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
Christopher Kelly
9939b267d2
Added switching to fallback linear operator in reliable update CG, and added recalculation of b parameter on update.
2017-07-31 13:39:44 -04:00
Christopher Kelly
9f280b82c4
Added mixed-precision CG with reliable updates
2017-07-25 11:30:41 -04:00
paboyle
e504260f3d
Able to run a test job splitting into multiple MPI subdomains.
2017-06-22 18:53:11 +01:00
paboyle
e8b95bd35b
Clean up finished. Could shrink Lanczos to around 400 lines at a push
2017-06-21 02:50:09 +01:00
paboyle
7e35286860
Simplified lanczos, added Eigen diagonalisation.
...
Curious if we can deprecate dependencly on BLAS.
Will see when we get 48^3 running on our BG/Q port
2017-06-21 02:26:03 +01:00
paboyle
0486ff8e79
Improved the lancos
2017-06-20 18:46:01 +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
cfe3cd76d1
Block solver improvements
2017-06-19 14:04:21 +01:00
paboyle
2439999ec8
Warning elimination; drop to -O2 on G++ bad versions
2017-05-06 14:44:49 +01:00
Guido Cossu
3344788fa1
Merge branch 'develop' into feature/hmc_generalise
2017-05-01 12:13:56 +01:00
paboyle
8e161152e4
MultiRHS solver improvements with slice operations moved into lattice and sped up.
...
Block solver requires a lot of performance work.
2017-04-18 10:51:55 +01:00
paboyle
3141ebac10
MultiRHS working, starting to optimise. Block doesn't and I thought it already was; puzzled.
2017-04-17 10:50:19 +01:00
paboyle
7ede696126
Non compile of tests fixed
2017-04-16 23:40:00 +01:00
paboyle
a8db024c92
Cleaning up the dense matrix and lanczos sector
2017-04-15 08:54:11 +01:00
paboyle
b12dc89d26
Commenting and clean up
2017-04-10 20:38:20 +09:00
paboyle
d80d802f9d
MultiRHS solver test
2017-04-10 00:12:12 +09:00
paboyle
3d99b09dba
Start of blockCG
2017-04-09 23:42:10 +09:00
Guido Cossu
e0571c872b
Merge branch 'develop' into feature/hmc_generalise
2017-02-09 16:12:00 +00:00
Christopher Kelly
c94133af49
Added iteration reporting to CG and mixed CG
...
Added ability to manually change the initial CG inner tolerance in mixed CG
Added .hpp files to filelist script
2017-02-02 17:04:42 -05:00
fad743fbb1
Build system sanity check: corrected several headers not in the <Grid/*> format
2017-01-26 17:00:41 -08:00
Guido Cossu
0bd296dda4
Adding check of the Dag part in the benchmark
2016-12-14 03:15:09 +00:00
Guido Cossu
5c74b6028b
Commit for debugging, lot of IO
2016-12-13 06:35:30 +00:00
97cddda49e
Merge branch 'feature/gen-simd' into feature/doxygen
...
# Conflicts:
# Makefile.am
# configure.ac
2016-11-19 13:11:13 +01:00
paboyle
33dc1f51b5
Final sign off commits from Cori-1
2016-11-09 04:11:03 -08:00
e74417ca12
big build system polish
2016-10-31 16:31:27 +00:00
cb02b7088f
Merge branch 'develop' into feature/doxygen
...
# Conflicts:
# configure.ac
2016-10-09 13:35:44 +01:00
Guido Cossu
2e453dfbf5
Added some instrumentation to benchmark the force computation
2016-10-06 17:52:45 +01:00
b1cfb4d661
first try at a nicer Doxygen implementation
2016-08-05 15:29:18 +01:00
paboyle
f4dd5062d7
Merge branch 'develop' of https://github.com/paboyle/Grid into develop
2016-07-15 19:26:06 +01:00
paboyle
a0676beeb1
Open up dependency on Eigen and FFTW
2016-07-07 22:31:07 +01:00
Christopher Kelly
c5106d0c03
Bugfix
2016-07-07 16:06:30 -04:00
Christopher Kelly
25fafa9a89
Comment
2016-07-06 16:19:41 -04:00
Christopher Kelly
85ed8175cb
Implemented mixed precision CG. Fixed filelist to exclude lib/Old directory and include Config.h.
2016-07-06 15:57:04 -04:00
Christopher Kelly
15f22425c8
Added option to prevent CG from exiting when it fails to converge
2016-07-06 14:50:01 -04:00
fa59789580
ConjugateGradient: cleaner output
2016-05-01 18:29:20 -07:00
92c2c7d3b5
SchurRedBlackDiagMooeeSolve: fix: guess was not initialised from input
2016-05-01 16:07:55 -07:00
Christopher Kelly
af9c8d1372
-Checkerboard fixes for Lanczos
2016-04-06 13:50:56 -04:00
paboyle
340428a1fe
Eigen fixes and HDCR work
2016-03-30 00:16:02 -07:00
paboyle
3920b2c0ab
HDCR updates
2016-03-16 02:29:58 -07:00
paboyle
090e7aa930
Merge remote-tracking branch 'origin/chulwoo-dec12-2015'
...
Merge Chulwoo's Lanczos related improvements.
Merge Nd!=4 fixes for pure gauge HMC from Evan.
2016-03-08 09:55:14 +00:00
paboyle
2dce9c3cff
HDCR running on 16^3 with 2x-3x speed up.
2016-03-08 01:01:50 -08:00
paboyle
dc72293398
More timing info
2016-03-06 10:46:55 -08:00
paboyle
ff129d9ad9
Redundant operations removed
2016-03-02 07:02:37 -08:00
Jung
9f0d9ade68
Added configure flag for LAPACK. Tested ImplicitlyRestartedLanczos::calc()
...
Checking in before cleaning up
2016-02-20 02:50:32 -05:00
Jung
bd84c23298
definitions reconciled.
2016-01-25 16:30:59 -05:00
Jung
b8fb05a422
Addtional routines for Lanczos (SYM2, Chebyshef)..
2016-01-25 01:26:25 -05:00
Jung
5c57d4f403
Merge branch 'master' of https://github.com/paboyle/Grid into scidac1_2
...
Conflicts:
lib/qcd/action/fermion/WilsonKernels.h
2016-01-11 11:36:45 -05:00
Jung
5924e5a562
Merge branch 'master' of https://github.com/paboyle/Grid into scidac1_2
...
Conflicts:
configure
lib/qcd/action/Actions.h
lib/qcd/action/fermion/WilsonKernels.h
2016-01-06 03:44:57 -05:00
paboyle
aae8bf31a7
Global edit adding copyright and license info to every source file.
2016-01-02 14:51:32 +00:00
paboyle
841a37f941
Fix to WilsonCompressor that fixes a bug in comms phase due to the sign change on gamma
...
matrix in hopping term.
Add logging of time spent in CG.
2015-12-29 23:49:41 +00:00
paboyle
08edbb5cbe
HMC bit repro across checkpoints. Fixed parallel RNG issue with threading.
...
Conclusion: c++11 distributions not thread safe and must us distinct dist as well as distinct engine
per site. Makes sense when you think of box muller. Also added a reset of dist on fill to ensure
repro across checkpoints.
2015-12-22 08:54:40 +00:00
paboyle
8c9010d0f4
Isnan check on guess and convergence assert on result
2015-12-21 15:29:46 +00:00
paboyle
93356fd246
No compile fixes on gcc/Cray
2015-11-29 03:14:44 -08:00
Peter Boyle
11cf0f08f3
This file is not yet debugged.
2015-11-29 00:32:45 +00:00
Peter Boyle
3b7576ad53
Switch off for now
2015-11-04 05:13:29 -06:00
Peter Boyle
2d95dac6b6
Lanczos untested/partially tested additions. In middle of shake out but at least compiles
2015-10-09 00:40:25 +02:00
Peter Boyle
5ef42add2d
Changes to remove warnings under icc; disambiguate AVX512 from IMCI correctly
...
and drop swizzles in AVX512. Don't know why these compiled.
2015-09-23 05:23:45 -07:00
Peter Boyle
fc160eeccc
Added one flavour rational ratios (unprec)
2015-08-19 04:58:40 +01:00
Peter Boyle
69ce87fbe4
Preconditioned conjugate residual
2015-08-11 06:24:53 +01:00
Peter Boyle
4fe110bd07
Two flavour HMC for Wilson/Wilson is conserving energy.
...
Still to check plaq and <e(-dH)>, but nevertheless this is
progress
2015-07-29 17:53:39 +09:00
Peter Boyle
d1afebf71e
Sizable improvement in multigrid for unsquared.
...
6000 matmuls CG unprec
2000 matmuls CG prec (4000 eo muls)
1050 matmuls PGCR on 16^3 x 32 x 8 m=.01
Substantial effort on timing and logging infrastructure
2015-07-24 01:31:13 +09:00
Peter Boyle
0007669381
Printing change
2015-07-21 13:51:56 +09:00
Peter Boyle
8a7b7f1e2b
verbose
2015-07-21 12:12:29 +09:00
Peter Boyle
8eaf657f95
VPGCR updates
2015-06-30 15:02:27 +01:00
Peter Boyle
a17684ebe2
Some small steps towards a multigrid
2015-06-22 12:49:44 +01:00
Azusa Yamaguchi
3b4118f33e
Variable preconditioned GCR with restarting.
...
Orthogonalisation depth and restart frequency is controllable via constructor
2015-06-21 10:58:46 +01:00
Peter Boyle
960f29c0b1
Will start this as a two level algorithm
2015-06-20 22:24:21 +01:00
Peter Boyle
fb07ee5781
HDCG but this is not complete and placeholder for later completion
2015-06-20 22:23:57 +01:00
Peter Boyle
b4a6dbfa65
Patches for beginnings of an overlap multigrid
2015-06-20 22:22:56 +01:00
Peter Boyle
3111f50f2f
Merge branch 'master' of https://github.com/paboyle/Grid
...
Conflicts:
lib/Make.inc
tests/Make.inc
tests/Test_remez.cc
2015-06-08 12:08:09 +01:00
Peter Boyle
d6f1ddf99c
Conjugate residual algorithm; some more unary functions
2015-06-08 12:04:59 +01:00
Azusa Yamaguchi
8688ff8b3a
multishift conjugate gradient added and a strong test: take a diagonal
...
but non-identity matrix
l1 0 0 0 ....
0 l2 0 0 ....
0 0 l3 0 ...
. . .
. . .
. . .
And apply the multishift CG to it. Sum the poles and residues.
Insist that this be the same as the exactly taken square root
where l1,l2,l3 >= 0.
2015-06-08 11:52:44 +01:00
Peter Boyle
1a05882d7c
Conjugate residual added
2015-06-05 18:16:25 +01:00
Peter Boyle
7f6304fac3
Rework the linop support to get different forms of red black schur solver
...
Moo on diag, or MooInv Moe MeeInv Meo
2015-06-05 10:17:10 +01:00
Peter Boyle
b1b412f63c
comment improvement
2015-06-05 05:31:27 +01:00
Peter Boyle
5644ab1e19
Large scale change to support 5d fermion formulations.
...
Have 5d replicated wilson with 4d gauge working and matrix regressing
to Ls copies of wilson.
2015-05-31 15:09:02 +01:00
Peter Boyle
624c0ac3ef
Updates now schur red black solver working
2015-05-25 13:43:58 +01:00
Peter Boyle
d0d41b8bce
Didn't like a print statement
2015-05-21 06:36:15 +01:00
Peter Boyle
4dba8522a1
Got unpreconditioned conjugate gradient to run and converge on a random (uniform random,
...
not even SU(3) for now) gauge field. Convergence history is correctly indepdendent of decomposition
on 1,2,4,8,16 mpi tasks.
Found a couple of simd bugs which required fixed and enhanced the Grid_simd.cc test suite.
Implemented the Mdag, M, MdagM, Meooe Mooee schur type stuff in the wilson dop.
2015-05-19 13:57:35 +01:00
Peter Boyle
11cb3e9a01
Getting closer to having a wilson solver... introducing a first and untested
...
cut at Conjugate gradient. Also copied in Remez, Zolotarev, Chebyshev from
Mike Clark, Tony Kennedy and my BFM package respectively since we know we will
need these. I wanted the structure of
algorithms/approx
algorithms/iterative
etc.. to start taking shape.
2015-05-18 07:47:05 +01:00
Peter Boyle
7992346190
Working towards solvers
2015-05-17 00:19:03 +01:00
