paboyle
285deab432
Coordinate handling GPU friendly. Avoid std::vector
2018-02-24 22:19:28 +00:00
paboyle
bb7d87d0a0
Coordinate handling gpu friendly
2018-02-24 22:18:33 +00:00
paboyle
b9b5bdfc3a
Proper offload (accelerator access) will require a mutable copy lambda.
2018-02-02 11:38:19 +00:00
paboyle
51eb2c5dfc
Make referencign the stencil and all info required to evaluate the kernel
...
accelerator marked up
2018-02-02 11:37:13 +00:00
paboyle
ede0dff794
Mark up as an accelerator function
2018-02-02 11:36:44 +00:00
paboyle
aa6de818e2
Copy data needed by Kernels out of the grid object to avoid host reference
2018-02-02 11:36:11 +00:00
paboyle
dcf6517a93
Accelerator offload and copy Opt into the kernel for GPU host var safety
2018-02-02 11:35:35 +00:00
paboyle
a308dff410
accelerator loop, copy Opt into the GPU
2018-02-02 11:34:37 +00:00
paboyle
14ba20898a
Accelerator loop the key kernel call
2018-02-02 11:30:07 +00:00
paboyle
a53d3ee19a
Add Opt to the lambda capture to get it into the GPU
2018-02-02 11:28:39 +00:00
paboyle
5df435319d
Use constexpr
2018-02-02 11:27:56 +00:00
paboyle
0da2d3e222
accelerator off load some more stuff
2018-02-02 11:27:35 +00:00
paboyle
9c9dfbfa78
Force accelerator
2018-02-02 11:25:09 +00:00
paboyle
e4df025d01
Accelerator related
2018-02-01 23:20:05 +00:00
paboyle
cfeda9d536
constexpr on const ints
2018-02-01 22:59:12 +00:00
paboyle
4450b1993a
Offload
2018-02-01 22:45:47 +00:00
paboyle
d03ce5c2a4
Provide a way to get around std::vector for a known type on device.
...
Use template specialisation to access a private member in the Clang++ STL implementation
2018-02-01 22:44:25 +00:00
paboyle
7d6522c1ef
Accelerator inline
2018-02-01 22:43:56 +00:00
paboyle
b96832a922
Accelerator inline
2018-02-01 22:43:26 +00:00
paboyle
5d7af47b05
accelerator_inline
2018-02-01 22:42:54 +00:00
paboyle
053ef25c90
constexpr makes GPU happy
2018-02-01 22:42:29 +00:00
paboyle
8ae77d3706
Small simplification of FermionOperatorImpl towards GPU but not there yet
2018-02-01 22:41:54 +00:00
paboyle
79b50feacf
fixme updates
2018-01-29 16:00:40 +00:00
paboyle
c67c1544cd
abs no compile on travis fix attempt
2018-01-28 10:26:04 +00:00
paboyle
e657f9a344
OMP collapse changes to make NVCC happy
2018-01-28 01:21:53 +00:00
paboyle
b6ebf35af5
Intel compiler doesn't like Nvidia error disable pragmas
2018-01-28 01:03:10 +00:00
paboyle
604c05f4b8
parallel_for elimination -> thread_loop
2018-01-28 01:01:36 +00:00
paboyle
70e276e1ab
parallel_for elimination -> thread_loop
2018-01-28 01:01:14 +00:00
paboyle
9472b02771
Parallel_for elimination -> thread_loop.
2018-01-28 01:00:55 +00:00
paboyle
9597ab94eb
Zero changes, swap on lattice type.
2018-01-27 23:51:40 +00:00
paboyle
ce4da83bc2
Zero changes, literally
2018-01-27 23:51:10 +00:00
paboyle
d557f3ef77
Zero changes (literally) and also a warning elimination
2018-01-27 23:50:43 +00:00
paboyle
f574c20118
Zero changes, __VA_ARGS__ and swap
2018-01-27 23:50:17 +00:00
paboyle
f102897385
VA_ARGS to make comma safe automatic
2018-01-27 23:49:47 +00:00
paboyle
d6fce3e498
Zero changes, literally
2018-01-27 23:48:01 +00:00
paboyle
2d0bcc2606
Zero changes, acceleartor on kernels and some thread loop changes
2018-01-27 23:47:38 +00:00
paboyle
45df59720e
Zero changes and VA_ARGS changes
2018-01-27 23:46:58 +00:00
paboyle
44ef5bc207
Zero changes (literally speaking).
2018-01-27 23:46:28 +00:00
paboyle
98af36217a
Zero changes. (I mean literally)
2018-01-27 23:46:02 +00:00
paboyle
be7b37b9c9
Mistake on openmp
2018-01-27 00:05:11 +00:00
paboyle
c4f82e072b
_grid becomes private ; use Grid()§
2018-01-27 00:04:12 +00:00
paboyle
3f9654e397
Hiding internals
2018-01-26 23:09:03 +00:00
paboyle
912b50f6fa
Hiding lattice internals
2018-01-26 23:08:45 +00:00
paboyle
2a4a0e43c1
Hide internals
2018-01-26 23:08:27 +00:00
paboyle
32523a229c
Hide internals
2018-01-26 23:08:02 +00:00
paboyle
1ebd56c3fb
Hide internal data
2018-01-26 23:07:34 +00:00
paboyle
8dccffdfd5
Hide internal data
2018-01-26 23:06:51 +00:00
paboyle
5642ea270f
Hide internal data
2018-01-26 23:06:28 +00:00
paboyle
43cea62855
Hide internal data
2018-01-26 23:06:03 +00:00
paboyle
2b4067bb71
Hide internal data
2018-01-26 23:05:32 +00:00
paboyle
85771e97e9
Hide internal data
2018-01-26 23:04:46 +00:00
paboyle
8b371ffa94
Hide internal data
2018-01-26 23:03:54 +00:00
paboyle
bf659dfd92
Hide the ._odata
2018-01-26 22:27:47 +00:00
paboyle
76a4dd36d9
Fix no compile of test serialisation
2018-01-26 00:13:21 +00:00
paboyle
f4010023ca
Warning fixes
2018-01-25 23:46:47 +00:00
paboyle
24a4589def
Changes to interface a little
2018-01-25 23:37:34 +00:00
paboyle
c904822e74
Warning removal
2018-01-25 23:37:15 +00:00
paboyle
40ee1e1957
Zero()
2018-01-25 23:36:58 +00:00
paboyle
461df78a3f
Better to use Zero(), and not zero static data
2018-01-25 23:36:22 +00:00
paboyle
db9c9475d4
const
2018-01-25 23:36:06 +00:00
paboyle
214f7a6f13
Drop std::vector container for the lattice data
2018-01-25 23:35:04 +00:00
paboyle
c844cfcda8
Remove commAllocator; make more simple; option to switch off the pointer caceh
2018-01-25 23:33:57 +00:00
paboyle
a3e3034e6f
Host compile
2018-01-25 23:33:00 +00:00
paboyle
e7cba358c2
Temporary update to reflect the new dropping of std::vector in Lattice
...
Will update again to hide the internals in an interface
2018-01-25 23:31:41 +00:00
paboyle
99329197ee
Rename header to .h
2018-01-24 14:10:09 +00:00
paboyle
421401af55
Remove IMCI as really don't support
2018-01-24 13:53:21 +00:00
paboyle
0626c1e39e
Accelerator flaggina dn thrust complex for NVCC
2018-01-24 13:50:41 +00:00
paboyle
725f03e2e2
Accelerator markup and thrust complex on nvcc
2018-01-24 13:50:10 +00:00
paboyle
65f77112e0
Thread loops done properly
2018-01-24 13:49:39 +00:00
paboyle
408b868475
Generic for GPU needs accelerator markup of functions
2018-01-24 13:49:12 +00:00
paboyle
1c797deb04
Accelerator tweaks
2018-01-24 13:43:43 +00:00
paboyle
b9d5a42b57
Should be able to eliminate the COMMA_SAFE with VA_ARGS trick ; revisit this file
2018-01-24 13:42:06 +00:00
paboyle
e737591918
Accelerator loops
2018-01-24 13:41:12 +00:00
paboyle
ba5ea5830b
Acceleartor loops
2018-01-24 13:40:56 +00:00
paboyle
43f244badf
Thread loops for now; figure out what can be GPU accelerated later here
2018-01-24 13:40:30 +00:00
paboyle
e9c8ba5ef7
Accelerator loosp
2018-01-24 13:39:54 +00:00
paboyle
d70709a8e8
Thread construct changes
2018-01-24 13:39:06 +00:00
paboyle
733f8ff0b2
Still using parallel_for -- don't know how to implement reduction on GPU yet. Look at some sample code is best.
2018-01-24 13:38:13 +00:00
paboyle
0bfa5bb213
Accelerator loosp
2018-01-24 13:37:26 +00:00
paboyle
1f26a234f9
CPU loops explicit for peek poke
2018-01-24 13:36:31 +00:00
paboyle
13f0116425
Accelerator loops
2018-01-24 13:35:55 +00:00
paboyle
25f589b064
Accelerator loops
2018-01-24 13:35:36 +00:00
paboyle
210c50a278
Accelerator prep work
2018-01-24 13:35:13 +00:00
paboyle
549a143e78
Accelerator related
2018-01-24 13:34:46 +00:00
paboyle
277301486d
Simple warning elimination
2018-01-24 13:34:15 +00:00
paboyle
c851b39a49
Nicer way of including aggregate
2018-01-24 13:33:34 +00:00
paboyle
15cc12eb6c
Delete the old non ET file
2018-01-24 13:33:07 +00:00
paboyle
ae4f1f8c12
New file, split out two from Lattice_reduction
2018-01-24 13:32:43 +00:00
paboyle
5609624b44
Threading constructs replaced
2018-01-24 13:32:24 +00:00
paboyle
b5a947dd79
Change to make NVCC happy
2018-01-24 13:32:02 +00:00
paboyle
ee16f62322
stray semicolon elimination. NVCC is picky, but eventually picked up these diags
...
with a pragma to suppress
2018-01-24 13:31:17 +00:00
paboyle
3318de27d6
Thread macro changes
2018-01-24 13:30:23 +00:00
paboyle
ac56965306
GPU changes and threading macros replaced
2018-01-24 13:28:30 +00:00
paboyle
8e99264f40
Accelerator mark up of entire tensore space for offload
2018-01-24 13:27:30 +00:00
paboyle
69327db9a9
Improviements for NVCC. Eigen is not compat with CUDA 9 and must hack to disable device
...
compilation
2018-01-24 13:25:07 +00:00
paboyle
7331ee2d80
Warnings control to overpower the NVCC compiler
2018-01-24 13:24:36 +00:00
paboyle
918c105c57
NVCC warning elimination
2018-01-24 13:23:59 +00:00
paboyle
be1511d469
Remove old macros for threading
2018-01-24 13:23:24 +00:00
paboyle
f1c31df9d2
updated Eigen version. Still didn't fix CUDA 9 no compile.
...
Worked around by switching off __NVCC__ during the include of Eigen and switching it
back on after. No Eigen code can be offloaded, note as a rsult of this. No harm done.
2018-01-24 13:19:29 +00:00
paboyle
ff7b587fad
Ugly... nvcc needs -x cu to compile .cc as cuda.
...
Since CXXFLAGS is Also passed to linker, and -x cu breaks link phase must replace
CXX and CXXLD with nvcc -x cu and nvcc -link respectively.
2018-01-24 13:18:19 +00:00
paboyle