice<vobj> &rhs,std::vector<typename cobj::scalar_
 
   int e1=rhs._grid->_slice_nblock[dimension];
   int e2=rhs._grid->_slice_block[dimension];
-  
+  int n1=rhs._grid->_slice_stride[dimension];
+  int n2=rhs._grid->_slice_block[dimension];
   if ( cbmask ==0x3){
 PARALLEL_NESTED_LOOP2
     for(int n=0;n<e1;n++){
       for(int b=0;b<e2;b++){
 
-	int o=n*rhs._grid->_slice_stride[dimension];
-	int offset = b+n*rhs._grid->_slice_block[dimension];
-
-	cobj temp =compress(rhs._odata[so+o+b],dimension,plane,so+o+b,rhs._grid);
+	int o      =   n*n1;
+	int offset = b+n*n2;
+	cobj temp =compress(rhs._odata[so+o+b]);
 	extract<cobj>(temp,pointers,offset);
 
       }
@@ -94,7 +121,7 @@ PARALLEL_NESTED_LOOP2
 	int offset = b+n*rhs._grid->_slice_block[dimension];
 
 	if ( ocb & cbmask ) {
-	  cobj temp =compress(rhs._odata[so+o+b],dimension,plane,so+o+b,rhs._grid);
+	  cobj temp =compress(rhs._odata[so+o+b]);
 	  extract<cobj>(temp,pointers,offset);
 	}
       }
@@ -216,13 +243,13 @@ template<class vobj> void Copy_plane(Lattice<vobj>& lhs,const Lattice<vobj> &rhs
 
   int e1=rhs._grid->_slice_nblock[dimension]; // clearly loop invariant for icpc
   int e2=rhs._grid->_slice_block[dimension];
-
+  int stride = rhs._grid->_slice_stride[dimension];
   if(cbmask == 0x3 ){
 PARALLEL_NESTED_LOOP2
     for(int n=0;n<e1;n++){
       for(int b=0;b<e2;b++){
  
-        int o =n*rhs._grid->_slice_stride[dimension]+b;
+        int o =n*stride+b;
   	//lhs._odata[lo+o]=rhs._odata[ro+o];
 	vstream(lhs._odata[lo+o],rhs._odata[ro+o]);
       }
@@ -232,7 +259,7 @@ PARALLEL_NESTED_LOOP2
     for(int n=0;n<e1;n++){
       for(int b=0;b<e2;b++){
  
-        int o =n*rhs._grid->_slice_stride[dimension]+b;
+        int o =n*stride+b;
         int ocb=1<<lhs._grid->CheckerBoardFromOindex(o);
         if ( ocb&cbmask ) {
   	//lhs._odata[lo+o]=rhs._odata[ro+o];
@@ -258,11 +285,12 @@ template<class vobj> void Copy_plane_permute(Lattice<vobj>& lhs,const Lattice<vo
 
   int e1=rhs._grid->_slice_nblock[dimension];
   int e2=rhs._grid->_slice_block [dimension];
+  int stride = rhs._grid->_slice_stride[dimension];
 PARALLEL_NESTED_LOOP2
   for(int n=0;n<e1;n++){
   for(int b=0;b<e2;b++){
 
-      int o  =n*rhs._grid->_slice_stride[dimension];
+      int o  =n*stride;
       int ocb=1<<lhs._grid->CheckerBoardFromOindex(o+b);
       if ( ocb&cbmask ) {
 	permute(lhs._odata[lo+o+b],rhs._odata[ro+o+b],permute_type);
@@ -296,6 +324,7 @@ template<class vobj> Lattice<vobj> Cshift_local(Lattice<vobj> &ret,const Lattice
   int rd = grid->_rdimensions[dimension];
   int ld = grid->_ldimensions[dimension];
   int gd = grid->_gdimensions[dimension];
+  int ly = grid->_simd_layout[dimension];
 
   // Map to always positive shift modulo global full dimension.
   shift = (shift+fd)%fd;
@@ -304,6 +333,7 @@ template<class vobj> Lattice<vobj> Cshift_local(Lattice<vobj> &ret,const Lattice
   // the permute type
   int permute_dim =grid->PermuteDim(dimension);
   int permute_type=grid->PermuteType(dimension);
+  int permute_type_dist;
 
   for(int x=0;x<rd;x++){       
 
@@ -315,15 +345,31 @@ template<class vobj> Lattice<vobj> Cshift_local(Lattice<vobj> &ret,const Lattice
     int sshift = grid->CheckerBoardShiftForCB(rhs.checkerboard,dimension,shift,cb);
     int sx     = (x+sshift)%rd;
 
+    // FIXME : This must change where we have a 
+    // Rotate slice.
+    
+    // Document how this works ; why didn't I do this when I first wrote it...
+    // wrap is whether sshift > rd.
+    //  num is sshift mod rd.
+    // 
     int permute_slice=0;
     if(permute_dim){
       int wrap = sshift/rd;
       int  num = sshift%rd;
+
       if ( x< rd-num ) permute_slice=wrap;
-      else permute_slice = 1-wrap;
+      else permute_slice = (wrap+1)%ly;
+
+      if ( (ly>2) && (permute_slice) ) {
+	assert(permute_type & RotateBit);
+	permute_type_dist = permute_type|permute_slice;
+      } else {
+	permute_type_dist = permute_type;
+      }
+      
     }
 
-    if ( permute_slice ) Copy_plane_permute(ret,rhs,dimension,x,sx,cbmask,permute_type);
+    if ( permute_slice ) Copy_plane_permute(ret,rhs,dimension,x,sx,cbmask,permute_type_dist);
     else                 Copy_plane(ret,rhs,dimension,x,sx,cbmask); 
 
   

lib/cshift/Cshift_mpi.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/cshift/Cshift_mpi.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef _GRID_CSHIFT_MPI_H_
 #define _GRID_CSHIFT_MPI_H_
 
@@ -163,11 +191,12 @@ template<class vobj> void  Cshift_comms_simd(Lattice<vobj> &ret,const Lattice<vo
   int buffer_size = grid->_slice_nblock[dimension]*grid->_slice_block[dimension];
   int words = sizeof(vobj)/sizeof(vector_type);
 
-  std::vector<std::vector<scalar_object> > send_buf_extract(Nsimd,std::vector<scalar_object>(buffer_size) );
-  std::vector<std::vector<scalar_object> > recv_buf_extract(Nsimd,std::vector<scalar_object>(buffer_size) );
+  std::vector<Vector<scalar_object> >   send_buf_extract(Nsimd,Vector<scalar_object>(buffer_size) );
+  std::vector<Vector<scalar_object> >   recv_buf_extract(Nsimd,Vector<scalar_object>(buffer_size) );
+
   int bytes = buffer_size*sizeof(scalar_object);
 
-  std::vector<scalar_object *>  pointers(Nsimd);  // 
+  std::vector<scalar_object *>  pointers(Nsimd); // 
   std::vector<scalar_object *> rpointers(Nsimd); // received pointers
 
   ///////////////////////////////////////////

lib/cshift/Cshift_none.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/cshift/Cshift_none.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef _GRID_CSHIFT_NONE_H_
 #define _GRID_CSHIFT_NONE_H_
 namespace Grid {

lib/lattice/Lattice_ET.h

@@ -1,3 +1,32 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_ET.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: neo <cossu@post.kek.jp>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_ET_H
 #define GRID_LATTICE_ET_H
 

lib/lattice/Lattice_arith.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_arith.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_ARITH_H
 #define GRID_LATTICE_ARITH_H
 

lib/lattice/Lattice_base.h

@@ -1,3 +1,32 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_base.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_BASE_H
 #define GRID_LATTICE_BASE_H
 
@@ -26,7 +55,13 @@ extern int GridCshiftPermuteMap[4][16];
 // Basic expressions used in Expression Template
 ////////////////////////////////////////////////
 
-class LatticeBase {};
+class LatticeBase
+{
+public:
+    virtual ~LatticeBase(void) = default;
+    GridBase *_grid;
+};
+    
 class LatticeExpressionBase {};
 
 template<class T> using Vector = std::vector<T,alignedAllocator<T> >;               // Aligned allocator??
@@ -59,8 +94,6 @@ template<class vobj>
 class Lattice : public LatticeBase
 {
 public:
-
-    GridBase *_grid;
     int checkerboard;
     Vector<vobj> _odata;
     
@@ -73,7 +106,6 @@ public:
     typedef typename vobj::scalar_type scalar_type;
     typedef typename vobj::vector_type vector_type;
     typedef vobj vector_object;
- 
    
   ////////////////////////////////////////////////////////////////////////////////
   // Expression Template closure support
@@ -149,8 +181,8 @@ PARALLEL_FOR_LOOP
   }
   //GridFromExpression is tricky to do
   template<class Op,class T1>
-    Lattice(const LatticeUnaryExpression<Op,T1> & expr):    _grid(nullptr){
-
+    Lattice(const LatticeUnaryExpression<Op,T1> & expr) {
+    _grid = nullptr;
     GridFromExpression(_grid,expr);
     assert(_grid!=nullptr);
 
@@ -171,7 +203,8 @@ PARALLEL_FOR_LOOP
     }
   };
   template<class Op,class T1, class T2>
-  Lattice(const LatticeBinaryExpression<Op,T1,T2> & expr):    _grid(nullptr){
+  Lattice(const LatticeBinaryExpression<Op,T1,T2> & expr) {
+    _grid = nullptr;
     GridFromExpression(_grid,expr);
     assert(_grid!=nullptr);
 
@@ -192,7 +225,8 @@ PARALLEL_FOR_LOOP
     }
   };
   template<class Op,class T1, class T2, class T3>
-  Lattice(const LatticeTrinaryExpression<Op,T1,T2,T3> & expr):    _grid(nullptr){
+  Lattice(const LatticeTrinaryExpression<Op,T1,T2,T3> & expr) {
+    _grid = nullptr;
     GridFromExpression(_grid,expr);
     assert(_grid!=nullptr);
 
@@ -212,14 +246,17 @@ PARALLEL_FOR_LOOP
     // Constructor requires "grid" passed.
     // what about a default grid?
     //////////////////////////////////////////////////////////////////
-    Lattice(GridBase *grid) : _grid(grid), _odata(_grid->oSites()) {
-      //        _odata.reserve(_grid->oSites());
-      //        _odata.resize(_grid->oSites());
+    Lattice(GridBase *grid) : _odata(grid->oSites()) {
+        _grid = grid;
+    //        _odata.reserve(_grid->oSites());
+    //        _odata.resize(_grid->oSites());
     //      std::cout << "Constructing lattice object with Grid pointer "<<_grid<<std::endl;
         assert((((uint64_t)&_odata[0])&0xF) ==0);
         checkerboard=0;
     }
 
+    virtual ~Lattice(void) = default;
+    
     template<class sobj> strong_inline Lattice<vobj> & operator = (const sobj & r){
 PARALLEL_FOR_LOOP
         for(int ss=0;ss<_grid->oSites();ss++){

lib/lattice/Lattice_comparison.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_comparison.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_COMPARISON_H
 #define GRID_LATTICE_COMPARISON_H
 

lib/lattice/Lattice_comparison_utils.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_comparison_utils.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_COMPARISON_H
 #define GRID_COMPARISON_H
 

lib/lattice/Lattice_conformable.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_conformable.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_CONFORMABLE_H
 #define GRID_LATTICE_CONFORMABLE_H
 

lib/lattice/Lattice_coordinate.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_coordinate.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_COORDINATE_H
 #define GRID_LATTICE_COORDINATE_H
 

lib/lattice/Lattice_local.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_local.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_LOCALREDUCTION_H
 #define GRID_LATTICE_LOCALREDUCTION_H
 

lib/lattice/Lattice_overload.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_overload.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_OVERLOAD_H
 #define GRID_LATTICE_OVERLOAD_H
 

lib/lattice/Lattice_peekpoke.h

@@ -1,3 +1,32 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_peekpoke.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: Peter Boyle <peterboyle@Peters-MacBook-Pro-2.local>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_PEEK_H
 #define GRID_LATTICE_PEEK_H
 
@@ -123,7 +152,7 @@ PARALLEL_FOR_LOOP
     // Peek a scalar object from the SIMD array
     //////////////////////////////////////////////////////////
     template<class vobj,class sobj>
-    void peekLocalSite(sobj &s,Lattice<vobj> &l,std::vector<int> &site){
+    void peekLocalSite(sobj &s,const Lattice<vobj> &l,std::vector<int> &site){
         
       GridBase *grid=l._grid;
 

lib/lattice/Lattice_reality.h

@@ -1,3 +1,32 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_reality.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: neo <cossu@post.kek.jp>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_REALITY_H
 #define GRID_LATTICE_REALITY_H
 

lib/lattice/Lattice_reduction.h

@@ -1,3 +1,32 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_reduction.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_REDUCTION_H
 #define GRID_LATTICE_REDUCTION_H
 
@@ -19,7 +48,6 @@ namespace Grid {
     {
       typedef typename vobj::scalar_type scalar_type;
       typedef typename vobj::vector_type vector_type;
-      vector_type vnrm;
       scalar_type  nrm;
 
       GridBase *grid = left._grid;
@@ -31,7 +59,6 @@ namespace Grid {
 
 PARALLEL_FOR_LOOP
       for(int thr=0;thr<grid->SumArraySize();thr++){
-
 	int nwork, mywork, myoff;
 	GridThread::GetWork(left._grid->oSites(),thr,mywork,myoff);
 	
@@ -125,7 +152,7 @@ template<class vobj> inline void sliceSum(const Lattice<vobj> &Data,std::vector<
   assert(grid!=NULL);
 
   // FIXME
-  std::cout<<GridLogMessage<<"WARNING ! SliceSum is unthreaded "<<grid->SumArraySize()<<" threads "<<std::endl;
+  // std::cout<<GridLogMessage<<"WARNING ! SliceSum is unthreaded "<<grid->SumArraySize()<<" threads "<<std::endl;
 
   const int    Nd = grid->_ndimension;
   const int Nsimd = grid->Nsimd();
@@ -151,7 +178,7 @@ template<class vobj> inline void sliceSum(const Lattice<vobj> &Data,std::vector<
   // sum over reduced dimension planes, breaking out orthog dir
 
   for(int ss=0;ss<grid->oSites();ss++){
-    GridBase::CoorFromIndex(coor,ss,grid->_rdimensions);
+    Lexicographic::CoorFromIndex(coor,ss,grid->_rdimensions);
     int r = coor[orthogdim];
     lvSum[r]=lvSum[r]+Data._odata[ss];
   }  

lib/lattice/Lattice_rng.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_rng.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_RNG_H
 #define GRID_LATTICE_RNG_H
 
@@ -22,13 +50,16 @@ namespace Grid {
       assert(fine->_processors[d]==1);
     }
 
-    // local and global volumes subdivide cleanly after SIMDization
     int multiplicity=1;
+    for(int d=0;d<lowerdims;d++){
+      multiplicity=multiplicity*fine->_rdimensions[d];
+    }
+    // local and global volumes subdivide cleanly after SIMDization
     for(int d=0;d<rngdims;d++){
       int fd= d+lowerdims;
       assert(coarse->_processors[d]  == fine->_processors[fd]);
       assert(coarse->_simd_layout[d] == fine->_simd_layout[fd]);
-      assert((fine->_rdimensions[fd] / coarse->_rdimensions[d])* coarse->_rdimensions[d]==fine->_rdimensions[fd]); 
+      assert(((fine->_rdimensions[fd] / coarse->_rdimensions[d])* coarse->_rdimensions[d])==fine->_rdimensions[fd]); 
 
       multiplicity = multiplicity *fine->_rdimensions[fd] / coarse->_rdimensions[d]; 
     }
@@ -44,7 +75,7 @@ namespace Grid {
 
     std::seed_seq src;
     
-    fixedSeed(std::vector<int> &seeds) : src(seeds.begin(),seeds.end()) {};
+    fixedSeed(const std::vector<int> &seeds) : src(seeds.begin(),seeds.end()) {};
 
     result_type operator () (void){
 
@@ -76,16 +107,41 @@ namespace Grid {
 
   public:
 
-   GridRNGbase() : _uniform{0,1}, _gaussian(0.0,1.0) {};
-
     int _seeded;
     // One generator per site.
     // Uniform and Gaussian distributions from these generators.
-    std::vector<std::ranlux48>             _generators;
-    std::uniform_real_distribution<double> _uniform;
-    std::normal_distribution<double>       _gaussian;
-
+#ifdef RNG_RANLUX
+    typedef uint64_t      RngStateType;
+    typedef std::ranlux48 RngEngine;
+    static const int RngStateCount = 15;
+#else
+    typedef std::mt19937 RngEngine;
+    typedef uint32_t     RngStateType;
+    static const int     RngStateCount = std::mt19937::state_size;
+#endif
+    std::vector<RngEngine>                             _generators;
+    std::vector<std::uniform_real_distribution<RealD>> _uniform;
+    std::vector<std::normal_distribution<RealD>>       _gaussian;
+    std::vector<std::discrete_distribution<int32_t>>     _bernoulli;
 
+    void GetState(std::vector<RngStateType> & saved,int gen) {
+      saved.resize(RngStateCount);
+      std::stringstream ss;
+      ss<<_generators[gen];
+      ss.seekg(0,ss.beg);
+      for(int i=0;i<RngStateCount;i++){
+	ss>>saved[i];
+      }
+    }
+    void SetState(std::vector<RngStateType> & saved,int gen){
+      assert(saved.size()==RngStateCount);
+      std::stringstream ss;
+      for(int i=0;i<RngStateCount;i++){
+	ss<< saved[i]<<" ";
+      }
+      ss.seekg(0,ss.beg);
+      ss>>_generators[gen];
+    }
   };
 
   class GridSerialRNG : public GridRNGbase {
@@ -98,18 +154,21 @@ namespace Grid {
     {
       typename source::result_type init = src();
       CartesianCommunicator::BroadcastWorld(0,(void *)&init,sizeof(init));
-      _generators[0] = std::ranlux48(init);
+      _generators[0] = RngEngine(init);
       _seeded=1;
     }    
 
     GridSerialRNG() : GridRNGbase() {
       _generators.resize(1);
+      _uniform.resize(1,std::uniform_real_distribution<RealD>{0,1});
+      _gaussian.resize(1,std::normal_distribution<RealD>(0.0,1.0) );
+      _bernoulli.resize(1,std::discrete_distribution<int32_t>{1,1});
       _seeded=0;
     }
 
 
 
-    template <class sobj,class distribution> inline void fill(sobj &l,distribution &dist){
+    template <class sobj,class distribution> inline void fill(sobj &l,std::vector<distribution> &dist){
 
       typedef typename sobj::scalar_type scalar_type;
  
@@ -117,56 +176,65 @@ namespace Grid {
 
       scalar_type *buf = (scalar_type *) & l;
 
+      dist[0].reset();
       for(int idx=0;idx<words;idx++){
-	fillScalar(buf[idx],dist,_generators[0]);
+	fillScalar(buf[idx],dist[0],_generators[0]);
       }
       
       CartesianCommunicator::BroadcastWorld(0,(void *)&l,sizeof(l));
 
     };
 
-    template <class distribution>  inline void fill(ComplexF &l,distribution &dist){
-      fillScalar(l,dist,_generators[0]);
+    template <class distribution>  inline void fill(ComplexF &l,std::vector<distribution> &dist){
+      dist[0].reset();
+      fillScalar(l,dist[0],_generators[0]);
       CartesianCommunicator::BroadcastWorld(0,(void *)&l,sizeof(l));
     }
-    template <class distribution>  inline void fill(ComplexD &l,distribution &dist){
-      fillScalar(l,dist,_generators[0]);
+    template <class distribution>  inline void fill(ComplexD &l,std::vector<distribution> &dist){
+      dist[0].reset();
+      fillScalar(l,dist[0],_generators[0]);
       CartesianCommunicator::BroadcastWorld(0,(void *)&l,sizeof(l));
     }
-    template <class distribution>  inline void fill(RealF &l,distribution &dist){
-      fillScalar(l,dist,_generators[0]);
+    template <class distribution>  inline void fill(RealF &l,std::vector<distribution> &dist){
+      dist[0].reset();
+      fillScalar(l,dist[0],_generators[0]);
       CartesianCommunicator::BroadcastWorld(0,(void *)&l,sizeof(l));
     }
-    template <class distribution>  inline void fill(RealD &l,distribution &dist){
-      fillScalar(l,dist,_generators[0]);
+    template <class distribution>  inline void fill(RealD &l,std::vector<distribution> &dist){
+      dist[0].reset();
+      fillScalar(l,dist[0],_generators[0]);
       CartesianCommunicator::BroadcastWorld(0,(void *)&l,sizeof(l));
     }
     // vector fill
-    template <class distribution>  inline void fill(vComplexF &l,distribution &dist){
+    template <class distribution>  inline void fill(vComplexF &l,std::vector<distribution> &dist){
       RealF *pointer=(RealF *)&l;
+      dist[0].reset();
       for(int i=0;i<2*vComplexF::Nsimd();i++){
-	fillScalar(pointer[i],dist,_generators[0]);
+	fillScalar(pointer[i],dist[0],_generators[0]);
       }
       CartesianCommunicator::BroadcastWorld(0,(void *)&l,sizeof(l));
     }
-    template <class distribution>  inline void fill(vComplexD &l,distribution &dist){
+    template <class distribution>  inline void fill(vComplexD &l,std::vector<distribution> &dist){
       RealD *pointer=(RealD *)&l;
+      dist[0].reset();
       for(int i=0;i<2*vComplexD::Nsimd();i++){
-	fillScalar(pointer[i],dist,_generators[0]);
+	fillScalar(pointer[i],dist[0],_generators[0]);
       }
       CartesianCommunicator::BroadcastWorld(0,(void *)&l,sizeof(l));
     }
-    template <class distribution>  inline void fill(vRealF &l,distribution &dist){
+    template <class distribution>  inline void fill(vRealF &l,std::vector<distribution> &dist){
       RealF *pointer=(RealF *)&l;
+      dist[0].reset();
       for(int i=0;i<vRealF::Nsimd();i++){
-	fillScalar(pointer[i],dist,_generators[0]);
+	fillScalar(pointer[i],dist[0],_generators[0]);
       }
       CartesianCommunicator::BroadcastWorld(0,(void *)&l,sizeof(l));
     }
-    template <class distribution>  inline void fill(vRealD &l,distribution &dist){
+    template <class distribution>  inline void fill(vRealD &l,std::vector<distribution> &dist){
       RealD *pointer=(RealD *)&l;
+      dist[0].reset();
       for(int i=0;i<vRealD::Nsimd();i++){
-	fillScalar(pointer[i],dist,_generators[0]);
+	fillScalar(pointer[i],dist[0],_generators[0]);
       }
       CartesianCommunicator::BroadcastWorld(0,(void *)&l,sizeof(l));
     }
@@ -176,7 +244,7 @@ namespace Grid {
       std::random_device rd;
       Seed(rd);
     }
-    void SeedFixedIntegers(std::vector<int> &seeds){
+    void SeedFixedIntegers(const std::vector<int> &seeds){
       fixedSeed src(seeds);
       Seed(src);
     }
@@ -185,12 +253,6 @@ namespace Grid {
 
   class GridParallelRNG : public GridRNGbase {
   public:
-    // One generator per site.
-    std::vector<std::ranlux48>             _generators;
-    
-    // Uniform and Gaussian distributions from these generators.
-    std::uniform_real_distribution<double> _uniform;
-    std::normal_distribution<double>       _gaussian;
 
     GridBase *_grid;
     int _vol;
@@ -202,7 +264,11 @@ namespace Grid {
     GridParallelRNG(GridBase *grid) : GridRNGbase() {
       _grid=grid;
       _vol =_grid->iSites()*_grid->oSites();
+
       _generators.resize(_vol);
+      _uniform.resize(_vol,std::uniform_real_distribution<RealD>{0,1});
+      _gaussian.resize(_vol,std::normal_distribution<RealD>(0.0,1.0) );
+      _bernoulli.resize(_vol,std::discrete_distribution<int32_t>{1,1});
       _seeded=0;
     }
 
@@ -220,7 +286,7 @@ namespace Grid {
       int gsites = _grid->_gsites;
 
       typename source::result_type init = src();
-      std::ranlux48 pseeder(init);
+      RngEngine pseeder(init);
       std::uniform_int_distribution<uint64_t> ui;
 
       for(int gidx=0;gidx<gsites;gidx++){
@@ -241,7 +307,7 @@ namespace Grid {
 	if( rank == _grid->ThisRank() ){
 	  fixedSeed ssrc(site_seeds);
 	  typename source::result_type sinit = ssrc();
-	  _generators[l_idx] = std::ranlux48(sinit);
+	  _generators[l_idx] = RngEngine(sinit);
 	}
       }
       _seeded=1;
@@ -251,7 +317,7 @@ namespace Grid {
     //void SaveState(const std::string<char> &file);
     //void LoadState(const std::string<char> &file);
 
-    template <class vobj,class distribution> inline void fill(Lattice<vobj> &l,distribution &dist){
+    template <class vobj,class distribution> inline void fill(Lattice<vobj> &l,std::vector<distribution> &dist){
 
       typedef typename vobj::scalar_object scalar_object;
       typedef typename vobj::scalar_type scalar_type;
@@ -275,8 +341,9 @@ PARALLEL_FOR_LOOP
 	  for(int si=0;si<Nsimd;si++){
 	    int gdx = generator_idx(ss,si); // index of generator state
 	    scalar_type *pointer = (scalar_type *)&buf[si];
+	    dist[gdx].reset();
 	    for(int idx=0;idx<words;idx++){
-	      fillScalar(pointer[idx],dist,_generators[gdx]);
+	      fillScalar(pointer[idx],dist[gdx],_generators[gdx]);
 	    }
 	  }
 
@@ -290,7 +357,7 @@ PARALLEL_FOR_LOOP
       std::random_device rd;
       Seed(rd);
     }
-    void SeedFixedIntegers(std::vector<int> &seeds){
+    void SeedFixedIntegers(const std::vector<int> &seeds){
       fixedSeed src(seeds);
       Seed(src);
     }
@@ -304,14 +371,22 @@ PARALLEL_FOR_LOOP
   template <class vobj> inline void gaussian(GridParallelRNG &rng,Lattice<vobj> &l){
     rng.fill(l,rng._gaussian);
   }
-
+  
+  template <class vobj> inline void bernoulli(GridParallelRNG &rng,Lattice<vobj> &l){
+    rng.fill(l,rng._bernoulli);
+  }
 
   template <class sobj> inline void random(GridSerialRNG &rng,sobj &l){
     rng.fill(l,rng._uniform);
   }
+  
   template <class sobj> inline void gaussian(GridSerialRNG &rng,sobj &l){
     rng.fill(l,rng._gaussian);
   }
+  
+  template <class sobj> inline void bernoulli(GridSerialRNG &rng,sobj &l){
+    rng.fill(l,rng._bernoulli);
+  }
 
 }
 #endif

lib/lattice/Lattice_trace.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_trace.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_TRACE_H
 #define GRID_LATTICE_TRACE_H
 

lib/lattice/Lattice_transfer.h

@@ -1,3 +1,30 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_transfer.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_TRANSFER_H
 #define GRID_LATTICE_TRANSFER_H
 
@@ -17,7 +44,7 @@ inline void subdivides(GridBase *coarse,GridBase *fine)
   }
 }
 
-
+ 
   ////////////////////////////////////////////////////////////////////////////////////////////
   // remove and insert a half checkerboard
   ////////////////////////////////////////////////////////////////////////////////////////////
@@ -88,9 +115,9 @@ inline void blockProject(Lattice<iVector<CComplex,nbasis > > &coarseData,
     int sc;
     std::vector<int> coor_c(_ndimension);
     std::vector<int> coor_f(_ndimension);
-    GridBase::CoorFromIndex(coor_f,sf,fine->_rdimensions);
+    Lexicographic::CoorFromIndex(coor_f,sf,fine->_rdimensions);
     for(int d=0;d<_ndimension;d++) coor_c[d]=coor_f[d]/block_r[d];
-    GridBase::IndexFromCoor(coor_c,sc,coarse->_rdimensions);
+    Lexicographic::IndexFromCoor(coor_c,sc,coarse->_rdimensions);
 
     for(int i=0;i<nbasis;i++) {
       
@@ -133,9 +160,9 @@ PARALLEL_FOR_LOOP
     std::vector<int> coor_c(_ndimension);
     std::vector<int> coor_f(_ndimension);
 
-    GridBase::CoorFromIndex(coor_f,sf,fine->_rdimensions);
+    Lexicographic::CoorFromIndex(coor_f,sf,fine->_rdimensions);
     for(int d=0;d<_ndimension;d++) coor_c[d]=coor_f[d]/block_r[d];
-    GridBase::IndexFromCoor(coor_c,sc,coarse->_rdimensions);
+    Lexicographic::IndexFromCoor(coor_c,sc,coarse->_rdimensions);
 
     // z = A x + y
     fineZ._odata[sf]=coarseA._odata[sc]*fineX._odata[sf]+fineY._odata[sf];
@@ -198,9 +225,9 @@ inline void blockSum(Lattice<vobj> &coarseData,const Lattice<vobj> &fineData)
     std::vector<int> coor_c(_ndimension);
     std::vector<int> coor_f(_ndimension);
 
-    GridBase::CoorFromIndex(coor_f,sf,fine->_rdimensions);
+    Lexicographic::CoorFromIndex(coor_f,sf,fine->_rdimensions);
     for(int d=0;d<_ndimension;d++) coor_c[d]=coor_f[d]/block_r[d];
-    GridBase::IndexFromCoor(coor_c,sc,coarse->_rdimensions);
+    Lexicographic::IndexFromCoor(coor_c,sc,coarse->_rdimensions);
 
     coarseData._odata[sc]=coarseData._odata[sc]+fineData._odata[sf];
 
@@ -284,9 +311,9 @@ inline void blockPromote(const Lattice<iVector<CComplex,nbasis > > &coarseData,
     std::vector<int> coor_c(_ndimension);
     std::vector<int> coor_f(_ndimension);
 
-    GridBase::CoorFromIndex(coor_f,sf,fine->_rdimensions);
+    Lexicographic::CoorFromIndex(coor_f,sf,fine->_rdimensions);
     for(int d=0;d<_ndimension;d++) coor_c[d]=coor_f[d]/block_r[d];
-    GridBase::IndexFromCoor(coor_c,sc,coarse->_rdimensions);
+    Lexicographic::IndexFromCoor(coor_c,sc,coarse->_rdimensions);
 
     for(int i=0;i<nbasis;i++) {
       if(i==0) fineData._odata[sf]=coarseData._odata[sc](i) * Basis[i]._odata[sf];
@@ -298,6 +325,126 @@ inline void blockPromote(const Lattice<iVector<CComplex,nbasis > > &coarseData,
   
 }
 
+// Useful for precision conversion, or indeed anything where an operator= does a conversion on scalars.
+// Simd layouts need not match since we use peek/poke Local
+template<class vobj,class vvobj>
+void localConvert(const Lattice<vobj> &in,Lattice<vvobj> &out)
+{
+  typedef typename vobj::scalar_object sobj;
+  typedef typename vvobj::scalar_object ssobj;
+
+  sobj s;
+  ssobj ss;
+
+  GridBase *ig = in._grid;
+  GridBase *og = out._grid;
+
+  int ni = ig->_ndimension;
+  int no = og->_ndimension;
+
+  assert(ni == no);
+
+  for(int d=0;d<no;d++){
+    assert(ig->_processors[d]  == og->_processors[d]);
+    assert(ig->_ldimensions[d] == og->_ldimensions[d]);
+  }
+
+PARALLEL_FOR_LOOP
+  for(int idx=0;idx<ig->lSites();idx++){
+    std::vector<int> lcoor(ni);
+    ig->LocalIndexToLocalCoor(idx,lcoor);
+    peekLocalSite(s,in,lcoor);
+    ss=s;
+    pokeLocalSite(ss,out,lcoor);
+  }
+}
+
+
+template<class vobj>
+void InsertSlice(Lattice<vobj> &lowDim,Lattice<vobj> & higherDim,int slice, int orthog)
+{
+  typedef typename vobj::scalar_object sobj;
+  sobj s;
+
+  GridBase *lg = lowDim._grid;
+  GridBase *hg = higherDim._grid;
+  int nl = lg->_ndimension;
+  int nh = hg->_ndimension;
+
+  assert(nl+1 == nh);
+  assert(orthog<nh);
+  assert(orthog>=0);
+  assert(hg->_processors[orthog]==1);
+
+  int dl; dl = 0;
+  for(int d=0;d<nh;d++){
+    if ( d != orthog) {
+      assert(lg->_processors[dl]  == hg->_processors[d]);
+      assert(lg->_ldimensions[dl] == hg->_ldimensions[d]);
+      dl++;
+    }
+  }
+
+  // the above should guarantee that the operations are local
+PARALLEL_FOR_LOOP
+  for(int idx=0;idx<lg->lSites();idx++){
+    std::vector<int> lcoor(nl);
+    std::vector<int> hcoor(nh);
+    lg->LocalIndexToLocalCoor(idx,lcoor);
+    dl=0;
+    hcoor[orthog] = slice;
+    for(int d=0;d<nh;d++){
+      if ( d!=orthog ) { 
+	hcoor[d]=lcoor[dl++];
+      }
+    }
+    peekLocalSite(s,lowDim,lcoor);
+    pokeLocalSite(s,higherDim,hcoor);
+  }
+}
+
+template<class vobj>
+void ExtractSlice(Lattice<vobj> &lowDim, Lattice<vobj> & higherDim,int slice, int orthog)
+{
+  typedef typename vobj::scalar_object sobj;
+  sobj s;
+
+  GridBase *lg = lowDim._grid;
+  GridBase *hg = higherDim._grid;
+  int nl = lg->_ndimension;
+  int nh = hg->_ndimension;
+
+  assert(nl+1 == nh);
+  assert(orthog<nh);
+  assert(orthog>=0);
+  assert(hg->_processors[orthog]==1);
+
+  int dl; dl = 0;
+  for(int d=0;d<nh;d++){
+    if ( d != orthog) {
+      assert(lg->_processors[dl]  == hg->_processors[d]);
+      assert(lg->_ldimensions[dl] == hg->_ldimensions[d]);
+      dl++;
+    }
+  }
+  // the above should guarantee that the operations are local
+PARALLEL_FOR_LOOP
+  for(int idx=0;idx<lg->lSites();idx++){
+    std::vector<int> lcoor(nl);
+    std::vector<int> hcoor(nh);
+    lg->LocalIndexToLocalCoor(idx,lcoor);
+    dl=0;
+    hcoor[orthog] = slice;
+    for(int d=0;d<nh;d++){
+      if ( d!=orthog ) { 
+	hcoor[d]=lcoor[dl++];
+      }
+    }
+    peekLocalSite(s,higherDim,hcoor);
+    pokeLocalSite(s,lowDim,lcoor);
+  }
+
+}
 
 template<class vobj>
 void Replicate(Lattice<vobj> &coarse,Lattice<vobj> & fine)

lib/lattice/Lattice_transpose.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_transpose.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_TRANSPOSE_H
 #define GRID_LATTICE_TRANSPOSE_H
 

lib/lattice/Lattice_unary.h

@@ -1,3 +1,33 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_unary.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: neo <cossu@post.kek.jp>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_UNARY_H
 #define GRID_LATTICE_UNARY_H
 

lib/lattice/Lattice_where.h

@@ -1,3 +1,32 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/lattice/Lattice_where.h
+
+    Copyright (C) 2015
+
+Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: Peter Boyle <peterboyle@Peters-MacBook-Pro-2.local>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_LATTICE_WHERE_H
 #define GRID_LATTICE_WHERE_H
 namespace Grid {

lib/parallelIO/BinaryIO.h

@@ -1,3 +1,31 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/parallelIO/BinaryIO.h
+
+    Copyright (C) 2015
+
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_BINARY_IO_H
 #define GRID_BINARY_IO_H
 
@@ -107,7 +135,7 @@ class BinaryIO {
     }
   }
 
-  template<class vobj,class fobj,class munger> static inline void Uint32Checksum(Lattice<vobj> lat,munger munge,uint32_t &csum)
+  template<class vobj,class fobj,class munger> static inline void Uint32Checksum(Lattice<vobj> &lat,munger munge,uint32_t &csum)
   {
     typedef typename vobj::scalar_object sobj;
     GridBase *grid = lat._grid ;
@@ -118,7 +146,7 @@ class BinaryIO {
     csum = 0;
     std::vector<int> lcoor;
     for(int l=0;l<grid->lSites();l++){
-      grid->CoorFromIndex(lcoor,l,grid->_ldimensions);
+      Lexicographic::CoorFromIndex(lcoor,l,grid->_ldimensions);
       peekLocalSite(siteObj,lat,lcoor);
       munge(siteObj,fileObj,csum);
     }
@@ -140,6 +168,7 @@ class BinaryIO {
     GridBase *grid = Umu._grid;
 
     std::cout<< GridLogMessage<< "Serial read I/O "<< file<< std::endl;
+    GridStopWatch timer; timer.Start();
 
     int ieee32big = (format == std::string("IEEE32BIG"));
     int ieee32    = (format == std::string("IEEE32"));
@@ -154,6 +183,7 @@ class BinaryIO {
 
     Umu = zero;
     uint32_t csum=0;
+    uint64_t bytes=0;
     fobj file_object;
     sobj munged;
     
@@ -166,7 +196,7 @@ class BinaryIO {
 
       if ( grid->IsBoss() ) {
 	fin.read((char *)&file_object,sizeof(file_object));
-	
+	bytes += sizeof(file_object);
 	if(ieee32big) be32toh_v((void *)&file_object,sizeof(file_object));
 	if(ieee32)    le32toh_v((void *)&file_object,sizeof(file_object));
 	if(ieee64big) be64toh_v((void *)&file_object,sizeof(file_object));
@@ -177,6 +207,10 @@ class BinaryIO {
       // The boss who read the file has their value poked
       pokeSite(munged,Umu,site);
     }}}}
+    timer.Stop();
+    std::cout<<GridLogPerformance<<"readObjectSerial: read "<< bytes <<" bytes in "<<timer.Elapsed() <<" "
+	     << (double)bytes/ (double)timer.useconds() <<" MB/s "  <<std::endl;
+
     return csum;
   }
 
@@ -196,13 +230,14 @@ class BinaryIO {
     // Serialise through node zero
     //////////////////////////////////////////////////
     std::cout<< GridLogMessage<< "Serial write I/O "<< file<<std::endl;
+    GridStopWatch timer; timer.Start();
 
     std::ofstream fout;
     if ( grid->IsBoss() ) {
       fout.open(file,std::ios::binary|std::ios::out|std::ios::in);
       fout.seekp(offset);
     }
-    
+    uint64_t bytes=0;
     uint32_t csum=0;
     fobj file_object;
     sobj unmunged;
@@ -224,14 +259,126 @@ class BinaryIO {
 	if(ieee32)    htole32_v((void *)&file_object,sizeof(file_object));
 	if(ieee64big) htobe64_v((void *)&file_object,sizeof(file_object));
 	if(ieee64)    htole64_v((void *)&file_object,sizeof(file_object));
-	
+
+	// NB could gather an xstrip as an optimisation.
 	fout.write((char *)&file_object,sizeof(file_object));
+	bytes+=sizeof(file_object);
       }
     }}}}
+    timer.Stop();
+    std::cout<<GridLogPerformance<<"writeObjectSerial: wrote "<< bytes <<" bytes in "<<timer.Elapsed() <<" "
+	     << (double)bytes/timer.useconds() <<" MB/s "  <<std::endl;
 
     return csum;
   }
 
+  static inline uint32_t writeRNGSerial(GridSerialRNG &serial,GridParallelRNG &parallel,std::string file,int offset)
+  {
+    typedef typename GridSerialRNG::RngStateType RngStateType;
+    const int RngStateCount = GridSerialRNG::RngStateCount;
+
+
+    GridBase *grid = parallel._grid;
+    int gsites = grid->_gsites;
+
+    //////////////////////////////////////////////////
+    // Serialise through node zero
+    //////////////////////////////////////////////////
+    std::cout<< GridLogMessage<< "Serial RNG write I/O "<< file<<std::endl;
+
+    std::ofstream fout;
+    if ( grid->IsBoss() ) {
+      fout.open(file,std::ios::binary|std::ios::out|std::ios::in);
+      fout.seekp(offset);
+    }
+    
+    uint32_t csum=0;
+    std::vector<RngStateType> saved(RngStateCount);
+    int bytes = sizeof(RngStateType)*saved.size();
+    std::vector<int> gcoor;
+
+    for(int gidx=0;gidx<gsites;gidx++){
+
+      int rank,o_idx,i_idx;
+      grid->GlobalIndexToGlobalCoor(gidx,gcoor);
+      grid->GlobalCoorToRankIndex(rank,o_idx,i_idx,gcoor);
+      int l_idx=parallel.generator_idx(o_idx,i_idx);
+
+      if( rank == grid->ThisRank() ){
+	//	std::cout << "rank" << rank<<" Getting state for index "<<l_idx<<std::endl;
+	parallel.GetState(saved,l_idx);
+      }
+
+      grid->Broadcast(rank,(void *)&saved[0],bytes);
+
+      if ( grid->IsBoss() ) {
+	Uint32Checksum((uint32_t *)&saved[0],bytes,csum);
+	fout.write((char *)&saved[0],bytes);
+      }
+
+    }
+
+    if ( grid->IsBoss() ) {
+      serial.GetState(saved,0);
+      Uint32Checksum((uint32_t *)&saved[0],bytes,csum);
+      fout.write((char *)&saved[0],bytes);
+    }
+    grid->Broadcast(0,(void *)&csum,sizeof(csum));
+    return csum;
+  }
+  static inline uint32_t readRNGSerial(GridSerialRNG &serial,GridParallelRNG &parallel,std::string file,int offset)
+  {
+    typedef typename GridSerialRNG::RngStateType RngStateType;
+    const int RngStateCount = GridSerialRNG::RngStateCount;
+
+    GridBase *grid = parallel._grid;
+    int gsites = grid->_gsites;
+
+    //////////////////////////////////////////////////
+    // Serialise through node zero
+    //////////////////////////////////////////////////
+    std::cout<< GridLogMessage<< "Serial RNG read I/O "<< file<<std::endl;
+
+    std::ifstream fin(file,std::ios::binary|std::ios::in);
+    fin.seekg(offset);
+    
+    uint32_t csum=0;
+    std::vector<RngStateType> saved(RngStateCount);
+    int bytes = sizeof(RngStateType)*saved.size();
+    std::vector<int> gcoor;
+
+    for(int gidx=0;gidx<gsites;gidx++){
+
+      int rank,o_idx,i_idx;
+      grid->GlobalIndexToGlobalCoor(gidx,gcoor);
+      grid->GlobalCoorToRankIndex(rank,o_idx,i_idx,gcoor);
+      int l_idx=parallel.generator_idx(o_idx,i_idx);
+
+      if ( grid->IsBoss() ) {
+	fin.read((char *)&saved[0],bytes);
+	Uint32Checksum((uint32_t *)&saved[0],bytes,csum);
+      }
+
+      grid->Broadcast(0,(void *)&saved[0],bytes);
+
+      if( rank == grid->ThisRank() ){
+	parallel.SetState(saved,l_idx);
+      }
+
+    }
+
+    if ( grid->IsBoss() ) {
+      fin.read((char *)&saved[0],bytes);
+      serial.SetState(saved,0);
+      Uint32Checksum((uint32_t *)&saved[0],bytes,csum);
+    }
+
+    grid->Broadcast(0,(void *)&csum,sizeof(csum));
+
+    return csum;
+  }
+
+
   template<class vobj,class fobj,class munger>
   static inline uint32_t readObjectParallel(Lattice<vobj> &Umu,std::string file,munger munge,int offset,const std::string &format)
   {
@@ -266,7 +413,7 @@ class BinaryIO {
     int IOnode = 1;
     for(int d=0;d<grid->_ndimension;d++) {
 
-      if ( d==0 ) parallel[d] = 0;
+      if ( d == 0 ) parallel[d] = 0;
       if (parallel[d]) {
 	range[d] = grid->_ldimensions[d];
 	start[d] = grid->_processor_coor[d]*range[d];
@@ -284,6 +431,7 @@ class BinaryIO {
     {
       uint32_t tmp = IOnode;
       grid->GlobalSum(tmp);
+      std::cout<< std::dec ;
       std::cout<< GridLogMessage<< "Parallel read I/O to "<< file << " with " <<tmp<< " IOnodes for subslice ";
       for(int d=0;d<grid->_ndimension;d++){
 	std::cout<< range[d];
@@ -293,6 +441,9 @@ class BinaryIO {
       std::cout << std::endl;
     }
 
+    GridStopWatch timer; timer.Start();
+    uint64_t bytes=0;
+
     int myrank = grid->ThisRank();
     int iorank = grid->RankFromProcessorCoor(ioproc);
 
@@ -306,9 +457,9 @@ class BinaryIO {
     // available (how short sighted is that?)
     //////////////////////////////////////////////////////////
     Umu = zero;
-    uint32_t csum=0;
+    static uint32_t csum=0;
     fobj fileObj;
-    sobj siteObj;
+    static sobj siteObj; // Static to place in symmetric region for SHMEM
 
       // need to implement these loops in Nd independent way with a lexico conversion
     for(int tlex=0;tlex<slice_vol;tlex++){
@@ -318,7 +469,7 @@ class BinaryIO {
       std::vector<int> lsite(nd);
       std::vector<int> iosite(nd);
 
-      grid->CoorFromIndex(tsite,tlex,range);
+      Lexicographic::CoorFromIndex(tsite,tlex,range);
 
       for(int d=0;d<nd;d++){
 	lsite[d] = tsite[d]%grid->_ldimensions[d];  // local site
@@ -328,7 +479,7 @@ class BinaryIO {
       /////////////////////////
       // Get the rank of owner of data
       /////////////////////////
-	int rank, o_idx,i_idx, g_idx;
+      int rank, o_idx,i_idx, g_idx;
       grid->GlobalCoorToRankIndex(rank,o_idx,i_idx,gsite);
       grid->GlobalCoorToGlobalIndex(gsite,g_idx);
       
@@ -339,6 +490,7 @@ class BinaryIO {
 	
 	fin.seekg(offset+g_idx*sizeof(fileObj));
 	fin.read((char *)&fileObj,sizeof(fileObj));
+	bytes+=sizeof(fileObj);
 	
 	if(ieee32big) be32toh_v((void *)&fileObj,sizeof(fileObj));
 	if(ieee32)    le32toh_v((void *)&fileObj,sizeof(fileObj));
@@ -346,23 +498,29 @@ class BinaryIO {
 	if(ieee64)    le64toh_v((void *)&fileObj,sizeof(fileObj));
 	
 	munge(fileObj,siteObj,csum);
-	
-	if ( rank != myrank ) {
-	  grid->SendTo((void *)&siteObj,rank,sizeof(siteObj));
-	} else { 
-	  pokeLocalSite(siteObj,Umu,lsite);
+
+      }	
+
+      // Possibly do transport through pt2pt 
+      if ( rank != iorank ) { 
+	if ( (myrank == rank) || (myrank==iorank) ) {
+	  grid->SendRecvPacket((void *)&siteObj,(void *)&siteObj,iorank,rank,sizeof(siteObj));
 	}
-	 
-      } else { 
-	if ( myrank == rank ) {
-	  grid->RecvFrom((void *)&siteObj,iorank,sizeof(siteObj));
+      }
+      // Poke at destination
+      if ( myrank == rank ) {
 	  pokeLocalSite(siteObj,Umu,lsite);
-	} 
       }
       grid->Barrier(); // necessary?
     }
 
     grid->GlobalSum(csum);
+    grid->GlobalSum(bytes);
+    grid->Barrier();
+
+    timer.Stop();
+    std::cout<<GridLogPerformance<<"readObjectParallel: read "<< bytes <<" bytes in "<<timer.Elapsed() <<" "
+	     << (double)bytes/timer.useconds() <<" MB/s "  <<std::endl;
     
     return csum;
   }
@@ -397,7 +555,7 @@ class BinaryIO {
 
     for(int d=0;d<grid->_ndimension;d++) {
 
-      if ( d==0 ) parallel[d] = 0;
+      if ( d!= grid->_ndimension-1 ) parallel[d] = 0;
 
       if (parallel[d]) {
 	range[d] = grid->_ldimensions[d];
@@ -426,6 +584,9 @@ class BinaryIO {
       std::cout << std::endl;
     }
 
+    GridStopWatch timer; timer.Start();
+    uint64_t bytes=0;
+
     int myrank = grid->ThisRank();
     int iorank = grid->RankFromProcessorCoor(ioproc);
 
@@ -444,10 +605,10 @@ class BinaryIO {
 
     uint32_t csum=0;
     fobj fileObj;
-    sobj siteObj;
+    static sobj siteObj; // static for SHMEM target; otherwise dynamic allocate with AlignedAllocator
 
-
-      // need to implement these loops in Nd independent way with a lexico conversion
+    // should aggregate a whole chunk and then write.
+    // need to implement these loops in Nd independent way with a lexico conversion
     for(int tlex=0;tlex<slice_vol;tlex++){
 	
       std::vector<int> tsite(nd); // temporary mixed up site
@@ -455,7 +616,7 @@ class BinaryIO {
       std::vector<int> lsite(nd);
       std::vector<int> iosite(nd);
 
-      grid->CoorFromIndex(tsite,tlex,range);
+      Lexicographic::CoorFromIndex(tsite,tlex,range);
 
       for(int d=0;d<nd;d++){
 	lsite[d] = tsite[d]%grid->_ldimensions[d];  // local site
@@ -473,13 +634,21 @@ class BinaryIO {
       ////////////////////////////////
       // iorank writes from the seek
       ////////////////////////////////
-      if (myrank == iorank) {
+      
+      // Owner of data peeks it
+      peekLocalSite(siteObj,Umu,lsite);
 
-	if ( rank != myrank ) {
-	  grid->RecvFrom((void *)&siteObj,rank,sizeof(siteObj));
-	} else { 
-	  peekLocalSite(siteObj,Umu,lsite);
+      // Pair of nodes may need to do pt2pt send
+      if ( rank != iorank ) { // comms is necessary
+	if ( (myrank == rank) || (myrank==iorank) ) { // and we have to do it
+	  // Send to IOrank 
+	  grid->SendRecvPacket((void *)&siteObj,(void *)&siteObj,rank,iorank,sizeof(siteObj));
 	}
+      }
+
+      grid->Barrier(); // necessary?
+
+      if (myrank == iorank) {
 	
 	munge(siteObj,fileObj,csum);
 
@@ -490,17 +659,16 @@ class BinaryIO {
 	
 	fout.seekp(offset+g_idx*sizeof(fileObj));
 	fout.write((char *)&fileObj,sizeof(fileObj));
-
-      } else { 
-	if ( myrank == rank ) {
-	  peekLocalSite(siteObj,Umu,lsite);
-	  grid->SendTo((void *)&siteObj,iorank,sizeof(siteObj));
-	} 
+	bytes+=sizeof(fileObj);
       }
-      grid->Barrier(); // necessary// or every 16 packets to rate throttle??
     }
 
     grid->GlobalSum(csum);
+    grid->GlobalSum(bytes);
+
+    timer.Stop();
+    std::cout<<GridLogPerformance<<"writeObjectParallel: wrote "<< bytes <<" bytes in "<<timer.Elapsed() <<" "
+	     << (double)bytes/timer.useconds() <<" MB/s "  <<std::endl;
 
     return csum;
   }

lib/parallelIO/NerscIO.h

@@ -1,3 +1,32 @@
+    /*************************************************************************************
+
+    Grid physics library, www.github.com/paboyle/Grid 
+
+    Source file: ./lib/parallelIO/NerscIO.h
+
+    Copyright (C) 2015
+
+Author: Matt Spraggs <matthew.spraggs@gmail.com>
+Author: Peter Boyle <paboyle@ph.ed.ac.uk>
+Author: paboyle <paboyle@ph.ed.ac.uk>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+    See the full license in the file "LICENSE" in the top level distribution directory
+    *************************************************************************************/
+    /*  END LEGAL */
 #ifndef GRID_NERSC_IO_H
 #define GRID_NERSC_IO_H
 
@@ -51,7 +80,6 @@ class NerscField {
     std::string floating_point;
 };
 
-
 //////////////////////////////////////////////////////////////////////
 // Bit and Physical Checksumming and QA of data
 //////////////////////////////////////////////////////////////////////
@@ -69,8 +97,9 @@ inline void NerscGrid(GridBase *grid,NerscField &header)
 template<class GaugeField>
 inline void NerscStatistics(GaugeField & data,NerscField &header)
 {
-  header.link_trace=Grid::QCD::WilsonLoops<GaugeField>::linkTrace(data);
-  header.plaquette =Grid::QCD::WilsonLoops<GaugeField>::avgPlaquette(data);
+  // How to convert data precision etc...
+  header.link_trace=Grid::QCD::WilsonLoops<PeriodicGimplR>::linkTrace(data);
+  header.plaquette =Grid::QCD::WilsonLoops<PeriodicGimplR>::avgPlaquette(data);
 }
 
 inline void NerscMachineCharacteristics(NerscField &header)
@@ -181,38 +210,41 @@ inline void NerscMachineCharacteristics(NerscField &header)
 class NerscIO : public BinaryIO { 
  public:
 
+  static inline void truncate(std::string file){
+    std::ofstream fout(file,std::ios::out);
+  }
+  
+  #define dump_nersc_header(field, s)\
+  s << "BEGIN_HEADER"      << std::endl;\
+  s << "HDR_VERSION = "    << field.hdr_version    << std::endl;\
+  s << "DATATYPE = "       << field.data_type      << std::endl;\
+  s << "STORAGE_FORMAT = " << field.storage_format << std::endl;\
+  for(int i=0;i<4;i++){\
+    s << "DIMENSION_" << i+1 << " = " << field.dimension[i] << std::endl ;\
+  }\
+  s << "LINK_TRACE = " << std::setprecision(10) << field.link_trace << std::endl;\
+  s << "PLAQUETTE  = " << std::setprecision(10) << field.plaquette  << std::endl;\
+  for(int i=0;i<4;i++){\
+    s << "BOUNDARY_"<<i+1<<" = " << field.boundary[i] << std::endl;\
+  }\
+  \
+  s << "CHECKSUM = "<< std::hex << std::setw(10) << field.checksum << std::dec<<std::endl;\
+  s << "ENSEMBLE_ID = "     << field.ensemble_id      << std::endl;\
+  s << "ENSEMBLE_LABEL = "  << field.ensemble_label   << std::endl;\
+  s << "SEQUENCE_NUMBER = " << field.sequence_number  << std::endl;\
+  s << "CREATOR = "         << field.creator          << std::endl;\
+  s << "CREATOR_HARDWARE = "<< field.creator_hardware << std::endl;\
+  s << "CREATION_DATE = "   << field.creation_date    << std::endl;\
+  s << "ARCHIVE_DATE = "    << field.archive_date     << std::endl;\
+  s << "FLOATING_POINT = "  << field.floating_point   << std::endl;\
+  s << "END_HEADER"         << std::endl;
+  
   static inline unsigned int writeHeader(NerscField &field,std::string file)
   {
-    std::ofstream fout(file,std::ios::out);
+    std::ofstream fout(file,std::ios::out|std::ios::in);
   
     fout.seekp(0,std::ios::beg);
-    fout << "BEGIN_HEADER"      << std::endl;
-    fout << "HDR_VERSION = "    << field.hdr_version    << std::endl;
-    fout << "DATATYPE = "       << field.data_type      << std::endl;
-    fout << "STORAGE_FORMAT = " << field.storage_format << std::endl;
-
-    for(int i=0;i<4;i++){
-      fout << "DIMENSION_" << i+1 << " = " << field.dimension[i] << std::endl ;
-    }
-    // just to keep the space and write it later
-    fout << "LINK_TRACE = " << std::setprecision(10) << field.link_trace << std::endl;
-    fout << "PLAQUETTE  = " << std::setprecision(10) << field.plaquette  << std::endl;
-    for(int i=0;i<4;i++){
-      fout << "BOUNDARY_"<<i+1<<" = " << field.boundary[i] << std::endl;
-    }
-
-    fout << "CHECKSUM = "<< std::hex << std::setw(10) << field.checksum << std::endl;
-    fout << std::dec;
-
-    fout << "ENSEMBLE_ID = "     << field.ensemble_id      << std::endl;
-    fout << "ENSEMBLE_LABEL = "  << field.ensemble_label   << std::endl;
-    fout << "SEQUENCE_NUMBER = " << field.sequence_number  << std::endl;
-    fout << "CREATOR = "         << field.creator          << std::endl;
-    fout << "CREATOR_HARDWARE = "<< field.creator_hardware << std::endl;
-    fout << "CREATION_DATE = "   << field.creation_date    << std::endl;
-    fout << "ARCHIVE_DATE = "    << field.archive_date     << std::endl;
-    fout << "FLOATING_POINT = "  << field.floating_point   << std::endl;
-    fout << "END_HEADER"         << std::endl;
+    dump_nersc_header(field, fout);
     field.data_start = fout.tellp();
     return field.data_start;
 }
@@ -287,9 +319,6 @@ static inline int readHeader(std::string file,GridBase *grid,  NerscField &field
   return field.data_start;
 }
 
-
-
-
 /////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
 // Now the meat: the object readers
 /////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
@@ -317,17 +346,17 @@ static inline void readConfiguration(Lattice<iLorentzColourMatrix<vsimd> > &Umu,
   if ( header.data_type == std::string("4D_SU3_GAUGE") ) {
     if ( ieee32 || ieee32big ) {
       //      csum=BinaryIO::readObjectSerial<iLorentzColourMatrix<vsimd>, LorentzColour2x3F> 
-      csum=BinaryIO::readObjectParallel<iLorentzColourMatrix<vsimd>, LorentzColour2x3F> 
+	csum=BinaryIO::readObjectParallel<iLorentzColourMatrix<vsimd>, LorentzColour2x3F> 
 	(Umu,file,Nersc3x2munger<LorentzColour2x3F,LorentzColourMatrix>(), offset,format);
     }
     if ( ieee64 || ieee64big ) {
-      //      csum=BinaryIO::readObjectSerial<iLorentzColourMatrix<vsimd>, LorentzColour2x3D> 
+      //csum=BinaryIO::readObjectSerial<iLorentzColourMatrix<vsimd>, LorentzColour2x3D> 
       csum=BinaryIO::readObjectParallel<iLorentzColourMatrix<vsimd>, LorentzColour2x3D> 
-	(Umu,file,Nersc3x2munger<LorentzColour2x3D,LorentzColourMatrix>(),offset,format);
+      	(Umu,file,Nersc3x2munger<LorentzColour2x3D,LorentzColourMatrix>(),offset,format);
     }
-  } else if ( header.data_type == std::string("4D_SU3_GAUGE_3X3") ) {
+  } else if ( header.data_type == std::string("4D_SU3_GAUGE_3x3") ) {
     if ( ieee32 || ieee32big ) {
-      //      csum=BinaryIO::readObjectSerial<iLorentzColourMatrix<vsimd>,LorentzColourMatrixF>
+      //csum=BinaryIO::readObjectSerial<iLorentzColourMatrix<vsimd>,LorentzColourMatrixF>
       csum=BinaryIO::readObjectParallel<iLorentzColourMatrix<vsimd>,LorentzColourMatrixF>
 	(Umu,file,NerscSimpleMunger<LorentzColourMatrixF,LorentzColourMatrix>(),offset,format);
     }
@@ -344,6 +373,7 @@ static inline void readConfiguration(Lattice<iLorentzColourMatrix<vsimd> > &Umu,
 
   assert(fabs(clone.plaquette -header.plaquette ) < 1.0e-5 );
   assert(fabs(clone.link_trace-header.link_trace) < 1.0e-6 );
+
   assert(csum == header.checksum );
 
   std::cout<<GridLogMessage <<"Read NERSC Configuration "<<file<< " and plaquette, link trace, and checksum agree"<<std::endl;
@@ -377,6 +407,8 @@ static inline void writeConfiguration(Lattice<iLorentzColourMatrix<vsimd> > &Umu
   uint32_t csum;
   int offset;
   
+  truncate(file);
+
   if ( two_row ) { 
 
     header.floating_point = std::string("IEEE64BIG");
@@ -389,28 +421,105 @@ static inline void writeConfiguration(Lattice<iLorentzColourMatrix<vsimd> > &Umu
     std::string file1 = file+"para";
     int offset1 = writeHeader(header,file1);
     int csum1=BinaryIO::writeObjectParallel<vobj,fobj2D>(Umu,file1,munge,offset,header.floating_point);
+    //int csum1=BinaryIO::writeObjectSerial<vobj,fobj2D>(Umu,file1,munge,offset,header.floating_point);
 
     
     std::cout << GridLogMessage << " TESTING PARALLEL WRITE offsets " << offset1 << " "<< offset << std::endl;
-    std::cout << GridLogMessage <<std::hex<< " TESTING PARALLEL WRITE csums   " << csum1 << " "<< csum << std::endl;
-    std::cout << std::dec;
+    std::cout << GridLogMessage << " TESTING PARALLEL WRITE csums   " << csum1 << " "<<std::hex<< csum << std::dec<< std::endl;
 
     assert(offset1==offset);  
     assert(csum1==csum);  
 
   } else { 
     header.floating_point = std::string("IEEE64BIG");
-    header.data_type      = std::string("4D_SU3_GAUGE_3X3");
+    header.data_type      = std::string("4D_SU3_GAUGE_3x3");
     NerscSimpleUnmunger<fobj3D,sobj> munge;
     BinaryIO::Uint32Checksum<vobj,fobj3D>(Umu, munge,header.checksum);
     offset = writeHeader(header,file);
-    csum=BinaryIO::writeObjectSerial<vobj,fobj3D>(Umu,file,munge,offset,header.floating_point);
+    //    csum=BinaryIO::writeObjectSerial<vobj,fobj3D>(Umu,file,munge,offset,header.floating_point);
+    csum=BinaryIO::writeObjectParallel<vobj,fobj3D>(Umu,file,munge,offset,header.floating_point);
   }
 
   std::cout<<GridLogMessage <<"Written NERSC Configuration "<<file<< " checksum "<<std::hex<<csum<< std::dec<<" plaq "<< header.plaquette <<std::endl;
 
  }
+
+
+    ///////////////////////////////
+    // RNG state
+    ///////////////////////////////
+static inline void writeRNGState(GridSerialRNG &serial,GridParallelRNG &parallel,std::string file)
+{
+  typedef typename GridParallelRNG::RngStateType RngStateType;
+
+  // Following should become arguments
+  NerscField header;
+  header.sequence_number = 1;
+  header.ensemble_id     = "UKQCD";
+  header.ensemble_label  = "DWF";
+
+  GridBase *grid = parallel._grid;
+
+  NerscGrid(grid,header);
+  header.link_trace=0.0;
+  header.plaquette=0.0;
+  NerscMachineCharacteristics(header);
+
+  uint32_t csum;
+  int offset;
+  
+#ifdef RNG_RANLUX
+    header.floating_point = std::string("UINT64");
+    header.data_type      = std::string("RANLUX48");
+#else
+    header.floating_point = std::string("UINT32");
+    header.data_type      = std::string("MT19937");
+#endif
+
+  truncate(file);
+  offset = writeHeader(header,file);
+  csum=BinaryIO::writeRNGSerial(serial,parallel,file,offset);
+  header.checksum = csum;
+  offset = writeHeader(header,file);
+
+  std::cout<<GridLogMessage <<"Written NERSC RNG STATE "<<file<< " checksum "<<std::hex<<csum<<std::dec<<std::endl;
+
+ }
+    
+static inline void readRNGState(GridSerialRNG &serial,GridParallelRNG & parallel,NerscField& header,std::string file)
+{
+  typedef typename GridParallelRNG::RngStateType RngStateType;
+
+  GridBase *grid = parallel._grid;
+
+  int offset = readHeader(file,grid,header);
+
+  NerscField clone(header);
+
+  std::string format(header.floating_point);
+  std::string data_type(header.data_type);
+
+#ifdef RNG_RANLUX
+  assert(format == std::string("UINT64"));
+  assert(data_type == std::string("RANLUX48"));
+#else
+  assert(format == std::string("UINT32"));
+  assert(data_type == std::string("MT19937"));
+#endif
+
+  // depending on datatype, set up munger;
+  // munger is a function of <floating point, Real, data_type>
+  uint32_t csum=BinaryIO::readRNGSerial(serial,parallel,file,offset);
+
+  std::cerr<<" Csum "<< csum << " "<< header.checksum <<std::endl;
+
+  assert(csum == header.checksum );
+
+  std::cout<<GridLogMessage <<"Read NERSC RNG file "<<file<< " format "<< data_type <<std::endl;
+}
+
 };
 
+
 }}
 #endif