da59379 | Peter Boyle | 26 March 2024, 17:03:20 UTC | Large reg file for double | 26 March 2024, 17:03:20 UTC |
3ef2a41 | Peter Boyle | 26 March 2024, 14:50:32 UTC | ifdef guard ommitted | 26 March 2024, 14:50:32 UTC |
aa96f42 | Peter Boyle | 26 March 2024, 14:41:25 UTC | Acclerator ware MPI guard on the Unix domain sockets | 26 March 2024, 14:41:25 UTC |
49e9e4e | Peter Boyle | 26 March 2024, 14:14:06 UTC | Fences | 26 March 2024, 14:14:06 UTC |
f7b8163 | Peter Boyle | 26 March 2024, 14:11:40 UTC | Deterministic MPI reduce options | 26 March 2024, 14:11:40 UTC |
93769ea | Peter Boyle | 26 March 2024, 14:10:24 UTC | Updated configure for bounce through host | 26 March 2024, 14:10:24 UTC |
59b0cc1 | Peter Boyle | 26 March 2024, 00:42:40 UTC | REduce the time in single | 26 March 2024, 00:42:40 UTC |
f32c275 | Peter Boyle | 26 March 2024, 00:42:00 UTC | Updated config options for MPI not being aware of GPU | 26 March 2024, 00:42:00 UTC |
5404fc6 | Peter Boyle | 26 March 2024, 00:38:41 UTC | Merge needs a fence on SYCL | 26 March 2024, 00:38:41 UTC |
1f53458 | Peter Boyle | 26 March 2024, 00:37:19 UTC | Options to bounce through a host buffer if --disable-accelerator-aware-mpi | 26 March 2024, 00:37:19 UTC |
434c3e7 | Peter Boyle | 25 March 2024, 14:32:44 UTC | We have a choice of GET or PUT across NVlink | 25 March 2024, 14:32:44 UTC |
500b119 | Peter Boyle | 22 March 2024, 15:55:23 UTC | Deterministic MPI | 22 March 2024, 15:55:23 UTC |
4b87259 | Peter Boyle | 22 March 2024, 15:43:49 UTC | New config command for sunspot | 22 March 2024, 15:43:49 UTC |
503dec3 | Peter Boyle | 22 March 2024, 15:43:30 UTC | This appears working now on Sunspot | 22 March 2024, 15:43:30 UTC |
d1e9fe5 | Peter Boyle | 22 March 2024, 15:42:57 UTC | Xor csum for repro testing | 22 March 2024, 15:42:57 UTC |
d01e5fa | Peter Boyle | 22 March 2024, 15:42:32 UTC | Improved FlightRecorder | 22 March 2024, 15:42:32 UTC |
a477c25 | Peter Boyle | 22 March 2024, 15:42:11 UTC | Sunspot repro tests | 22 March 2024, 15:42:11 UTC |
1bd20cd | Peter Boyle | 22 March 2024, 15:40:01 UTC | FlightRecorder | 22 March 2024, 15:40:01 UTC |
e49e95b | Peter Boyle | 22 March 2024, 15:39:27 UTC | Upgrade of the Britney test with flight recorder and fast xor checksum | 22 March 2024, 15:39:27 UTC |
6f59fed | Peter Boyle | 22 March 2024, 15:32:32 UTC | Flight recorder, resurrecting the "world famous" Britney test | 22 March 2024, 15:32:32 UTC |
60b7f6c | Peter Boyle | 22 March 2024, 15:32:26 UTC | Flight recorder, resurrecting the "world famous" Britney test | 22 March 2024, 15:32:26 UTC |
b92dfcc | Peter Boyle | 22 March 2024, 15:30:27 UTC | Flight recorder, resurrecting the "world famous" Britney test | 22 March 2024, 15:30:27 UTC |
f6fd6dd | Peter Boyle | 22 March 2024, 15:30:01 UTC | Flight recorder, resurrecting the "world famous" Britney test | 22 March 2024, 15:30:01 UTC |
79ad567 | Peter Boyle | 19 March 2024, 15:43:42 UTC | Merge branch 'develop' of https://github.com/paboyle/Grid into develop | 19 March 2024, 15:43:42 UTC |
fab1efb | Peter Boyle | 19 March 2024, 14:36:21 UTC | More britney logging improvements | 19 March 2024, 14:36:21 UTC |
660eb76 | Peter Boyle | 19 March 2024, 14:28:33 UTC | FFTW from OneAPI | 19 March 2024, 14:28:33 UTC |
62e7bf0 | Peter Boyle | 12 March 2024, 20:10:04 UTC | Updated flight logging for Britney test | 12 March 2024, 20:10:04 UTC |
95f3d69 | Peter Boyle | 12 March 2024, 20:09:37 UTC | Extra hardware test hook | 12 March 2024, 20:09:37 UTC |
89c0519 | Peter Boyle | 12 March 2024, 16:11:33 UTC | Repro test | 12 March 2024, 16:11:33 UTC |
2704b82 | Peter Boyle | 12 March 2024, 15:16:24 UTC | Merge branch 'develop' of https://github.com/paboyle/Grid into develop | 12 March 2024, 15:16:24 UTC |
cf8632b | Peter Boyle | 12 March 2024, 15:15:35 UTC | Britney test option | 12 March 2024, 15:15:35 UTC |
d224297 | Peter Boyle | 12 March 2024, 15:15:16 UTC | PBS scripts | 12 March 2024, 15:15:16 UTC |
a4d11a6 | Peter Boyle | 07 March 2024, 12:50:25 UTC | Merge pull request #458 from paboyle/fix/HOST_NAME_MAX fallback to _POSIX_HOST_NAME_MAX if HOST_NAME_MAX is not defined | 07 March 2024, 12:50:25 UTC |
2b4399f | Antonin Portelli | 07 March 2024, 06:26:01 UTC | more HOST_NAME_MAX fix | 07 March 2024, 06:26:01 UTC |
f17b8de | Antonin Portelli | 07 March 2024, 06:22:08 UTC | fallback to _POSIX_HOST_NAME_MAX if HOST_NAME_MAX is not defined | 07 March 2024, 06:22:08 UTC |
7e5bd46 | Peter Boyle | 06 March 2024, 18:03:45 UTC | Booster update | 06 March 2024, 18:03:45 UTC |
228bbb9 | Peter Boyle | 06 March 2024, 18:03:35 UTC | Benchmark results | 06 March 2024, 18:03:35 UTC |
b812a7b | Peter Boyle | 06 March 2024, 01:32:40 UTC | Staggered launch script | 06 March 2024, 01:32:40 UTC |
891a366 | Peter Boyle | 06 March 2024, 01:22:55 UTC | Repro CG script | 06 March 2024, 01:22:55 UTC |
10116b3 | Peter Boyle | 06 March 2024, 01:13:27 UTC | Force device copyable and tell SYCL to shut it. | 06 March 2024, 01:13:27 UTC |
a46a0f0 | Peter Boyle | 06 March 2024, 01:12:49 UTC | force device copyable and don't take crap from SYCL | 06 March 2024, 01:12:49 UTC |
a26a8a3 | Peter Boyle | 06 March 2024, 00:05:00 UTC | Merge branch 'develop' of https://github.com/paboyle/Grid into develop | 06 March 2024, 00:05:00 UTC |
7435315 | Peter Boyle | 06 March 2024, 00:03:59 UTC | More blasted shell variables | 06 March 2024, 00:03:59 UTC |
9b5f741 | Peter Boyle | 06 March 2024, 00:03:16 UTC | Reproducing CG can be more useful now | 06 March 2024, 00:03:16 UTC |
517822f | Peter Boyle | 06 March 2024, 00:02:27 UTC | SPR HBM benchmarking right and also PVC batched GEMM | 06 March 2024, 00:02:27 UTC |
1b93a9b | Peter Boyle | 06 March 2024, 00:01:58 UTC | Print out the hostname | 06 March 2024, 00:01:58 UTC |
783a66b | Peter Boyle | 06 March 2024, 00:01:37 UTC | Deterministic reduction please | 06 March 2024, 00:01:37 UTC |
976c3e9 | Peter Boyle | 05 March 2024, 23:59:57 UTC | Hack for flight logging CG inner products. Can be made to work, but could put in some more serious infrastructure for repro testing and blame attribution (Britney test) if necessary | 05 March 2024, 23:59:57 UTC |
f8ca971 | Peter Boyle | 05 March 2024, 23:59:13 UTC | Use of a bare PRECISION macro is not namespace safe and collides with SYCL | 05 March 2024, 23:59:13 UTC |
21bc8c2 | Peter Boyle | 05 March 2024, 23:58:20 UTC | OneMKL batched blas starting | 05 March 2024, 23:58:20 UTC |
3022821 | Peter Boyle | 05 March 2024, 23:56:10 UTC | SYCL conflict with Eigen | 05 March 2024, 23:56:10 UTC |
2ae980a | Peter Boyle | 05 March 2024, 18:39:18 UTC | Update sourceme.sh | 05 March 2024, 18:39:18 UTC |
6153dec | Peter Boyle | 05 March 2024, 18:38:32 UTC | Update setup.sh | 05 March 2024, 18:38:32 UTC |
c805f86 | Peter Boyle | 01 March 2024, 05:05:04 UTC | USQCD benchmark | 01 March 2024, 05:05:04 UTC |
04ca065 | Peter Boyle | 01 March 2024, 01:09:11 UTC | Only one rank opens | 01 March 2024, 01:09:11 UTC |
88d8fa4 | Peter Boyle | 01 March 2024, 01:01:44 UTC | Benchmark development | 01 March 2024, 01:01:44 UTC |
3c49762 | Peter Boyle | 29 February 2024, 20:33:06 UTC | Propagate in the blas routine | 29 February 2024, 20:33:06 UTC |
436bf1d | Peter Boyle | 29 February 2024, 20:29:39 UTC | Merge pull request #455 from clarkedavida/hisq_fat_links Hisq fat links | 29 February 2024, 20:29:39 UTC |
f70df6e | david clarke | 29 February 2024, 19:29:30 UTC | changed NO_SHIFT and BACKWARD_CONST from define to enum | 29 February 2024, 19:29:30 UTC |
fce3852 | Peter Boyle | 28 February 2024, 23:03:37 UTC | Merge pull request #451 from paboyle/feature/eigen-3.4.0-update updating Eigen to 3.4.0 | 28 February 2024, 23:03:37 UTC |
ee1b8bb | Peter Boyle | 28 February 2024, 19:05:27 UTC | Merge pull request #454 from edbennett/adjoint-broke fix HMC for non-fundamental representations | 28 February 2024, 19:05:27 UTC |
3f16366 | Peter Boyle | 28 February 2024, 19:04:43 UTC | Merge pull request #453 from dbollweg/feature/sliceSum_gpu Feature/slice sum gpu | 28 February 2024, 19:04:43 UTC |
2e570f5 | Peter Boyle | 28 February 2024, 18:59:04 UTC | Merge pull request #457 from lehner/feature/gpt Import GPT-related updates | 28 February 2024, 18:59:04 UTC |
9f89486 | Christoph Lehner | 28 February 2024, 18:56:23 UTC | remove unnecessary code path | 28 February 2024, 18:56:23 UTC |
22b43b8 | Christoph Lehner | 28 February 2024, 11:57:17 UTC | Make GPT test suite work with SYCL | 28 February 2024, 11:57:17 UTC |
3c90126 | dbollweg | 27 February 2024, 17:41:45 UTC | CUDA cub refuses to reduce vSpinColourMatrix, breaking up into smaller parts like already done for HIP case. | 27 February 2024, 17:41:45 UTC |
b507fe2 | Dennis Bollweg | 27 February 2024, 16:28:32 UTC | Added SpinColourMatrix case to sliceSum Test | 27 February 2024, 16:28:32 UTC |
6cd2d8f | Dennis Bollweg | 26 February 2024, 14:55:07 UTC | Replace cuda/hip memcpy with Grid functions | 26 February 2024, 14:55:07 UTC |
b02d022 | david clarke | 24 February 2024, 00:14:28 UTC | fixed race condition (thx michael) | 24 February 2024, 00:14:28 UTC |
94581e3 | david clarke | 23 February 2024, 22:58:33 UTC | accelerator_for is broken | 23 February 2024, 22:58:33 UTC |
88b52cc | david clarke | 23 February 2024, 21:47:15 UTC | Merge branch 'develop' into hisq_fat_links | 23 February 2024, 21:47:15 UTC |
0a816b5 | dbollweg | 23 February 2024, 02:43:06 UTC | Merge branch 'feature/sliceSum_gpu' of https://github.com/dbollweg/Grid into feature/sliceSum_gpu | 23 February 2024, 02:43:06 UTC |
1c8b807 | dbollweg | 23 February 2024, 02:42:44 UTC | free malloc'd memory | 23 February 2024, 02:42:44 UTC |
66391f8 | Christoph Lehner | 21 February 2024, 18:05:00 UTC | Merge branch 'feature/gpt' of ../Grid into develop | 21 February 2024, 18:05:00 UTC |
97f7a9e | Ed Bennett | 21 February 2024, 08:27:55 UTC | fix HMC for non-fundamental representations | 21 February 2024, 08:27:55 UTC |
15878f7 | Dennis Bollweg | 16 February 2024, 18:55:21 UTC | sliceSumReduction_cub_large now also faster than CPU on Frontier | 16 February 2024, 18:55:21 UTC |
e0d5e3c | dbollweg | 16 February 2024, 18:16:37 UTC | Merge branch 'paboyle:develop' into feature/sliceSum_gpu | 16 February 2024, 18:16:37 UTC |
6f34559 | dbollweg | 16 February 2024, 18:15:02 UTC | Adding sliceSumReduction_cub_small/large since hipcub cannot deal with arb. large vobjs | 16 February 2024, 18:15:02 UTC |
56827d6 | david clarke | 14 February 2024, 20:56:57 UTC | accelerator_inline bug | 14 February 2024, 20:56:57 UTC |
73c0b29 | Peter Boyle | 13 February 2024, 20:19:32 UTC | Merge branch 'develop' of https://github.com/paboyle/Grid into develop | 13 February 2024, 20:19:32 UTC |
303b83c | Peter Boyle | 13 February 2024, 19:48:03 UTC | Scaling benchmarks, verbosity and MPICH aware in acceleratorInit() For some reason Dirichlet benchmark fails on several nodes; need to debug this. | 13 February 2024, 19:48:03 UTC |
5ef4da3 | Peter Boyle | 13 February 2024, 19:47:36 UTC | Silence verbose | 13 February 2024, 19:47:36 UTC |
1502860 | Peter Boyle | 13 February 2024, 19:47:02 UTC | Benchmark scripts | 13 February 2024, 19:47:02 UTC |
585efc6 | Peter Boyle | 13 February 2024, 19:40:49 UTC | More benchmark scripts | 13 February 2024, 19:40:49 UTC |
62055e0 | Antonin Portelli | 13 February 2024, 17:18:27 UTC | missing semicolon generates error with some compilers | 13 February 2024, 17:18:27 UTC |
e4a641b | Antonin Portelli | 13 February 2024, 09:37:14 UTC | removing old Eigen tensor patch | 13 February 2024, 09:37:14 UTC |
8849f18 | Antonin Portelli | 13 February 2024, 09:30:22 UTC | updating Eigen to 3.4.0 | 13 February 2024, 09:30:22 UTC |
db42052 | david clarke | 12 February 2024, 22:03:53 UTC | fix Simd::Nsimd typo | 12 February 2024, 22:03:53 UTC |
b5659d1 | dbollweg | 09 February 2024, 18:37:14 UTC | more test cases | 09 February 2024, 18:37:14 UTC |
4b43307 | dbollweg | 09 February 2024, 18:07:56 UTC | Undo include path changes for level zero api header | 09 February 2024, 18:07:56 UTC |
09af8c2 | dbollweg | 09 February 2024, 18:02:59 UTC | Merge branch 'paboyle:develop' into feature/sliceSum_gpu | 09 February 2024, 18:02:59 UTC |
9514035 | dbollweg | 09 February 2024, 18:02:28 UTC | refactor slicesum: slicesum uses GPU version by default now | 09 February 2024, 18:02:28 UTC |
2da09ae | david clarke | 07 February 2024, 01:40:13 UTC | acceleration compiles and doesn't break scalar mode | 07 February 2024, 01:40:13 UTC |
a38fb0e | david clarke | 07 February 2024, 01:24:55 UTC | first effort toward accelerators | 07 February 2024, 01:24:55 UTC |
7019916 | Peter Boyle | 07 February 2024, 00:56:39 UTC | RNG seed change safer for large volumes; this is a long term solution | 07 February 2024, 00:56:39 UTC |
1514b4f | dbollweg | 07 February 2024, 00:08:44 UTC | slicesum_sycl passes test | 07 February 2024, 00:08:44 UTC |
91cf5ee | Peter Boyle | 06 February 2024, 23:45:10 UTC | Updated bench script | 06 February 2024, 23:45:10 UTC |
0a6e2f4 | david clarke | 06 February 2024, 23:32:07 UTC | small amount of cleanup | 06 February 2024, 23:32:07 UTC |
ab2de13 | dbollweg | 06 February 2024, 18:24:45 UTC | work towards sliceSum for sycl backend | 06 February 2024, 18:24:45 UTC |
5bfa88b | Peter Boyle | 06 February 2024, 16:28:40 UTC | Aurora MPI standalone benchmake and options that work well | 06 February 2024, 16:28:40 UTC |