Peter Boyle
adbdc4e65b
Half comms not working on GPU yet, so disable.
2018-09-11 05:15:22 +01:00
Peter Boyle
e4deea4b94
Weird bug appears with Vector<Vector<>>.
...
"fix" with std::vector<Vector<>>
Lies in the face table code. But think there is some latent problem.
Possibly in my allocator since it is caching, but could simplify or eliminate the caching
option and retest. One to look at later.
2018-09-11 04:36:57 +01:00
Peter Boyle
94d721a20b
Comments on further topology discovery work
2018-09-11 04:20:04 +01:00
Peter Boyle
7bf82f5b37
Offload the face handling to GPU
2018-09-10 11:28:42 +01:00
Peter Boyle
f02c7ea534
Peer to peer on GPU's setup
2018-09-10 11:26:20 +01:00
Peter Boyle
bc503b60e6
Offloadable gather code
2018-09-10 11:21:25 +01:00
Peter Boyle
704ca162c1
Offloadable compression
2018-09-10 11:20:50 +01:00
Peter Boyle
b5329d8852
Protect against zero length loops giving a kernel call failure
2018-09-10 11:20:07 +01:00
Peter Boyle
f27b9347ff
Better unquiesce MPI coverage
2018-09-10 11:19:39 +01:00
Peter Boyle
b4967f0231
Verbose and error trapping cleaner
2018-09-09 14:28:02 +01:00
Peter Boyle
6d0f1aabb1
Fix the multi-node path
2018-09-09 14:27:37 +01:00
Peter Boyle
f4bfeb835d
Drop back to smaller Ls
2018-09-09 14:25:06 +01:00
Peter Boyle
394b7b6276
Verbose decrease
2018-09-09 14:24:46 +01:00
Peter Boyle
da17a015c7
Pack the stencil smaller for 128 bit access
2018-07-23 06:12:45 -04:00
Peter Boyle
1fd08c21ac
make simd width configure time option for GPU
2018-07-23 06:10:55 -04:00
Peter Boyle
28db0631ff
Hack to force 128bit accesses
2018-07-23 06:10:27 -04:00
Peter Boyle
b35401b86b
Fix CUDA_ARCH. Need to simplify. See when new eigen release happens
2018-07-23 06:09:33 -04:00
Peter Boyle
a0714de8ec
Define vector length for GPU
2018-07-23 06:09:05 -04:00
Peter Boyle
21a1710b43
Verbose vector length
2018-07-23 06:08:39 -04:00
Peter Boyle
b2b5137d28
Finally starting to get decent performance on Volta
2018-07-13 12:06:18 -04:00
Peter Boyle
2cc07450f4
Fastest option for the dslash
2018-07-05 09:57:55 -04:00
Peter Boyle
c0e8bc9da9
Current version gets 250 - 320 GF/s on Volta on the target 12^4 volume.
2018-07-05 07:10:25 -04:00
Peter Boyle
b1265ae867
Prettify code
2018-07-05 07:08:06 -04:00
Peter Boyle
32bb85ea4c
Standard extractLane is fast
2018-07-05 07:07:30 -04:00
Peter Boyle
ca0607b6ef
Clearer kernel call meaning
2018-07-05 07:06:15 -04:00
Peter Boyle
19b527e83f
Better extract merge for GPU. Let the SIMD header files define the pointer type for
...
access. GPU redirects through builtin float2, double2 for complex
2018-07-05 07:05:13 -04:00
Peter Boyle
4730d4692a
Fast lane extract, saturates bandwidth on Volta for SU3 benchmarks
2018-07-05 07:03:33 -04:00
Peter Boyle
1bb456c0c5
Minor GPU vector width changeÂ
2018-07-05 07:02:04 -04:00
Peter Boyle
4b04ae3611
Printing improvement
2018-07-05 06:59:38 -04:00
Peter Boyle
2f776d51c6
Gpu specific benchmark saturates memory. Can enhance Grid to do this for expressions,
...
but a bitof (known) work.
2018-07-05 06:58:37 -04:00
paboyle
3a50afe7e7
GPU dslash updates
2018-06-27 22:32:21 +01:00
paboyle
f8e880b445
Loop for s and xyzt offlow
2018-06-27 21:49:57 +01:00
paboyle
3e947527cb
Move looping over "s" and "site" into kernels for GPU optimisatoin
2018-06-27 21:29:43 +01:00
paboyle
31f65beac8
Move site and Ls looping into the kernels
2018-06-27 21:28:48 +01:00
paboyle
38e2a32ac9
Single SIMD lane operations for CUDA
2018-06-27 21:28:06 +01:00
paboyle
efa84ca50a
Keep Cuda 9.1 happy
2018-06-27 21:27:32 +01:00
paboyle
5e96d6d04c
Keep CUDA happy
2018-06-27 21:27:11 +01:00
paboyle
df30bdc599
CUDA happy
2018-06-27 21:26:49 +01:00
paboyle
7f45222924
Diagnostics on memory alloc fail
2018-06-27 21:26:20 +01:00
paboyle
dd891f5e3b
Use NVCC to suppress device Eigen
2018-06-27 21:25:17 +01:00
paboyle
6c97a6a071
Coalescing version of the kernel
2018-06-13 20:52:29 +01:00
paboyle
73bb2d5128
Ugly hack to speed up compile on GPU; we don't use the hand kernels on GPU anyway so why compile
2018-06-13 20:35:28 +01:00
paboyle
b710fec6ea
Gpu code first version of specialised kernel
2018-06-13 20:34:39 +01:00
paboyle
b2a8cd60f5
Doubled gauge field is useful
2018-06-13 20:27:47 +01:00
paboyle
867ee364ab
Explicit instantiation hooks
2018-06-13 20:27:12 +01:00
paboyle
25becc9324
GPU tweaks for benchmarking; really necessary?
2018-06-13 20:26:07 +01:00
paboyle
94d1ae4c82
Some prep work for GPU shared memory. Need to be careful, as will try GPU direct
...
RDMA and inter-GPU memory sharing on SUmmit later
2018-06-13 20:24:06 +01:00
paboyle
2075b177ef
CUDA_ARCH more carefule treatment
2018-06-13 20:22:34 +01:00
paboyle
847c761ccc
Move sfw IEEE fp16 into central location
2018-06-13 20:22:01 +01:00
paboyle
8287ed8383
New GPU vector targets
2018-06-13 20:21:35 +01:00
paboyle
e6be7416f4
Use managed memory
2018-06-13 20:14:00 +01:00
paboyle
26863b6d95
User Managed memory
2018-06-13 20:13:42 +01:00
paboyle
ebd730bd54
Adding 2D loops
2018-06-13 20:13:01 +01:00
paboyle
066be31a3b
Optional GPU target SIMD types; work in progress and trying experiments
2018-06-13 20:07:55 +01:00
paboyle
7a4c142955
Add GPU specific simd targets
2018-06-13 19:55:30 +01:00
Peter Boyle
eb7d34a4cc
GPU version
2018-05-14 19:41:47 -04:00
Peter Boyle
aab27a655a
Start of GPU kernels
2018-05-14 19:41:17 -04:00
Peter Boyle
93280bae85
Gpu option
2018-05-14 19:40:58 -04:00
Peter Boyle
c5f93abcd7
GPU clean up
2018-05-14 19:40:33 -04:00
Peter Boyle
d5deef782d
Useful debug comments
2018-05-14 19:39:52 -04:00
Peter Boyle
5f50473c0d
Clean up
2018-05-14 19:39:11 -04:00
Peter Boyle
13f50406e3
Suppress print statement
2018-05-12 18:00:00 -04:00
Peter Boyle
09cd46d337
Lane by Lane operation
2018-05-12 17:59:35 -04:00
Peter Boyle
d3f51065c2
Give command line control of blocks/threads split
2018-05-12 17:58:56 -04:00
Peter Boyle
925ac4173d
Thread count control for warp scheduler thingy doodaa thing
2018-05-12 17:58:22 -04:00
Peter Boyle
eb921041d0
Perf count control
2018-05-12 17:57:32 -04:00
Peter Boyle
87c5c0271b
Ficxing eigen
2018-04-16 19:08:07 -04:00
Peter Boyle
a3f5a13591
Better Eigen handling
2018-04-16 18:02:55 -04:00
Peter Boyle
9fe28f00eb
Eigen sim link off head revision
2018-04-16 17:54:46 -04:00
Peter Boyle
a8a0bb85cc
Control scalar execution or vector under generic. Disable Eigen vectorisation on powerpc / SUmmit
2018-04-12 12:32:57 -04:00
Peter Boyle
6411caad67
work distribution
2018-04-12 11:41:41 -04:00
Peter Boyle
7533035a99
Control Eigen vectorisatoin
2018-04-12 11:40:56 -04:00
Peter Boyle
b15db11c60
Kernels -> pure static object to enable device execution
2018-03-24 19:35:20 -04:00
Peter Boyle
f6077f9d48
Kernels -> not instantiaed otherwise object ref on GPU
2018-03-24 19:33:44 -04:00
Peter Boyle
572954ef12
Kernels not an instantiated object, just static
2018-03-24 19:33:13 -04:00
Peter Boyle
cedeaae7db
Lebesge -> StencilView if necessary
2018-03-24 19:32:41 -04:00
Peter Boyle
e6cf0b1e17
View typedefs go to OperatorImpl
2018-03-24 19:32:11 -04:00
Peter Boyle
5412628ea6
begin end lamda
2018-03-24 19:31:45 -04:00
Peter Boyle
1f70cedbab
Have to make all kernel called routines static since object reference will be a host pointer on GPU
2018-03-24 19:29:26 -04:00
Peter Boyle
b50f37cfb4
Remove overlap comms flag
2018-03-24 19:28:53 -04:00
Peter Boyle
cb0d2a1b03
threaded rng init; I thought this was on
2018-03-24 19:28:17 -04:00
Peter Boyle
6fe9b28a82
Cosmetic
2018-03-24 19:27:14 -04:00
Peter Boyle
b002587d7c
Simplify
2018-03-24 19:26:44 -04:00
Peter Boyle
6c08385782
Simplify
2018-03-24 19:26:19 -04:00
Peter Boyle
4e1272fabf
Kernels need to be static to work on GPU. No reference to host resident data
2018-03-22 18:44:53 -04:00
Peter Boyle
607dc2d3c6
Remove lebesgue order
2018-03-22 18:23:09 -04:00
Peter Boyle
23c880b009
Remove lebesgue order; stick in stencil if need
2018-03-22 18:13:41 -04:00
Peter Boyle
334bb6792f
Lebesgue order removed. Stick in the stencil view
2018-03-22 18:12:12 -04:00
Peter Boyle
a3690071b4
Warm up GPu
2018-03-22 18:05:20 -04:00
Peter Boyle
299d119013
GPU work allocation improved
2018-03-22 18:04:24 -04:00
Peter Boyle
55be842d23
Dont force l1p.h so early
2018-03-22 18:01:43 -04:00
Peter Boyle
9875c446c6
Clean up pragmas
2018-03-20 07:19:17 -04:00
Peter Boyle
9c25eb35ca
Eigen develop branch for now
2018-03-20 07:18:56 -04:00
Peter Boyle
5ac96dbdc6
Warm behaviour in SU3 benchmark
2018-03-20 07:18:31 -04:00
Peter Boyle
5cc9aca85d
Use 64bit index for looping
2018-03-20 06:34:52 -04:00
Peter Boyle
ac29ebcb95
Clean up debug prints
2018-03-20 06:33:59 -04:00
Peter Boyle
a5cfb89304
Update eigen process direct from develop on github. Dangerous, but needed from GPU
2018-03-19 07:20:48 -04:00
Peter Boyle
f04a7251cc
Gpu welcome message and device info
2018-03-19 07:12:12 -04:00
Peter Boyle
d4ce7d9905
GPU friendly Stencil needs a view
2018-03-19 07:11:21 -04:00
Peter Boyle
8a1d303ab9
GPU friendly stencil improvements
2018-03-19 07:11:03 -04:00
Peter Boyle