Azusa Yamaguchi
a8b9109cc8
multishift conjugate gradient added and a strong test: take a diagonal
...
but non-identity matrix
l1 0 0 0 ....
0 l2 0 0 ....
0 0 l3 0 ...
. . .
. . .
. . .
And apply the multishift CG to it. Sum the poles and residues.
Insist that this be the same as the exactly taken square root
where l1,l2,l3 >= 0.
2015-06-08 11:52:44 +01:00
Azusa Yamaguchi
5f33cc3a95
Compile fix
2015-06-05 10:29:42 +01:00
Azusa Yamaguchi
bb36139fc2
Fix
2015-06-05 10:21:28 +01:00
Azusa Yamaguchi
cc5f518b21
Endif terminated
2015-06-05 10:19:42 +01:00
Peter Boyle
337b6e83af
Rework the linop support to get different forms of red black schur solver
...
Moo on diag, or MooInv Moe MeeInv Meo
2015-06-05 10:17:10 +01:00
Azusa Yamaguchi
8f9627520b
merge to the head
2015-06-05 10:15:31 +01:00
Azusa Yamaguchi
db84b19443
Merge branch 'master' of https://github.com/paboyle/Grid
2015-06-05 10:04:46 +01:00
Azusa Yamaguchi
7d984b9547
Adding some wilson loop support
2015-06-05 10:02:36 +01:00
Peter Boyle
f22170ad49
comment improvement
2015-06-05 05:31:27 +01:00
Peter Boyle
7678fbd30d
PartialFraction Hw with Zolo and Tanh approx converged under CG and passed EO breakdown
...
and hermiticity tests.
2015-06-04 13:28:37 +01:00
neo
7e47b0c6eb
Corrected small compilation bug in traceIndex for iVectors
2015-06-04 19:01:43 +09:00
neo
41e88c232b
Merge remote-tracking branch 'upstream/master'
2015-06-04 18:30:29 +09:00
neo
949b6a7afa
Added support for Ta to Lattice types
2015-06-04 18:29:55 +09:00
neo
bb73569fd6
Addedd Ta functionality to the tensor types
...
Merge remote-tracking branch 'upstream/master'
Conflicts:
configure
2015-06-04 18:11:32 +09:00
Peter Boyle
54c082dc35
Allow traceIndex on a different index to distribute replicated across a vector index
2015-06-04 09:41:16 +01:00
neo
c6f2ee91f6
Small modification to the configure files
2015-06-04 14:17:58 +09:00
Peter Boyle
9c1ab656d4
CG Tests work for wilson kernel cont frac zolo and tanh
2015-06-04 06:02:00 +01:00
Peter Boyle
1ad689e4d5
Implementing the Hw kernel continued fraction 5d overlap cases
2015-06-04 00:23:16 +01:00
Peter Boyle
802e94e9ca
First pass at continued fraction; solver and even odd decomposition tests pass.
...
Have to make ContFrac class virtual and derive end non-abstract actions for the particular
cases.
2015-06-04 00:00:45 +01:00
Peter Boyle
e68d087010
Assist for generating file lists contained in Make.inc files for convenience when things are added
2015-06-03 13:07:00 +01:00
Peter Boyle
3254bb2c8e
Make.inc needed in repo
2015-06-03 12:49:36 +01:00
Peter Boyle
f9b070d64d
Reorganise of file naming
2015-06-03 12:47:05 +01:00
Peter Boyle
6cb38dc5dc
Overlap Wilson Cayley tanh & zolo
2015-06-03 11:26:54 +01:00
Peter Boyle
c659c76053
Scaled Shamir and Scaled Shamir Zolotarev aliases for special cases of Mobius.
2015-06-03 09:51:06 +01:00
Peter Boyle
68e26140ee
Mobius Caley form, Mobius Zolotarev operators. Pass Even Odd vs unprec test and hermiticity checks
...
in tests/Grid_any_evenodd.cc; will work on inversion tests shortly.
2015-06-03 09:36:26 +01:00
Peter Boyle
0bc004de7c
Domain wall fermions now invert ; have the basis set up for
...
Tanh/Zolo * (Cayley/PartFrac/ContFrac) * (Mobius/Shamir/Wilson)
Approx Representation Kernel.
All are done with space-time taking part in checkerboarding, Ls uncheckerboarded
Have only so far tested the Domain Wall limit of mobius, and at that only checked
that it
i) Inverts
ii) 5dim DW == Ls copies of 4dim D2
iii) MeeInv Mee == 1
iv) Meo+Mee+Moe+Moo == M unprec.
v) MpcDagMpc is hermitan
vi) Mdag is the adjoint of M between stochastic vectors.
That said, the RB schur solve, RB MpcDagMpc solve, Unprec solve
all converge and the true residual becomes small; so pretty good tests.
2015-06-02 16:57:12 +01:00
Azusa Yamaguchi
8f87950dc1
FIx miistake
2015-06-01 12:26:20 +01:00
Azusa Yamaguchi
00bf7f4d42
Const safety
2015-06-01 12:25:59 +01:00
Azusa Yamaguchi
eb28a64c3c
No compile fix on mpi target
2015-05-31 22:50:03 +01:00
azusayamaguchi
328aa9ae49
Bug in Makefile.am fixed
2015-05-31 18:50:08 +01:00
Peter Boyle
66d997e031
Large scale change to support 5d fermion formulations.
...
Have 5d replicated wilson with 4d gauge working and matrix regressing
to Ls copies of wilson.
2015-05-31 15:09:02 +01:00
neo
661c7e3e37
Merge remote-tracking branch 'upstream/master'
2015-05-29 11:41:39 +09:00
Peter Boyle
62dccb3247
Weak scale the benchmarks automatically.
2015-05-28 13:47:01 +01:00
Peter Boyle
c0c1ebe757
Works now with Clang-avx, Clang-sse and ICPC-avx, ICPC-sse
2015-05-28 11:35:43 +01:00
Peter Boyle
bd81ac0f17
Improving the reduction to go through our on permute.
...
Must also do this for avx512
2015-05-27 16:07:17 +01:00
neo
d8b05e001d
Check at configure time if CPU supports the requested SIMD optimization
2015-05-27 18:30:11 +09:00
neo
9fd6506d1f
Included Gpermute in the new Grid_simd.h file style.
...
Now tested for SSE4. OK
2015-05-27 12:11:44 +09:00
neo
75442e48ce
Added a .gitignore file to eliminate autoconf files from commits
2015-05-27 11:10:51 +09:00
Guido Cossu
26ec41288d
Corrected AVX regression error. Tested.
2015-05-27 10:49:33 +09:00
neo
13707e0808
Merge remote-tracking branch 'upstream/master'
...
Conflicts:
Makefile.in
2015-05-27 10:41:33 +09:00
neo
12ae11ef62
Adding support for doxygen generation
2015-05-27 10:34:56 +09:00
Peter Boyle
22d073eb2b
Simd revert to Guido's commit. I edited concurrently and things went bad.
2015-05-26 22:20:09 +01:00
Peter Boyle
ccf10a973a
Merge branch 'master' of https://github.com/paboyle/Grid
...
Conflicts:
lib/Grid_simd.h
2015-05-26 20:04:08 +01:00
Peter Boyle
6ef0096dc9
Strip out the dslash kernel implementation
2015-05-26 19:55:18 +01:00
Peter Boyle
20100d0a40
Hand unrolled version of dslash in a separate class.
...
Useful to compare; raises Intel compiler from 9GFlop/s to 17.5 Gflops.
on ivybridge core. Raises Clang form 14.5 to 17.5
2015-05-26 19:54:03 +01:00
neo
c04cad92ac
More cleanup of Grid_simd.h
2015-05-26 13:54:34 +09:00
neo
aff978f60a
Cleaning up simd files
2015-05-26 13:31:10 +09:00
neo
9ad6d0c65f
Merge remote-tracking branch 'upstream/master'
...
Conflicts:
lib/math/Grid_math_tensors.h
lib/simd/Grid_vector_types.h
2015-05-26 13:14:06 +09:00
neo
377083e6ae
checked performance of new vector libaries.
...
Added check for c++11 support on the configure.ac
2015-05-26 12:02:54 +09:00
Peter Boyle
d7f5172860
Schur complement based red-black inversion working
2015-05-25 13:47:12 +01:00
