portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2026-07-17 15:43:27 +01:00
Code Releases Activity
3,244 Commits 95 Branches 14 Tags
a9e8758a01aa7866699496e1660707dc04d1e7b1
Commit Graph
100 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Azusa Yamaguchi a9e8758a01 Improvements to staggered tests timings 2018-04-26 10:08:05 +01:00
Azusa Yamaguchi 3e125c5b61 Faster linalg on CG optimised against staggered 
Sum overhead is bigger for staggered
 2018-04-26 10:07:19 +01:00
Azusa Yamaguchi eac6ec4b5e Faster reductions, important on single node staggered 2018-04-26 10:03:57 +01:00
Azusa Yamaguchi 213f8db6a2 Microsecond resultion 2018-04-26 10:01:39 +01:00
Azusa Yamaguchi b938202081 Overlapped Comm for Wilson DhopInternal 2018-03-07 14:08:43 +00:00
Azusa Yamaguchi 0f468e2179 OverlappedComm for Staggered 5D and 4D. 2018-02-22 12:50:09 +00:00
Azusa Yamaguchi 97b9c6f03d No option for interior/exterior split of asm kernels since different directions get interleaved 2018-01-22 11:04:19 +00:00
Azusa Yamaguchi 63982819c6 No option to overlap comms and compute for asm implementation since different directions are interleaved 
in the kernels, introducing if else structure would be too painful
 2018-01-22 11:03:39 +00:00
Azusa Yamaguchi 24162c9ead Staggered overlap comms comput 2018-01-09 13:02:52 +00:00
azusayamaguchi e55397bc13 Staggerd cg 2017-11-24 14:18:30 +00:00
Azusa Yamaguchi 1860b1698c Fixed the bag on MPI_T at Cam 2017-11-08 09:03:01 +00:00
Azusa Yamaguchi 9b8d1cc3da Staggered Schur decomposed matrix norm changed to not be the Schur anymore :( 
Carleton wanted this for multimass / multishift
 2017-11-07 14:48:45 +00:00
Azusa Yamaguchi 034de160bf Staggered updates : Schur fixed and added a unit test for Test_staggered_cg_schur.cc giving stronger check 2017-10-26 20:58:46 +01:00
Azusa Yamaguchi f9e28577f3 Merge branch 'develop' of https://github.com/paboyle/Grid into develop 2017-10-25 21:07:56 +01:00
Azusa Yamaguchi cac1750078 Merge branch 'develop' of https://github.com/paboyle/Grid into develop 2017-10-24 23:30:36 +01:00
Azusa Yamaguchi f1fa00b71b Merge branch 'develop' of https://github.com/paboyle/Grid into develop 2017-10-10 14:26:44 +01:00
Azusa Yamaguchi 1374c943d4 Correct Schur operator called 2017-10-10 13:59:50 +01:00
Azusa Yamaguchi dd949bc428 Merge branch 'feature/staggering' into develop 2017-10-10 13:02:51 +01:00
Azusa Yamaguchi bb7378cfc3 Schur for staggered 2017-10-10 12:02:18 +01:00
Azusa Yamaguchi f0e084a88c Schur staggered 2017-10-10 10:00:43 +01:00
Azusa Yamaguchi 09f4cdb11e Merge branch 'develop' of https://github.com/paboyle/Grid into feature/staggering 2017-10-04 10:51:16 +01:00
Azusa Yamaguchi 1e54882f71 Stagger 2017-10-04 10:51:06 +01:00
Azusa Yamaguchi eb6153080a Merge branch 'develop' of https://github.com/paboyle/Grid into feature/staggering 2017-10-02 08:56:33 +01:00
Azusa Yamaguchi a6eeea777b Merge branch 'develop' of https://github.com/paboyle/Grid into develop 2017-09-21 10:12:41 +01:00
Azusa Yamaguchi 77f7737ccc Merge branch 'develop' of https://github.com/paboyle/Grid into develop 2017-09-19 14:28:01 +01:00
Azusa Yamaguchi a5fe07c077 Merge branch 'develop' of https://github.com/paboyle/Grid into develop 2017-09-04 14:10:15 +01:00
Azusa Yamaguchi b83b2b1415 Stability improvement to BCG. Force m_rr hermitian beyond rounding. 2017-09-04 14:09:47 +01:00
Azusa Yamaguchi d9cd4f0273 Staggered multinode block cg debugged. Missing global sum. 
Code stalls and resumes on KNL at cambridge. Curious.

CG iterations 23ms each, then 3200 ms pauses. Mean bandwidth reports
as 200MB/s. Comms dominant in the report. However, the time behaviour suggests it
is *bursty*.... Could be swap to disk?
 2017-08-23 15:07:18 +01:00
azusayamaguchi 659d7d1a40 For test/solver 
Fixed
 2017-07-12 15:01:48 +01:00
azusayamaguchi dc6f078246 fixed the header file for mpi3 2017-07-11 14:15:08 +01:00
Azusa Yamaguchi e9cc21900f Block solver complete for staggered. Now stable on mass 0.003 and 
gives 8x (!) speed up on Haswell laptop vs. standard CG for 8 RHS solves.

166 iterations vs. 537 iterations so algorithmic gain + 2x in flop rate gain.

Better than a slap in the face with a wet kipper.
 2017-06-20 12:37:41 +01:00
Azusa Yamaguchi 0a8faac271 Fix make tests compile 2017-06-19 22:54:18 +01:00
Azusa Yamaguchi abc4de0fd2 No compile make tests fix 2017-06-19 22:03:03 +01:00
Azusa Yamaguchi cfe3cd76d1 Block solver improvements 2017-06-19 14:04:21 +01:00
Azusa Yamaguchi 3fa5e3109f Merge branch 'develop' of https://github.com/paboyle/Grid into develop 2017-06-19 14:01:44 +01:00
Azusa Yamaguchi 70ab598c96 Move gfix into utils 2017-06-08 22:22:23 +01:00
Azusa Yamaguchi 1d0ca65e28 Move Gfix into utils 2017-06-08 22:21:50 +01:00
Azusa Yamaguchi 2bc4d0a20e Move code into utils 2017-06-08 22:21:25 +01:00
Azusa Yamaguchi f46a67ffb3 No compile issue on clang on mac fixed. 
Compiler version was clang++-3.9 under mpicxx
 2017-05-17 10:51:01 +01:00
azusayamaguchi 1c30e9a961 Verified 2017-02-21 23:01:25 +00:00
azusayamaguchi bf7e3f20d4 Staggaered fermion optimised version 2017-02-21 14:35:42 +00:00
azusayamaguchi 05c1924819 Timing loop change 2017-01-23 10:43:45 +00:00
azusayamaguchi df9108154d Debugged 2 versions of assembler; ls vectorised, xyzt vectorised 2016-12-17 23:47:51 +00:00
azusayamaguchi b3e7f600da Partial implementation of 4d vectorisation assembler 2016-12-16 23:50:30 +00:00
azusayamaguchi d4071daf2a Template specialise 2016-12-16 22:28:29 +00:00
azusayamaguchi a2a6329094 AVX512 only for ASM compilation 2016-12-16 22:03:29 +00:00
azusayamaguchi eabc577940 Assembler possibly working 2016-12-16 16:55:36 +00:00
Azusa Yamaguchi 0cd6b1858c Merge branch 'develop' of https://github.com/paboyle/Grid into feature/staggering 2016-12-14 09:23:22 +00:00
Azusa Yamaguchi 426197e446 Nc=3 2016-12-12 09:10:54 +00:00
Azusa Yamaguchi 99e2c1e666 Kernels options 2016-12-12 09:08:53 +00:00
Azusa Yamaguchi 1440565a10 Decrease verbosity 2016-12-12 09:08:04 +00:00
Azusa Yamaguchi e9f0c0ea39 Staggered kernels options 2016-12-12 09:07:38 +00:00
Azusa Yamaguchi c097fd041a Merge branch 'develop' of https://github.com/paboyle/Grid into feature/staggering 2016-11-29 13:44:17 +00:00
Azusa Yamaguchi 77fb25fb29 Push 5d tests 2016-11-29 13:43:56 +00:00
Azusa Yamaguchi 389e0a77bd Staggerd Fermion 5D 2016-11-29 13:13:56 +00:00
Azusa Yamaguchi 95f43d27ae Merge branch 'develop' of https://github.com/paboyle/Grid into feature/staggering 2016-11-22 13:49:22 +00:00
Azusa Yamaguchi 668ca57702 Merge branch 'develop' of https://github.com/paboyle/Grid into feature/staggering 2016-11-22 13:49:11 +00:00
azusayamaguchi 50d277d8d9 Merge branch 'develop' into release/v0.6.0 2016-11-08 13:44:14 +00:00
azusayamaguchi 343f3e829f Fixes prerelease to make all tests 2016-11-08 13:42:12 +00:00
azusayamaguchi f85b35314d Fix a routine for single node processor coor from rank 2016-11-08 11:49:13 +00:00
azusayamaguchi 3dc2e05d6e Time as well since MKL returns zero for Mflops 2016-11-08 11:36:18 +00:00
azusayamaguchi 0cff8754d1 Usecs 2016-11-08 11:35:41 +00:00
azusayamaguchi 692b44dac1 Merge branch 'develop' into release/v0.6.0 2016-11-04 22:48:11 +00:00
azusayamaguchi 96ba42a297 omm buf 2016-11-04 22:47:25 +00:00
azusayamaguchi f7b60004f3 Merge branch 'develop' into release/v0.6.0 2016-11-04 16:08:07 +00:00
azusayamaguchi b7d55f7dfb Fix a typo in reorg of the --dslash-asm 2016-11-04 11:35:08 +00:00
azusayamaguchi 6e548a8ad5 Linux compile needed 2016-11-04 11:34:16 +00:00
Azusa Yamaguchi ee686a7d85 Compiles now 2016-11-03 16:58:23 +00:00
Azusa Yamaguchi 1c5b7a6be5 Staggered phases first cut, c1, c2, u0 2016-11-03 16:26:56 +00:00
Azusa Yamaguchi 164d3691db Staggered 2016-11-01 14:24:22 +00:00
Azusa Yamaguchi bca861e112 Note:FFT shoud be GridFFT (Not change yet). 
Gauge fix with FFt is added (tests/core)
 2016-10-25 14:21:48 +01:00
azusayamaguchi d7d92af09d Travis fail fix attempt 2016-10-25 01:45:53 +01:00
azusayamaguchi 460d0753a1 Merge branch 'develop' into feature/mpi3 
Conflicts:
	lib/simd/Grid_avx512.h
 2016-10-25 01:08:51 +01:00
azusayamaguchi 8f8058f8a5 More random bits on parallel seeding 2016-10-25 01:05:52 +01:00
azusayamaguchi d97a27f483 Verbose 2016-10-25 01:05:31 +01:00
azusayamaguchi 7c3363b91e Compiles all comms targets 2016-10-25 00:04:17 +01:00
azusayamaguchi b94478fa51 mpi, mpi3, shmem all compile. 
mpi, mpi3 pass single node multi-rank
 2016-10-24 23:45:31 +01:00
azusayamaguchi b6a65059a2 Update to use shared memory to contain the stencil comms buffers 
Tested on 2.1.1.1 1.2.1.1 4.1.1.1 1.4.1.1 2.2.1.1 subnode decompositions
 2016-10-24 17:30:43 +01:00
azusayamaguchi ea25a4d9ac Works 2016-10-23 06:10:05 +01:00
azusayamaguchi c190221fd3 Internal SHM comms in non-simd directions working 
Need to fix simd directions
 2016-10-22 18:14:27 +01:00
azusayamaguchi 0fcd2e7188 Simplify the comms structure prior to implementing Shared memory direct bouncs 2016-10-21 22:44:10 +01:00
azusayamaguchi 910b8dd6a1 use simd type 2016-10-21 22:35:29 +01:00
azusayamaguchi 75ebd3a0d1 Typo fixes and rotate for CLANG 2016-10-21 22:34:29 +01:00
azusayamaguchi 09fd5c43a7 Reasonably fast version 2016-10-21 15:17:39 +01:00
azusayamaguchi f22317748f Merge branch 'feature/mpi3' of https://github.com/paboyle/Grid into feature/mpi3 2016-10-21 13:36:35 +01:00
azusayamaguchi 6a9eae6b6b Reporting improvements 2016-10-21 13:36:18 +01:00
azusayamaguchi fad96cf250 StencilBufs 2016-10-21 13:36:00 +01:00
azusayamaguchi f331809c27 Use variable type for loop 2016-10-21 13:35:37 +01:00
azusayamaguchi 20a091c3ed Intel vs. Clang intrinsics differences absorbed 2016-10-21 09:08:36 +01:00
azusayamaguchi 202078eb1b Cray / OpenSHMEM ordering differs 2016-10-21 09:07:20 +01:00
azusayamaguchi 8b0d171c9a 32bit issue on the KNL code variant where byte offsets were stored 2016-10-12 17:49:32 +01:00
azusayamaguchi 1f293b76b4 Merge branch 'feature/knl-stats' into develop 2016-10-12 13:47:58 +01:00
azusayamaguchi 8bbd9ebc27 Reversing changes to Stencil class 2016-10-12 13:47:20 +01:00
azusayamaguchi 6472b431f0 __rdpmc needed for gcc, clang++ 2016-10-12 12:29:08 +01:00
azusayamaguchi bd205a3293 Fixing for non x86 and non KNL 2016-10-12 12:09:15 +01:00
azusayamaguchi 496beffa88 Fix non-KNL build 2016-10-12 12:06:08 +01:00
azusayamaguchi 9b63e97108 align not absolutely required and confuses clang++ 2016-10-12 11:51:21 +01:00
azusayamaguchi 81f2aeaece KNL streaming stores, and KNL performance coutners 2016-10-12 11:45:22 +01:00
Azusa Yamaguchi d9408893b3 Prefetching in the normal kernel implementation. 2016-06-08 05:43:48 -07:00
azusa f54e0ec9bd Try lanczos to set up hdcr subspace 2016-03-17 10:36:16 +00:00