 paboyle
|
41009cc142
|
Move excange into the stencil only; keep Cshift fully general
|
2017-02-20 17:48:04 -05:00 |
|
|
paboyle
|
37720c4db7
|
Count bytes off node only
|
2017-02-20 17:47:40 -05:00 |
|
|
paboyle
|
1a30455a10
|
1000 iters on bmark for more accurate timing
|
2017-02-20 17:47:01 -05:00 |
|
|
paboyle
|
cd0da81196
|
Merge branch 'feature/bgq-asm' of https://github.com/paboyle/Grid into feature/bgq-asm
|
2017-02-16 18:52:30 -05:00 |
|
|
paboyle
|
f246fe3304
|
Improvements to avx for invertible to avoid latent bug
|
2017-02-16 23:52:44 +00:00 |
|
|
paboyle
|
8a29c16bde
|
Faster gather exchange
|
2017-02-16 23:52:22 +00:00 |
|
|
paboyle
|
d68907fc3e
|
Debug temp
|
2017-02-16 18:51:35 -05:00 |
|
|
paboyle
|
5c0adf7bf2
|
Make clang happy with parenthesis
|
2017-02-16 23:51:33 +00:00 |
|
|
paboyle
|
be3a8249c6
|
Faster gather
|
2017-02-16 23:51:15 +00:00 |
|
|
paboyle
|
bd600702cf
|
Vectorise the XYZT face gathering better.
Hard coded for simd_layout <= 2 in any given spread out direction; full generality is inconsistent
with efficiency.
|
2017-02-15 11:11:04 +00:00 |
|
|
paboyle
|
aca7a3ef0a
|
Optimisation control improvements
|
2017-02-10 18:22:31 -05:00 |
|
|
paboyle
|
2c246551d0
|
Overlap comms and compute options in wilson kernels
|
2017-02-07 01:37:10 -05:00 |
|
|
paboyle
|
71ac2e7940
|
Faster RNG init
|
2017-02-07 01:33:23 -05:00 |
|
|
paboyle
|
2bf4688e83
|
Running on BNL KNL
|
2017-02-07 01:32:10 -05:00 |
|
|
paboyle
|
a48ee6f0f2
|
Don't use MPI3_leader any more. No real gain and complex
|
2017-02-07 01:31:24 -05:00 |
|
|
paboyle
|
73547cca66
|
MPI3 working i think
|
2017-02-07 01:30:02 -05:00 |
|
|
paboyle
|
123c673db7
|
Policy to control async or sync SendRecv
|
2017-02-07 01:24:54 -05:00 |
|
|
paboyle
|
61f82216e2
|
Communicator Policy, NodeCount distinct from Rank count
|
2017-02-07 01:22:53 -05:00 |
|
|
paboyle
|
8e7ca92278
|
Debugged cshift case
|
2017-02-07 01:21:32 -05:00 |
|
|
paboyle
|
485ad6fde0
|
Stencil working in SHM MPI3
|
2017-02-07 01:20:39 -05:00 |
|
|
paboyle
|
6ea2184e18
|
OMP define change
|
2017-02-07 01:17:16 -05:00 |
|
|
paboyle
|
fdc170b8a3
|
Parallel fors in lattice transfer
|
2017-02-07 01:16:39 -05:00 |
|
|
paboyle
|
060da786e9
|
Comms benchmark improvements
|
2017-02-07 01:07:39 -05:00 |
|
|
paboyle
|
85c7bc4321
|
Bug fixes for cases that physics code couldn't hit but latent
and discovered on KNL (long vector, y SIMD dir) and checker dir set to y.
Remove the assertions on these code paths now they are tested.
|
2017-02-07 01:01:15 -05:00 |
|
|
paboyle
|
0883d6a7ce
|
Overlap comms compute support; make reg naming consistent with bgq aasm
|
2017-02-07 00:59:32 -05:00 |
|
|
paboyle
|
9ff97b4711
|
Improved stencil tests passing all on KNL multinode
|
2017-02-07 00:58:34 -05:00 |
|
|
paboyle
|
b5e9c900a4
|
Better printing and signal handling options
|
2017-02-07 00:57:55 -05:00 |
|
|
paboyle
|
4bbdfb434c
|
Overlap comms compute modifications
|
2017-02-07 00:57:01 -05:00 |
|
|
Peter Boyle
|
c3b6d573b9
|
Merge branch 'feature/bgq-asm' of https://github.com/paboyle/Grid into feature/bgq-asm
|
2016-12-30 22:42:17 +00:00 |
|
|
Peter Boyle
|
1e179c903d
|
Worried about integer; suspect where statements are broken
|
2016-12-27 17:46:38 +00:00 |
|
|
Peter Boyle
|
669cfca9b7
|
No inline
|
2016-12-27 17:45:40 +00:00 |
|
|
Peter Boyle
|
ff2f559a57
|
Remove inline on gather optimised path
|
2016-12-27 17:45:19 +00:00 |
|
|
Peter Boyle
|
03c81bd902
|
Merge branch 'feature/bgq-asm' of https://github.com/paboyle/Grid into feature/bgq-asm
|
2016-12-27 11:25:35 +00:00 |
|
|
Peter Boyle
|
a869addef1
|
Stats switch off
|
2016-12-27 11:25:22 +00:00 |
|
|
Peter Boyle
|
1caa3fbc2d
|
LOCK UNLOCK only
|
2016-12-27 11:24:45 +00:00 |
|
|
Peter Boyle
|
3d21297bbb
|
Call the fast path compressor for wilson kernels to avoid if else on projector
|
2016-12-27 11:23:13 +00:00 |
|
|
Peter Boyle
|
25efefc5b4
|
Back to original thread policy post test
|
2016-12-23 09:49:04 +00:00 |
|
|
Peter Boyle
|
eabf316ed9
|
BGQ performance ASM
|
2016-12-22 21:56:08 +00:00 |
|
|
Peter Boyle
|
04ae7929a3
|
BGQ or KNL assembler now
|
2016-12-22 17:53:22 +00:00 |
|
|
Peter Boyle
|
caba0d42a5
|
L1p controls
|
2016-12-22 17:52:55 +00:00 |
|
|
Peter Boyle
|
9ae81c06d2
|
L1p controls for BG/Q
|
2016-12-22 17:52:21 +00:00 |
|
|
Peter Boyle
|
0903c48caa
|
Hot start SU3
|
2016-12-22 17:51:45 +00:00 |
|
|
Peter Boyle
|
7dc36628a1
|
QPX finishing
|
2016-12-22 17:50:48 +00:00 |
|
|
Peter Boyle
|
b8cdb3e90a
|
Debug hack; raises from 62GF/s to 72 GF/s per node on BG/Q
|
2016-12-22 17:50:14 +00:00 |
|
|
Peter Boyle
|
5241245534
|
Default to static scheduling
|
2016-12-22 17:49:21 +00:00 |
|
 Dr Peter Boyle
|
960316e207
|
type conversion in printf
|
2016-12-22 17:27:01 +00:00 |
|
|
paboyle
|
3f2d53a994
|
BGQ assembler beginning
|
2016-12-20 10:21:26 +00:00 |
|
|
paboyle
|
8a337f3070
|
Move cayley into mainstream tests
|
2016-12-18 02:35:31 +00:00 |
|
|
paboyle
|
a59f5374d7
|
Evade warning
|
2016-12-18 02:23:55 +00:00 |
|
|
paboyle
|
4b220972ac
|
Warning fix
|
2016-12-18 02:14:17 +00:00 |
|