58c5edae2f
Relax an overly strict tolerance in iterative_refiner_test
...
Change-Id: I01afef985c0d248a50df2cadb97e4be9cd8d7889
2018-04-16 12:54:18 -07:00
3b2f1f7d28
Add fake FloatCXSparseCholesky and FloatSuiteSparseCholesky
...
Change-Id: I8d61144d0678b8da92322a18517fa6924d950653
2018-04-12 11:22:49 -07:00
93ba16fefc
Simplify IterativeRefiner
...
Change the loop structure of IterativeRefiner to
unconditionally refine for max_num_iterations.
This is done for two reasons.
1. We expect to use this refinement for a small number of iterations
where the convergence test is useless.
2. Eliminating the convergence test means we can restructure the loop
and save on a sparse matrix-vector multiply, saving precious
compute.
Change-Id: I6347f453a5d19d234af2a2eb1bce811048963e06
2018-04-11 21:15:02 -07:00
05fc04490f
Converts std::unique_lock to std::lock_guard.
...
Tested by compiling for CXX threads, OpenMP, no threads, and TBB.
Change-Id: If1ba5cfce83e2ad4e1015354ce67f5b23e89101f
2018-04-10 17:22:15 -07:00
f973e107d2
Enable mixed precision solves.
...
1. Add Solver::Options::use_mixed_precision_solves,
and Solver::Options::max_num_refinement_iterations.
2. Make SparseCholesky::Create return a unique_ptr.
3. SparseCholesky::Create now takes LinearSolver::Options
as an argument.
4. IterativeRefiner's constructor does not require num_cols
as an argument.
5. SparseNormalCholeskySolver now uses a separate rhs vector.
This basic implementation results in a 10% reduction in solver time
and 30% reduction in linear solver memory usage.
Change-Id: I6830f32cae2febf082d2733262eb2c9f0482b0ea
2018-04-10 11:01:38 -07:00
81f413b720
Optimization for custom small blas multiplication with dynamic
...
template parameters in C level.
- unroll for loops
- matrix access more cache coherent
- platform independant
Briefly, this commit brings 1~50% performance improvments for
most cases in small_blas_gem(m/v)_benchmark, but a small drop
for corner cases with small dimensions especially 1,2,3. Here
we list the results partially, which show decrease percentage
of executing time, compared to unoptimized version.
Platform: desktop PC (i7-7700 CPU MP8@3.60GHz + ubuntu 17.10)
(Lenovo Research Device+ Lab, <yangfan34@lenovo.com >)
Benchmark Time CPU
-----------------------------------------------------------
BM_MatrixMatrixMultiplyDynamic/1/1/1 -0.0850 -0.0851
BM_MatrixMatrixMultiplyDynamic/1/1/2 -0.1444 -0.1446
BM_MatrixMatrixMultiplyDynamic/1/1/3 -0.1934 -0.1935
BM_MatrixMatrixMultiplyDynamic/1/1/4 -0.2933 -0.2934
BM_MatrixMatrixMultiplyDynamic/1/1/8 -0.1579 -0.1580
BM_MatrixMatrixMultiplyDynamic/1/1/12 -0.1556 -0.1558
BM_MatrixMatrixMultiplyDynamic/1/1/15 -0.1598 -0.1599
BM_MatrixMatrixMultiplyDynamic/1/2/1 -0.0797 -0.0799
BM_MatrixMatrixMultiplyDynamic/1/2/2 -0.2950 -0.2951
BM_MatrixMatrixMultiplyDynamic/1/2/3 -0.1363 -0.1364
BM_MatrixMatrixMultiplyDynamic/1/2/4 -0.2435 -0.2437
BM_MatrixMatrixMultiplyDynamic/1/2/8 -0.2299 -0.2300
BM_MatrixMatrixMultiplyDynamic/1/2/12 -0.2441 -0.2442
BM_MatrixMatrixMultiplyDynamic/1/2/15 -0.1671 -0.1673
BM_MatrixMatrixMultiplyDynamic/1/3/1 -0.0774 -0.0775
BM_MatrixMatrixMultiplyDynamic/1/3/2 -0.2761 -0.2762
BM_MatrixMatrixMultiplyDynamic/1/3/3 -0.0840 -0.0841
BM_MatrixMatrixMultiplyDynamic/1/3/4 -0.2027 -0.2028
BM_MatrixMatrixMultiplyDynamic/1/3/8 -0.2481 -0.2482
BM_MatrixMatrixMultiplyDynamic/1/3/12 -0.2629 -0.2630
BM_MatrixMatrixMultiplyDynamic/1/3/15 -0.1958 -0.1959
BM_MatrixMatrixMultiplyDynamic/1/4/1 -0.1260 -0.1261
BM_MatrixMatrixMultiplyDynamic/1/4/2 -0.1834 -0.1835
BM_MatrixMatrixMultiplyDynamic/1/4/3 -0.1379 -0.1380
BM_MatrixMatrixMultiplyDynamic/1/4/4 -0.2636 -0.2637
BM_MatrixMatrixMultiplyDynamic/1/4/8 -0.2838 -0.2839
BM_MatrixMatrixMultiplyDynamic/1/4/12 -0.3320 -0.3321
BM_MatrixMatrixMultiplyDynamic/1/4/15 -0.2464 -0.2465
BM_MatrixMatrixMultiplyDynamic/1/8/1 -0.0766 -0.0767
BM_MatrixMatrixMultiplyDynamic/1/8/2 -0.1713 -0.1714
BM_MatrixMatrixMultiplyDynamic/1/8/3 -0.1158 -0.1159
BM_MatrixMatrixMultiplyDynamic/1/8/4 -0.3205 -0.3206
BM_MatrixMatrixMultiplyDynamic/1/8/8 -0.3514 -0.3515
BM_MatrixMatrixMultiplyDynamic/1/8/12 -0.3658 -0.3658
BM_MatrixMatrixMultiplyDynamic/1/8/15 -0.3187 -0.3188
BM_MatrixMatrixMultiplyDynamic/1/12/1 -0.0424 -0.0425
BM_MatrixMatrixMultiplyDynamic/1/12/2 -0.1800 -0.1800
BM_MatrixMatrixMultiplyDynamic/1/12/3 -0.1457 -0.1457
BM_MatrixMatrixMultiplyDynamic/1/12/4 -0.3768 -0.3769
BM_MatrixMatrixMultiplyDynamic/1/12/8 -0.4072 -0.4073
BM_MatrixMatrixMultiplyDynamic/1/12/12 -0.4391 -0.4392
BM_MatrixMatrixMultiplyDynamic/1/12/15 -0.3383 -0.3383
BM_MatrixMatrixMultiplyDynamic/1/15/1 -0.0442 -0.0443
BM_MatrixMatrixMultiplyDynamic/1/15/2 -0.2378 -0.2379
BM_MatrixMatrixMultiplyDynamic/1/15/3 -0.1553 -0.1554
BM_MatrixMatrixMultiplyDynamic/1/15/4 -0.3954 -0.3955
BM_MatrixMatrixMultiplyDynamic/1/15/8 -0.4334 -0.4335
BM_MatrixMatrixMultiplyDynamic/1/15/12 -0.4175 -0.4175
BM_MatrixMatrixMultiplyDynamic/1/15/15 -0.3242 -0.3243
BM_MatrixVectorMultiply/1/1 +0.1613 +0.1613
BM_MatrixVectorMultiply/1/2 +0.1715 +0.1715
BM_MatrixVectorMultiply/1/3 +0.1051 +0.1051
BM_MatrixVectorMultiply/1/4 +0.1369 +0.1369
BM_MatrixVectorMultiply/1/8 +0.1180 +0.1180
BM_MatrixVectorMultiply/1/12 +0.0869 +0.0869
BM_MatrixVectorMultiply/1/15 +0.1887 +0.1886
BM_MatrixVectorMultiply/2/1 +0.1152 +0.1152
BM_MatrixVectorMultiply/2/2 +0.1520 +0.1520
BM_MatrixVectorMultiply/2/3 +0.1867 +0.1867
BM_MatrixVectorMultiply/2/4 +0.0173 +0.0173
BM_MatrixVectorMultiply/2/8 -0.0528 -0.0528
BM_MatrixVectorMultiply/2/12 -0.0176 -0.0176
BM_MatrixVectorMultiply/2/15 -0.0753 -0.0753
BM_MatrixVectorMultiply/3/1 +0.0844 +0.0844
BM_MatrixVectorMultiply/3/2 +0.0750 +0.0750
BM_MatrixVectorMultiply/3/3 -0.0153 -0.0153
BM_MatrixVectorMultiply/3/4 +0.0060 +0.0060
BM_MatrixVectorMultiply/3/8 +0.0152 +0.0152
BM_MatrixVectorMultiply/3/12 +0.0101 +0.0101
BM_MatrixVectorMultiply/3/15 -0.0795 -0.0795
BM_MatrixVectorMultiply/4/1 -0.1425 -0.1425
BM_MatrixVectorMultiply/4/2 -0.0869 -0.0869
BM_MatrixVectorMultiply/4/3 -0.1371 -0.1371
BM_MatrixVectorMultiply/4/4 -0.0088 -0.0088
BM_MatrixVectorMultiply/4/8 -0.1049 -0.1049
BM_MatrixVectorMultiply/4/12 -0.2566 -0.2566
BM_MatrixVectorMultiply/4/15 -0.2940 -0.2940
BM_MatrixVectorMultiply/6/1 -0.1798 -0.1798
BM_MatrixVectorMultiply/6/2 -0.0627 -0.0627
BM_MatrixVectorMultiply/6/3 -0.0389 -0.0389
BM_MatrixVectorMultiply/6/4 -0.1088 -0.1088
BM_MatrixVectorMultiply/6/8 -0.1815 -0.1815
BM_MatrixVectorMultiply/6/12 -0.1650 -0.1650
BM_MatrixVectorMultiply/6/15 -0.1855 -0.1855
BM_MatrixVectorMultiply/8/1 -0.1630 -0.1630
BM_MatrixVectorMultiply/8/2 -0.1248 -0.1248
BM_MatrixVectorMultiply/8/3 -0.1911 -0.1911
BM_MatrixVectorMultiply/8/4 -0.1996 -0.1996
BM_MatrixVectorMultiply/8/8 -0.2590 -0.2590
BM_MatrixVectorMultiply/8/12 -0.3266 -0.3266
BM_MatrixVectorMultiply/8/15 -0.3999 -0.3999
BM_MatrixTransposeVectorMultiply/1/1 -0.0234 -0.0234
BM_MatrixTransposeVectorMultiply/1/2 -0.0243 -0.0243
BM_MatrixTransposeVectorMultiply/1/3 -0.1324 -0.1324
BM_MatrixTransposeVectorMultiply/1/4 -0.2635 -0.2635
BM_MatrixTransposeVectorMultiply/1/8 -0.2461 -0.2461
BM_MatrixTransposeVectorMultiply/1/12 -0.2702 -0.2702
BM_MatrixTransposeVectorMultiply/1/15 -0.2538 -0.2538
BM_MatrixTransposeVectorMultiply/2/1 -0.0170 -0.0170
BM_MatrixTransposeVectorMultiply/2/2 -0.1475 -0.1475
BM_MatrixTransposeVectorMultiply/2/3 -0.1082 -0.1082
BM_MatrixTransposeVectorMultiply/2/4 -0.2594 -0.2595
BM_MatrixTransposeVectorMultiply/2/8 -0.2710 -0.2710
BM_MatrixTransposeVectorMultiply/2/12 -0.3053 -0.3053
BM_MatrixTransposeVectorMultiply/2/15 -0.2706 -0.2706
BM_MatrixTransposeVectorMultiply/3/1 -0.0096 -0.0096
BM_MatrixTransposeVectorMultiply/3/2 -0.2885 -0.2886
BM_MatrixTransposeVectorMultiply/3/3 -0.0790 -0.0790
BM_MatrixTransposeVectorMultiply/3/4 -0.2329 -0.2330
BM_MatrixTransposeVectorMultiply/3/8 -0.2742 -0.2742
BM_MatrixTransposeVectorMultiply/3/12 -0.3177 -0.3177
BM_MatrixTransposeVectorMultiply/3/15 -0.2610 -0.2610
BM_MatrixTransposeVectorMultiply/4/1 -0.0024 -0.0024
BM_MatrixTransposeVectorMultiply/4/2 -0.1578 -0.1578
BM_MatrixTransposeVectorMultiply/4/3 -0.0918 -0.0918
BM_MatrixTransposeVectorMultiply/4/4 -0.2570 -0.2570
BM_MatrixTransposeVectorMultiply/4/8 -0.3064 -0.3064
BM_MatrixTransposeVectorMultiply/4/12 -0.3316 -0.3316
BM_MatrixTransposeVectorMultiply/4/15 -0.2794 -0.2794
BM_MatrixTransposeVectorMultiply/6/1 -0.0484 -0.0484
BM_MatrixTransposeVectorMultiply/6/2 -0.1102 -0.1102
BM_MatrixTransposeVectorMultiply/6/3 -0.1188 -0.1188
BM_MatrixTransposeVectorMultiply/6/4 -0.2967 -0.2967
BM_MatrixTransposeVectorMultiply/6/8 -0.3190 -0.3190
BM_MatrixTransposeVectorMultiply/6/12 -0.3441 -0.3441
BM_MatrixTransposeVectorMultiply/6/15 -0.2723 -0.2723
BM_MatrixTransposeVectorMultiply/8/1 -0.0397 -0.0397
BM_MatrixTransposeVectorMultiply/8/2 -0.1453 -0.1453
BM_MatrixTransposeVectorMultiply/8/3 -0.1337 -0.1337
BM_MatrixTransposeVectorMultiply/8/4 -0.3084 -0.3084
BM_MatrixTransposeVectorMultiply/8/8 -0.3444 -0.3444
BM_MatrixTransposeVectorMultiply/8/12 -0.3717 -0.3717
BM_MatrixTransposeVectorMultiply/8/15 -0.3440 -0.3440
Change-Id: I17de05bf94699a07eea880b92a6d08daf1f038bb
2018-04-10 14:42:23 +08:00
bdda32bb16
Add MixedSparseCholesky.
...
A simple class that composes SparseCholesky with IterativeRefiner.
Change-Id: I4a67b8ca33a604aaa7b6a4bf511dad9501815f5b
2018-04-09 17:40:34 -07:00
7750c4c55c
Add a single precision variant of EigenSparseCholesky.
...
Given a double precision linear system, solve it using
a single precision Cholesky factorization.
Change-Id: I8a6e8b7a451e961a8a23a62dd7b5d8159c4db8ce
2018-04-09 15:59:45 -07:00
83f70e5c02
Remove unnecessary memory allocations when using SuiteSparse.
...
1. Add SuiteSparse::CreateDenseVectorView
2. Replace calls to SuiteSparse::CreateDenseVector with
SuiteSparse::CreateDenseVectorView.
2. Replace NULL with nullptr in suitesparse.cc and
dynamic_sparse_normal_cholesky_solver.cc
Change-Id: I94355c1dc27789e5b987a7b2850e9db6176a0914
2018-04-09 14:39:02 -07:00
21ec8f462c
Change the API description for SparseCholesky::Solve.
...
The description does not reflect reality and puts constraints
on the underlying implementation.
Change-Id: I5f64e776c952868ae82fb2144acf51c2dcdfab56
2018-04-09 14:20:58 -07:00
5d8b494557
Adds a ParallelFor wrapper for no threads and OpenMP.
...
With the addition of C++11 support we can simplify the parallel for code by
removing the ifdef branching. Converts coordinate_descent_minimizer.cc to use
the thread_id ParallelFor API.
Tested by building with OpenMP, C++11 threads, TBB, and no threads. All tests
pass.
Also compared timing via the bundle adjuster.
./bin/bundle_adjuster --input=../problem-744-543562-pre.txt
With OpenMP num_threads=8
Head:
Time (in seconds):
Residual only evaluation 0.807753 (5)
Jacobian & residual evaluation 4.489404 (6)
Linear solver 41.826481 (5)
Minimizer 50.745857
Total 73.294424
CL:
Time (in seconds):
Residual only evaluation 0.970483 (5)
Jacobian & residual evaluation 4.647438 (6)
Linear solver 41.781892 (5)
Minimizer 50.848904
Total 73.089983
With OpenMP num_threads=1
HEAD:
Time (in seconds):
Residual only evaluation 2.990246 (5)
Jacobian & residual evaluation 14.132090 (6)
Linear solver 79.631951 (5)
Minimizer 100.281847
Total 122.946267
CL:
Time (in seconds):
Residual only evaluation 3.075178 (5)
Jacobian & residual evaluation 13.966451 (6)
Linear solver 77.005441 (5)
Minimizer 97.568712
Total 120.410454
Change-Id: I1857d7943073be7465b6c6476bf46ab11c5475a3
2018-04-09 10:10:03 -07:00
8d7b2528bc
Improve the test coverage in small_blas_test
...
1. Add explicit tests for MatrixMatrixMultiplyNaive and
MatrixTransposeMatrixMultiplyNaive
2. Add tests that exercise a variety of matrix sizes for
MatrixVectorMultiply and MatrixTransposeVectorMultiply.
Change-Id: I0b25ec346b719f19b2067848f9d4bb64c9848750
2018-04-08 15:26:52 -07:00
df4de10ace
Revert "Optimization for custom small blas multiplication with dynamic"
...
This reverts commit 68cc71ce5d
2018-04-08 17:56:16 +00:00
865ac62d38
Change a rather strict tolerance in the iterative_refiner_test
...
Change-Id: I924ab56e24b18502797202167703e4e9bf555550
2018-04-08 10:42:10 -07:00
68cc71ce5d
Optimization for custom small blas multiplication with dynamic
...
template parameters in C level.
- unroll for loops
- matrix access more cache coherent
- platform independant
Briefly, this commit brings 1~50% performance improvments for
most cases in small_blas_gem(m/v)_benchmark, but a small drop
for corner cases with small dimensions especially 1,2,3. Here
we list the results partially, which show decrease percentage
of executing time, compared to unoptimized version.
Platform: desktop PC (i7-7700 CPU MP8@3.60GHz + ubuntu 17.10)
(Lenovo Research Device+ Lab, <yangfan34@lenovo.com >)
Benchmark Time CPU
-----------------------------------------------------------
BM_MatrixMatrixMultiplyDynamic/2/2/2 -0.1082 -0.1083
BM_MatrixMatrixMultiplyDynamic/2/2/15 -0.1270 -0.1270
BM_MatrixMatrixMultiplyDynamic/2/4/2 -0.1433 -0.1433
BM_MatrixMatrixMultiplyDynamic/2/4/15 -0.2069 -0.2068
BM_MatrixMatrixMultiplyDynamic/2/6/2 -0.1446 -0.1446
BM_MatrixMatrixMultiplyDynamic/2/6/15 -0.2156 -0.2156
BM_MatrixMatrixMultiplyDynamic/2/8/2 -0.1788 -0.1788
BM_MatrixMatrixMultiplyDynamic/2/8/15 -0.3316 -0.3316
BM_MatrixMatrixMultiplyDynamic/2/10/2 -0.2025 -0.2025
BM_MatrixMatrixMultiplyDynamic/2/10/15 -0.3444 -0.3444
BM_MatrixMatrixMultiplyDynamic/2/12/2 -0.0515 -0.0515
BM_MatrixMatrixMultiplyDynamic/2/12/15 -0.3733 -0.3733
BM_MatrixMatrixMultiplyDynamic/2/15/2 -0.2784 -0.2784
BM_MatrixMatrixMultiplyDynamic/2/15/15 -0.3704 -0.3704
BM_MatrixMatrixMultiplyDynamic/4/2/2 -0.1839 -0.1839
BM_MatrixMatrixMultiplyDynamic/4/2/15 -0.1922 -0.1922
BM_MatrixMatrixMultiplyDynamic/4/4/2 -0.2248 -0.2248
BM_MatrixMatrixMultiplyDynamic/4/4/15 -0.3132 -0.3132
BM_MatrixMatrixMultiplyDynamic/4/6/2 -0.2311 -0.2311
BM_MatrixMatrixMultiplyDynamic/4/6/15 -0.3239 -0.3239
BM_MatrixMatrixMultiplyDynamic/4/8/2 -0.0574 -0.0574
BM_MatrixMatrixMultiplyDynamic/4/8/15 -0.4173 -0.4173
BM_MatrixMatrixMultiplyDynamic/4/10/2 -0.2861 -0.2861
BM_MatrixMatrixMultiplyDynamic/4/10/15 -0.4065 -0.4064
BM_MatrixMatrixMultiplyDynamic/4/12/2 -0.2976 -0.2975
BM_MatrixMatrixMultiplyDynamic/4/12/15 -0.4218 -0.4218
BM_MatrixMatrixMultiplyDynamic/4/15/2 -0.3116 -0.3116
BM_MatrixMatrixMultiplyDynamic/4/15/15 -0.4242 -0.4241
BM_MatrixMatrixMultiplyDynamic/8/12/2 -0.3675 -0.3674
BM_MatrixMatrixMultiplyDynamic/8/12/4 -0.5055 -0.5055
BM_MatrixMatrixMultiplyDynamic/8/12/6 -0.4302 -0.4302
BM_MatrixMatrixMultiplyDynamic/8/12/8 -0.4854 -0.4854
BM_MatrixMatrixMultiplyDynamic/8/12/10 -0.4882 -0.4882
BM_MatrixMatrixMultiplyDynamic/8/12/12 -0.5209 -0.5209
BM_MatrixMatrixMultiplyDynamic/8/12/15 -0.4558 -0.4558
BM_MatrixMatrixMultiplyDynamic/8/15/2 -0.2319 -0.2319
BM_MatrixMatrixMultiplyDynamic/8/15/4 -0.5105 -0.5105
BM_MatrixMatrixMultiplyDynamic/8/15/6 -0.4477 -0.4477
BM_MatrixMatrixMultiplyDynamic/8/15/8 -0.5479 -0.5479
BM_MatrixMatrixMultiplyDynamic/8/15/10 -0.4843 -0.4843
BM_MatrixMatrixMultiplyDynamic/8/15/12 -0.5212 -0.5212
BM_MatrixMatrixMultiplyDynamic/8/15/15 -0.4459 -0.4459
BM_MatrixVectorMultiply/1/1 +0.0978 +0.0978
BM_MatrixVectorMultiply/1/2 +0.0551 +0.0551
BM_MatrixVectorMultiply/1/3 -0.0019 -0.0020
BM_MatrixVectorMultiply/1/4 +0.0563 +0.0562
BM_MatrixVectorMultiply/1/6 +0.1379 +0.1379
BM_MatrixVectorMultiply/1/7 +0.1090 +0.1090
BM_MatrixVectorMultiply/1/12 +0.0901 +0.0901
BM_MatrixVectorMultiply/1/16 +0.0493 +0.0493
BM_MatrixVectorMultiply/1/20 +0.2255 +0.2255
BM_MatrixVectorMultiply/2/1 +0.1261 +0.1261
BM_MatrixVectorMultiply/2/2 +0.2328 +0.2328
BM_MatrixVectorMultiply/2/3 +0.1404 +0.1403
BM_MatrixVectorMultiply/2/4 +0.0257 +0.0256
BM_MatrixVectorMultiply/2/6 -0.1691 -0.1691
BM_MatrixVectorMultiply/2/7 -0.2619 -0.2619
BM_MatrixVectorMultiply/2/12 -0.4261 -0.4261
BM_MatrixVectorMultiply/2/16 -0.5387 -0.5387
BM_MatrixVectorMultiply/2/20 -0.6171 -0.6171
BM_MatrixVectorMultiply/3/1 +0.1664 +0.1664
BM_MatrixVectorMultiply/3/2 +0.0848 +0.0848
BM_MatrixVectorMultiply/3/3 -0.0044 -0.0044
BM_MatrixVectorMultiply/3/4 -0.0683 -0.0684
BM_MatrixVectorMultiply/3/6 -0.1652 -0.1652
BM_MatrixVectorMultiply/3/7 -0.1633 -0.1633
BM_MatrixVectorMultiply/3/12 -0.1921 -0.1921
BM_MatrixVectorMultiply/3/16 -0.3659 -0.3659
BM_MatrixVectorMultiply/3/20 -0.4137 -0.4137
BM_MatrixVectorMultiply/4/1 -0.0577 -0.0577
BM_MatrixVectorMultiply/4/2 -0.1337 -0.1338
BM_MatrixVectorMultiply/4/3 -0.1443 -0.1443
BM_MatrixVectorMultiply/4/4 +0.0013 +0.0013
BM_MatrixVectorMultiply/4/6 -0.1071 -0.1071
BM_MatrixVectorMultiply/4/7 -0.1396 -0.1397
BM_MatrixVectorMultiply/4/12 -0.2792 -0.2792
BM_MatrixVectorMultiply/4/16 -0.4485 -0.4486
BM_MatrixVectorMultiply/4/20 -0.3588 -0.3588
Change-Id: I64a8cf11391e3d06341a2b8764cd1b4f1b8a23f1
2018-04-08 14:23:10 +00:00
f27082a174
Handle possible overflow in TrustRegionStepEvaluator.
...
Thanks to Volker Grabe for reporting the problem and suggesting
a fix.
Change-Id: I8072ffb275907baac62ee2ad84a02f17eb447c63
2018-04-06 20:13:16 -07:00
86814ab193
Add Iterative Refinement
...
Add a class IterativeRefiner which implements iterative refinement
for SPD linear systems.
Change-Id: I705d4e96cb7de9226ee35e2a9c11d98ffc0ee239
2018-04-06 16:51:14 -07:00
9814a91fcf
Use C++11's inline member initialization syntax
...
Migrate all Option and Summary structs to use
inline member initialization syntax.
This reduces the amount of code, and collocates the
default values with the documentation for the corresponding
member variable.
Change-Id: I8e6b9ee3b31464699d678667f6166ace5fc137c9
2018-04-06 16:50:42 -07:00
6851a63861
Fix lower-bound on result of minimising step-size polynomial.
...
- Previously we were requesting a step-size which minimised the
polynomial in: [previous.x, current.x * factor].
- This is incorrect c/f Nocedal & Wright p60, the bounds for the
minimising step-size should be: [current.x, current.x * factor].
- However, Nocedal & Wright's bounds are insufficient when the function
can return invalid values, which we support.
- In the case that f(current) is invalid, we need to contract the
step-size to lie within: [previous.x, current.x) given that we know
that previous.x is valid so a valid step must exist within that range.
Change-Id: I67b5aaa09cc6d54cf5f264e2cf894ddc2af3f3ad
2018-04-06 19:57:38 +01:00
8202edcbb4
Add missing <algorithm> headers.
...
Thanks to kea@google.com for reporting this.
Change-Id: I9bde5823e9ea89451f476694c9b860080dffdf17
2018-04-05 11:53:48 -07:00
47863a1e56
Remove unnecessary unique_ptr initializations
...
Change-Id: Ibcd3676907f8cb0662b80f600101bcd7bec8e785
2018-04-03 11:29:04 -07:00
e809cf0c28
Adds missing functional include in thread_pool.h
...
Fixes a build error when building with CXX11_THREADS=ON.
Change-Id: I5828dc408261c88a93745c2ade0b8da740a68e54
2018-04-03 11:14:41 -07:00
e49507cbea
More C++ification.
...
"> >" -> ">>"
"> > >" -> ">>>"
Change-Id: Id1ddd9dbf030fe21d57207741b4ca8403068e55b
2018-04-03 10:58:15 -07:00
7c4e8a454e
Replace scoped_ptr with C++11's unique_ptr
...
Change-Id: Ib5a504c491e3a79af52a95accf009df473470c6b
2018-04-02 14:47:47 -07:00
fc9336e12a
Deprecate mutex.h
...
Use C++11's std::mutex and std::lock_guard instead.
Change-Id: If8d513d5d8f2ee2a49c1ebb249ffa93e4a80086b
2018-04-02 06:03:06 -07:00
a1458f3348
More C++11ification.
...
1. Replace HashMap and HashSet with std::unordered_map and
std::unordered_set respectively.
2. Extract the pair hasher into a struct pair_hash.
3. Delete collections_port.h
4. Convert explicit iterator based loops to auto based
loops where sensible.
Change-Id: Ib88bcd13a7463d18435639d3b771abaa52080efb
2018-03-30 12:03:28 -07:00
ffae101c75
Removed CERES_USE_CXX11
...
Remove the special handling for tr1::shared_ptr.
Change-Id: Ic405dcfe881dd6560d236956f84bfb8911d0bcbb
2018-03-29 15:16:49 -07:00
282b8b5f35
Require >= C++11 & CMake >= 3.5 when building Ceres.
...
- Removes CXX11 option, and all associate paraphernalia. Ceres now
requires a compiler with full >= C++11 support. In MSVC terms this
means >= 2013 Release 4.
- This deprecates the use of CERES_STD_UNORDERED_MAP and CERES_USE_CXX11
as they will now always be defined. They will be removed from the
source in a future CL.
- For clients with CMake >= 3.8 we propagate via the exported/installed
Ceres target the CXX version that was specified when Ceres was built.
For versions < 3.8 (but >= 3.5) we specify the CXX features currently
used in the Ceres public API.
Change-Id: I535b545b10156e4426659c270a4a0649e071df0e
2018-03-29 09:32:59 +01:00
9d4d0bfe29
Make CompressedRowSparseMatrix sensitive to StorageType
...
Even though we added support for storing the upper and
lower triangular parts of symmetric matrices in
CompressedRowSparseMatrix. RightMultiply, LeftMultiply
and SquaredColumnNorm were not modified to account for this.
This CL changes their implementation and adds thorough
tests.
Also methods that cannot work correctly with symmetric
storage now CHECK and fail, because that indicates
programmer error.
Change-Id: I76288472c8bac98db7376a79bdb6259e346ef2b7
2018-03-26 20:46:57 +00:00
edffd386d2
Remove an unneeded virtual method from EigenSparseCholesky.
...
Change-Id: I4f9d1730feb98b21db8dea9cc1357a365fb8fcb8
2018-03-26 13:32:52 -07:00
0258b2a0b0
Update the sizes in small_blas_gem?_benchmark.cc
...
The sizes are reflective of the matrix sizes
that occur in production.
Change-Id: I8cd85d0640702df09afafced0d28ef416bab946f
2018-03-23 13:40:14 -07:00
7bdceb46cf
Evaluation callback API
...
This adds a callback mechanism to for users to get notified just
before jacobian and residual evaluations. This will enable
aggressive caching and sharing of compute between cost functions.
Change-Id: I67993726920218edf71ab9ae70c34c204756c71a
2018-03-07 12:41:22 -08:00
83098e12e8
Two changes to how we manage user's state.
...
1. When Solver::Options::update_state_every_iteration = true,
the StateUpdatingCallback only updates the state on a
successful iteration. This works fine but will not work
if the user provides an EvaluationCallback, because
every call to Evaluator::Evaluate will change the state
visible to the user. This means that on unsuccessful
iterations when the user's IterationCallback is called
the state visible to the user would be the last evaluation
which did not lead to an improved cost. This CL forces
the StateUpdatingCallback to unconditionally update
the user visible state.
2. When the minimizer terminates, we update the user visible
state only if the solution is usable, otherwise the user's
state will remain the same. To maintain this invariant
we will now cache the user's state and make sure we
use it to reset it upon return if the solver is not
successful.
Change-Id: Ic1f8fa6cb10d2753130ea752c70ff3d6b2f1462f
2018-03-06 19:12:14 +00:00
2f2e7222fa
Fix c++98 mode compilation for parallel_utils_test.cc
...
Change-Id: I009032978bca910704c52f4de693d04fa13c6e07
2018-03-02 16:41:18 -08:00
21f519d009
Improve threading in covariance.
...
Covariance computation wants to do a triangular iteration but as a
single loop. Right now it iterates over a square and does nothing half
the time, which is inefficient and has bad worst-case threading
performance. This adds a utility that allows waste-free linear iteration
over a triangle.
Change-Id: I881d5683c65882f87dc2b5f8449a855d22ace755
2018-03-02 13:33:19 -08:00
0a30117489
Add parallel bundle tests to CMake build
...
This converts the bundle_adjustment_test to the parallel
version with multiple binaries, as is already in place for
the Bazel build. Additionally, since there is no longer a
need for it, this deletes bundle_adjustment_test.cc.
The test suite now runs on my 6 year old desktop in ~60 seconds!
Change-Id: Ib4a59f8749e823f697e6da5d303977d284303ae3
2018-03-02 19:38:32 +00:00
660960294c
Remove an errant binary character
...
Change-Id: I22b6ef8975c3c8a34e7f96ca89e4ec90461f453c
2018-03-01 13:12:57 -08:00
f0c3b23684
Increases the performance of the C++11 threading.
...
Previously, the thread ID was acquired and released on every iteration
of the for loop. The C++11 concurrent queue implementation is much
slower than TBB's version and consequently this was a huge bottleneck.
This introduces another ParallelFor API which takes the thread ID as a
parameter in the evaluation function. This allows us to acquire and
release the thread ID for each block of work which drastically improves
the performance.
This change brings us on par with OpenMP and TBB. See below for a
timing comparison. Note: in this example this CLs C++11 version is
faster to compute the residuals because TBB still must acquire the
thread ID on every iteration, which has some overhead.
Tested by building and running tests for no threading, OpenMP, TBB, and
C++11 threads. Also ran bazel tests.
./bin/bundle_adjuster --input=problem-744-543562-pre.txt --num_threads=8
C++11 @Head
Time (in seconds):
Residual only evaluation 7.819692 (5)
Jacobian & residual evaluation 11.606063 (6)
Linear solver 47.860195 (5)
Minimizer 70.877072
Total 90.806338
---------------------------------------------------
C++11 (This CL)
Time (in seconds):
Residual only evaluation 1.217500 (5)
Jacobian & residual evaluation 5.796112 (6)
Linear solver 44.080873 (5)
Minimizer 54.635524
Total 77.640072
---------------------------------------------------
OpenMP
Time (in seconds):
Residual only evaluation 0.797023 (5)
Jacobian & residual evaluation 5.633916 (6)
Linear solver 43.280020 (5)
Minimizer 53.199058
Total 76.250861
---------------------------------------------------
TBB
Time (in seconds):
Residual only evaluation 1.911095 (5)
Jacobian & residual evaluation 5.557807 (6)
Linear solver 44.074680 (5)
Minimizer 55.002688
Total 78.052687
---------------------------------------------------
No Threads
Time (in seconds):
Residual only evaluation 2.939212 (5)
Jacobian & residual evaluation 18.519874 (6)
Linear solver 74.017837 (5)
Minimizer 98.980080
Total 122.216391
Change-Id: I3af959b0771bbdfe8cad8c13896191d6ac903181
2018-03-01 09:00:44 -08:00
8c81149e4d
Revert changes to MatrixTransposeMatrixMultiply.
...
This change while algebraically equivalent was causing non-trivial
changes in the numbers being reported by ceres and in some cases
they got worse.
Also fix a small header include in parallel_for_test.cc which
was discovered as part of testing this patch.
Change-Id: I8c8d61538819f0b25af6059fa75ad068fc22f5e4
2018-02-28 19:47:52 -08:00
3d933750a7
Deprecate Solver::Options::num_linear_solver_threads
...
1. Solver::Options::num_threads now controls parallelism in Ceres
Solver. The user specified value of
Solver::Options::num_linear_solver_threads is ignored.
2. If the user specifies Solver::Options::num_linear_solver_threads
and it is different from Solver::Options::num_threads,
a warning is printed.
3. Solver::Summary:num_linear_solver_threads_given and
Solver::Summary::num_linear_solver_threads_used are also
deprecated and are always set to Solver::Summary::num_threads_given
and Solver::Summary::num_threads_used.
Change-Id: I20b9336d9336e400e6f0a15b63857c0c43eb271c
2018-02-27 20:38:05 -08:00
d9d8de556a
Convert a nested loop into a linear loop in covariance_impl.h
...
This improves the readability and simplifies the logic
for interfacing with ParallelFor. More importantly, it paves the
way for ParallelFor refactoring to improve its performance.
Change-Id: I13b05596228900ee00d71f2ccce1db338844b9ab
2018-02-27 15:43:41 -08:00
c48937e8cf
Renable accidentally disabled tests in inner_product_computer_test
...
Change-Id: I25a73a2436c320ac5a3f84f9867258393fb2ffe9
2018-02-27 10:10:08 -08:00
973d30dffa
Revert "Improve the performance of MatrixTransposeVector Multiply."
...
This reverts commit 6c835f81a7
2018-02-27 14:40:45 +00:00
0e0d5b44b5
Make SubsetPreconditionerTest compatible with old Eigen.
...
The enum Eigen::UpLoType did not have a name in older versions
of Eigen, so a templated function using that enum type fails
to compile with earlier versions of Eigen.
This change replaces the enum in the template declaration with
an int.
Change-Id: Id128fd96b76818be347ee6ed5945c231936d9af8
2018-02-27 14:40:17 +00:00
7b2717582a
Adds C++11 parallel for implementation.
...
Implements ParallelFor using the C++11 based ThreadPool. The C++11
parallel for is 50-70% faster than single threaded, and 20-30% slower
than TBB.
Tested by compiling with OpenMP, TBB, and C++11 Threading support and
ran the unit tests. Ran bazel as well.
Change-Id: I7fd6c9037ff9f200ce6999b5f39918995bb6b8ea
2018-02-26 16:12:03 -08:00
f408f89e8b
Adds a Ceres Context structure.
...
A Ceres Context holds common global state that can be re-used within
Ceres. The Context current contains a thread pool if compiling with
C++11 threading support. Threads are expensive to create and destroy so
it is good to maintain across multiple Ceres solves.
Tested by compiling with and without TBB support and ran unit tests. Ran
bazel as well.
Change-Id: I82f598dfae642aa0e81a6039dc174608a5e8dbfb
2018-02-26 10:37:53 -08:00
f217ccfb69
Use MatrixTransposeMatrix multiply from small_blas.h
...
Prior to this Cheng Wang had to work the MatrixTransposeMatrixMultiply
function in inner_product_computer because of its lagging performance.
Now that the implementation of MatrixTransposeMatrixMultiply
has been updated to match the implementation in inner_product_computer
this CL removes the redundancy.
This CL also includes a fix to MatrixTransposeMatrixMultiply
which was missed in the earlier refactoring, because the test
coverage for small_blas is broken. The tests for InnerProductComputer
caught the problem.
Change-Id: Ia8890010da058c2d9fe738dc77f0be34af5618ab
2018-02-25 20:53:10 -08:00
656f9efbd3
Improve the performance of MatrixTransposeMatrix multiply.
...
Compute A'B as rank-1 sums of outer products of the rows of
A and B. This is more cache coherent.
Benchmark Time CPU Time Old Time New CPU Old CPU New
--------------------------------------------------------------------------------------------------------------------------------------
BM_MatrixMatrixMultiplyDynamic/2/2/2 +0.0031 +0.0033 27 27 27 27
BM_MatrixMatrixMultiplyDynamic/2/2/4 +0.0383 +0.0380 41 42 41 42
BM_MatrixMatrixMultiplyDynamic/2/2/6 +0.0001 +0.0019 58 58 57 58
BM_MatrixMatrixMultiplyDynamic/2/2/8 +0.0169 +0.0169 70 71 70 71
BM_MatrixMatrixMultiplyDynamic/2/2/10 +0.0162 +0.0161 83 85 83 85
BM_MatrixMatrixMultiplyDynamic/2/2/12 -0.0049 -0.0048 93 93 93 93
BM_MatrixMatrixMultiplyDynamic/2/2/15 +0.0131 +0.0129 114 115 114 115
BM_MatrixMatrixMultiplyDynamic/2/4/2 +0.0494 +0.0489 32 34 32 34
BM_MatrixMatrixMultiplyDynamic/2/4/4 +0.0121 +0.0121 51 52 51 52
BM_MatrixMatrixMultiplyDynamic/2/4/6 +0.0134 +0.0138 70 71 70 71
BM_MatrixMatrixMultiplyDynamic/2/4/8 +0.0006 +0.0005 90 90 90 90
BM_MatrixMatrixMultiplyDynamic/2/4/10 +0.0248 +0.0248 107 110 107 110
BM_MatrixMatrixMultiplyDynamic/2/4/12 -0.0239 -0.0234 125 122 125 122
BM_MatrixMatrixMultiplyDynamic/2/4/15 +0.0611 +0.0605 155 164 155 164
BM_MatrixMatrixMultiplyDynamic/2/6/2 -0.0017 -0.0017 39 39 39 39
BM_MatrixMatrixMultiplyDynamic/2/6/4 +0.0139 +0.0147 59 60 59 60
BM_MatrixMatrixMultiplyDynamic/2/6/6 -0.1062 -0.1003 89 80 89 80
BM_MatrixMatrixMultiplyDynamic/2/6/8 -0.0339 -0.0344 116 112 115 111
BM_MatrixMatrixMultiplyDynamic/2/6/10 +0.0568 +0.0455 142 150 141 148
BM_MatrixMatrixMultiplyDynamic/2/6/12 -0.0073 -0.0062 155 154 155 154
BM_MatrixMatrixMultiplyDynamic/2/6/15 +0.0219 +0.0215 187 191 187 191
BM_MatrixMatrixMultiplyDynamic/2/8/2 -0.0165 -0.0160 43 42 43 42
BM_MatrixMatrixMultiplyDynamic/2/8/4 -0.0247 -0.0245 71 69 70 69
BM_MatrixMatrixMultiplyDynamic/2/8/6 -0.0642 -0.0614 104 98 104 98
BM_MatrixMatrixMultiplyDynamic/2/8/8 -0.0607 -0.0599 133 125 133 125
BM_MatrixMatrixMultiplyDynamic/2/8/10 -0.0523 -0.0510 166 157 166 157
BM_MatrixMatrixMultiplyDynamic/2/8/12 -0.0381 -0.0376 189 182 189 182
BM_MatrixMatrixMultiplyDynamic/2/8/15 -0.0725 -0.0716 243 226 243 225
BM_MatrixMatrixMultiplyDynamic/2/10/2 -0.0421 -0.0416 50 48 50 48
BM_MatrixMatrixMultiplyDynamic/2/10/4 -0.0558 -0.0549 83 78 83 78
BM_MatrixMatrixMultiplyDynamic/2/10/6 -0.0527 -0.0518 115 109 115 109
BM_MatrixMatrixMultiplyDynamic/2/10/8 -0.0421 -0.0413 150 144 150 144
BM_MatrixMatrixMultiplyDynamic/2/10/10 -0.0606 -0.0601 202 190 202 190
BM_MatrixMatrixMultiplyDynamic/2/10/12 -0.0409 -0.0406 224 215 224 215
BM_MatrixMatrixMultiplyDynamic/2/10/15 -0.0308 -0.0302 277 268 277 268
BM_MatrixMatrixMultiplyDynamic/2/12/2 -0.0059 -0.0058 53 52 53 52
BM_MatrixMatrixMultiplyDynamic/2/12/4 +0.0019 +0.0023 87 87 87 87
BM_MatrixMatrixMultiplyDynamic/2/12/6 -0.0032 -0.0025 126 125 126 125
BM_MatrixMatrixMultiplyDynamic/2/12/8 -0.0039 -0.0038 159 158 159 158
BM_MatrixMatrixMultiplyDynamic/2/12/10 -0.0175 -0.0171 205 201 205 201
BM_MatrixMatrixMultiplyDynamic/2/12/12 -0.0643 -0.0633 295 276 294 276
BM_MatrixMatrixMultiplyDynamic/2/12/15 -0.0215 -0.0201 328 321 328 321
BM_MatrixMatrixMultiplyDynamic/2/15/2 +0.0092 +0.0103 58 58 58 58
BM_MatrixMatrixMultiplyDynamic/2/15/4 +0.0762 +0.0765 98 105 98 105
BM_MatrixMatrixMultiplyDynamic/2/15/6 -0.0155 -0.0172 138 136 138 136
BM_MatrixMatrixMultiplyDynamic/2/15/8 +0.0674 +0.0658 180 192 180 191
BM_MatrixMatrixMultiplyDynamic/2/15/10 +0.0254 +0.0255 236 242 236 242
BM_MatrixMatrixMultiplyDynamic/2/15/12 -0.0383 -0.0378 289 278 289 278
BM_MatrixMatrixMultiplyDynamic/2/15/15 -0.0313 -0.0308 407 394 406 394
BM_MatrixMatrixMultiplyDynamic/4/2/2 +0.0286 +0.0286 42 43 42 43
BM_MatrixMatrixMultiplyDynamic/4/2/4 -0.0455 -0.0439 71 68 71 68
BM_MatrixMatrixMultiplyDynamic/4/2/6 +0.0221 +0.0229 93 95 93 95
BM_MatrixMatrixMultiplyDynamic/4/2/8 -0.0226 -0.0225 120 118 120 118
BM_MatrixMatrixMultiplyDynamic/4/2/10 -0.1153 -0.1130 158 140 158 140
BM_MatrixMatrixMultiplyDynamic/4/2/12 -0.0145 -0.0143 164 162 164 162
BM_MatrixMatrixMultiplyDynamic/4/2/15 -0.0396 -0.0393 205 197 205 197
BM_MatrixMatrixMultiplyDynamic/4/4/2 -0.0696 -0.0682 57 53 57 53
BM_MatrixMatrixMultiplyDynamic/4/4/4 -0.0233 -0.0231 91 89 91 89
BM_MatrixMatrixMultiplyDynamic/4/4/6 -0.0711 -0.0702 130 121 130 121
BM_MatrixMatrixMultiplyDynamic/4/4/8 -0.1180 -0.1113 178 157 176 157
BM_MatrixMatrixMultiplyDynamic/4/4/10 -0.0301 -0.0282 194 188 194 188
BM_MatrixMatrixMultiplyDynamic/4/4/12 -0.1156 -0.1137 251 222 251 222
BM_MatrixMatrixMultiplyDynamic/4/4/15 -0.0182 -0.0179 307 302 307 301
BM_MatrixMatrixMultiplyDynamic/4/6/2 -0.0228 -0.0220 64 63 64 63
BM_MatrixMatrixMultiplyDynamic/4/6/4 +0.0048 +0.0045 106 107 106 107
BM_MatrixMatrixMultiplyDynamic/4/6/6 -0.0131 -0.0126 155 153 155 153
BM_MatrixMatrixMultiplyDynamic/4/6/8 -0.0526 -0.0502 211 200 210 200
BM_MatrixMatrixMultiplyDynamic/4/6/10 -0.0995 -0.0975 274 247 274 247
BM_MatrixMatrixMultiplyDynamic/4/6/12 -0.0436 -0.0430 297 284 297 284
BM_MatrixMatrixMultiplyDynamic/4/6/15 +0.0042 +0.0050 344 345 343 345
BM_MatrixMatrixMultiplyDynamic/4/8/2 -0.0283 -0.0279 74 72 74 72
BM_MatrixMatrixMultiplyDynamic/4/8/4 -0.0282 -0.0277 125 122 125 122
BM_MatrixMatrixMultiplyDynamic/4/8/6 -0.0756 -0.0693 185 171 184 171
BM_MatrixMatrixMultiplyDynamic/4/8/8 -0.0510 -0.0444 267 253 265 253
BM_MatrixMatrixMultiplyDynamic/4/8/10 -0.0512 -0.0460 301 285 299 285
BM_MatrixMatrixMultiplyDynamic/4/8/12 -0.0046 -0.0006 351 349 349 349
BM_MatrixMatrixMultiplyDynamic/4/8/15 -0.0795 -0.0785 452 416 451 416
BM_MatrixMatrixMultiplyDynamic/4/10/2 -0.0415 -0.0406 85 81 85 81
BM_MatrixMatrixMultiplyDynamic/4/10/4 -0.0118 -0.0118 143 141 143 141
BM_MatrixMatrixMultiplyDynamic/4/10/6 -0.0484 -0.0478 217 206 216 206
BM_MatrixMatrixMultiplyDynamic/4/10/8 -0.0373 -0.0372 280 269 280 269
BM_MatrixMatrixMultiplyDynamic/4/10/10 +0.0083 +0.0083 350 353 350 353
BM_MatrixMatrixMultiplyDynamic/4/10/12 -0.0364 -0.0365 436 420 436 420
BM_MatrixMatrixMultiplyDynamic/4/10/15 -0.0221 -0.0210 513 502 512 502
BM_MatrixMatrixMultiplyDynamic/4/12/2 -0.0569 -0.0565 94 89 94 89
BM_MatrixMatrixMultiplyDynamic/4/12/4 -0.0343 -0.0338 162 157 162 157
BM_MatrixMatrixMultiplyDynamic/4/12/6 +0.0197 +0.0198 227 232 227 231
BM_MatrixMatrixMultiplyDynamic/4/12/8 -0.0390 -0.0388 310 298 310 298
BM_MatrixMatrixMultiplyDynamic/4/12/10 -0.0088 -0.0085 389 386 389 386
BM_MatrixMatrixMultiplyDynamic/4/12/12 +0.0524 +0.0505 545 574 545 572
BM_MatrixMatrixMultiplyDynamic/4/12/15 +0.0801 +0.0764 597 645 597 642
BM_MatrixMatrixMultiplyDynamic/4/15/2 -0.0024 -0.0026 99 99 99 99
BM_MatrixMatrixMultiplyDynamic/4/15/4 -0.0289 -0.0270 187 181 186 181
BM_MatrixMatrixMultiplyDynamic/4/15/6 -0.0264 -0.0261 263 256 263 256
BM_MatrixMatrixMultiplyDynamic/4/15/8 -0.0344 -0.0334 349 337 348 337
BM_MatrixMatrixMultiplyDynamic/4/15/10 -0.0189 -0.0191 454 446 454 445
BM_MatrixMatrixMultiplyDynamic/4/15/12 -0.0483 -0.0470 524 499 523 498
BM_MatrixMatrixMultiplyDynamic/4/15/15 -0.1084 -0.1070 840 749 838 748
BM_MatrixMatrixMultiplyDynamic/6/2/2 -0.0574 -0.0561 59 56 59 56
BM_MatrixMatrixMultiplyDynamic/6/2/4 -0.0727 -0.0721 98 91 98 91
BM_MatrixMatrixMultiplyDynamic/6/2/6 -0.0292 -0.0283 133 129 133 129
BM_MatrixMatrixMultiplyDynamic/6/2/8 -0.0283 -0.0278 167 162 167 162
BM_MatrixMatrixMultiplyDynamic/6/2/10 -0.0482 -0.0476 216 206 216 206
BM_MatrixMatrixMultiplyDynamic/6/2/12 -0.0117 -0.0115 243 240 243 240
BM_MatrixMatrixMultiplyDynamic/6/2/15 -0.0234 -0.0225 299 292 299 292
BM_MatrixMatrixMultiplyDynamic/6/4/2 -0.0950 -0.0898 79 72 79 72
BM_MatrixMatrixMultiplyDynamic/6/4/4 -0.0300 -0.0298 130 127 130 127
BM_MatrixMatrixMultiplyDynamic/6/4/6 -0.0379 -0.0376 175 169 175 169
BM_MatrixMatrixMultiplyDynamic/6/4/8 -0.0340 -0.0319 236 228 236 228
BM_MatrixMatrixMultiplyDynamic/6/4/10 -0.0498 -0.0490 289 275 289 275
BM_MatrixMatrixMultiplyDynamic/6/4/12 +0.0134 +0.0134 325 330 325 330
BM_MatrixMatrixMultiplyDynamic/6/4/15 -0.0836 -0.0794 469 429 466 429
BM_MatrixMatrixMultiplyDynamic/6/6/2 -0.0614 -0.0563 91 86 91 86
BM_MatrixMatrixMultiplyDynamic/6/6/4 -0.0249 -0.0239 153 149 152 149
BM_MatrixMatrixMultiplyDynamic/6/6/6 -0.0442 -0.0438 235 224 234 224
BM_MatrixMatrixMultiplyDynamic/6/6/8 -0.0127 -0.0083 296 292 295 292
BM_MatrixMatrixMultiplyDynamic/6/6/10 +0.0305 +0.0302 360 371 360 371
BM_MatrixMatrixMultiplyDynamic/6/6/12 -0.0479 -0.0463 434 413 433 413
BM_MatrixMatrixMultiplyDynamic/6/6/15 -0.0774 -0.0742 542 500 540 500
BM_MatrixMatrixMultiplyDynamic/6/8/2 -0.0617 -0.0611 106 99 106 99
BM_MatrixMatrixMultiplyDynamic/6/8/4 -0.1153 -0.1147 202 179 202 179
BM_MatrixMatrixMultiplyDynamic/6/8/6 -0.0591 -0.0589 270 254 270 254
BM_MatrixMatrixMultiplyDynamic/6/8/8 -0.0677 -0.0671 381 355 381 355
BM_MatrixMatrixMultiplyDynamic/6/8/10 -0.0771 -0.0757 451 416 450 416
BM_MatrixMatrixMultiplyDynamic/6/8/12 -0.0520 -0.0519 522 495 522 495
BM_MatrixMatrixMultiplyDynamic/6/8/15 -0.0386 -0.0386 639 614 638 614
BM_MatrixMatrixMultiplyDynamic/6/10/2 -0.0764 -0.0756 121 111 120 111
BM_MatrixMatrixMultiplyDynamic/6/10/4 -0.0847 -0.0777 224 205 222 205
BM_MatrixMatrixMultiplyDynamic/6/10/6 -0.0762 -0.0748 320 295 319 295
BM_MatrixMatrixMultiplyDynamic/6/10/8 -0.0160 -0.0164 413 407 413 406
BM_MatrixMatrixMultiplyDynamic/6/10/10 +0.0438 +0.0429 526 549 525 548
BM_MatrixMatrixMultiplyDynamic/6/10/12 +0.0641 +0.0615 606 645 605 643
BM_MatrixMatrixMultiplyDynamic/6/10/15 +0.0983 +0.0965 741 814 740 812
BM_MatrixMatrixMultiplyDynamic/6/12/2 +0.0697 +0.0663 135 144 135 144
BM_MatrixMatrixMultiplyDynamic/6/12/4 -0.0230 -0.0240 251 245 250 244
BM_MatrixMatrixMultiplyDynamic/6/12/6 -0.0728 -0.0703 365 338 364 338
BM_MatrixMatrixMultiplyDynamic/6/12/8 -0.1444 -0.1402 514 440 511 440
BM_MatrixMatrixMultiplyDynamic/6/12/10 -0.1268 -0.1216 628 548 624 548
BM_MatrixMatrixMultiplyDynamic/6/12/12 -0.1323 -0.1250 883 766 875 766
BM_MatrixMatrixMultiplyDynamic/6/12/15 -0.1337 -0.1304 995 862 991 862
BM_MatrixMatrixMultiplyDynamic/6/15/2 -0.1163 -0.1140 158 140 158 140
BM_MatrixMatrixMultiplyDynamic/6/15/4 -0.0721 -0.0689 285 265 284 265
BM_MatrixMatrixMultiplyDynamic/6/15/6 -0.0597 -0.0582 414 389 413 389
BM_MatrixMatrixMultiplyDynamic/6/15/8 -0.1189 -0.1173 563 496 562 496
BM_MatrixMatrixMultiplyDynamic/6/15/10 -0.0754 -0.0737 749 693 747 692
BM_MatrixMatrixMultiplyDynamic/6/15/12 +0.0141 +0.0153 871 883 866 879
BM_MatrixMatrixMultiplyDynamic/6/15/15 -0.1184 -0.1152 1398 1233 1387 1228
BM_MatrixMatrixMultiplyDynamic/8/2/2 -0.0232 -0.0242 77 75 77 75
BM_MatrixMatrixMultiplyDynamic/8/2/4 +0.1114 +0.1085 120 133 119 132
BM_MatrixMatrixMultiplyDynamic/8/2/6 +0.0794 +0.0760 169 183 169 182
BM_MatrixMatrixMultiplyDynamic/8/2/8 +0.0833 +0.0780 221 239 221 238
BM_MatrixMatrixMultiplyDynamic/8/2/10 +0.0037 +0.0037 264 265 264 265
BM_MatrixMatrixMultiplyDynamic/8/2/12 +0.0114 +0.0115 301 304 301 304
BM_MatrixMatrixMultiplyDynamic/8/2/15 +0.0034 +0.0030 378 380 378 380
BM_MatrixMatrixMultiplyDynamic/8/4/2 +0.0197 +0.0198 90 91 90 91
BM_MatrixMatrixMultiplyDynamic/8/4/4 +0.0231 +0.0230 156 160 156 160
BM_MatrixMatrixMultiplyDynamic/8/4/6 +0.0233 +0.0235 225 230 225 230
BM_MatrixMatrixMultiplyDynamic/8/4/8 +0.0352 +0.0349 286 296 286 296
BM_MatrixMatrixMultiplyDynamic/8/4/10 +0.0569 +0.0551 362 382 362 382
BM_MatrixMatrixMultiplyDynamic/8/4/12 -0.0180 -0.0181 424 416 424 416
BM_MatrixMatrixMultiplyDynamic/8/4/15 -0.0807 -0.0763 627 576 623 576
BM_MatrixMatrixMultiplyDynamic/8/6/2 -0.0030 -0.0029 111 111 111 111
BM_MatrixMatrixMultiplyDynamic/8/6/4 +0.0053 +0.0051 198 199 198 199
BM_MatrixMatrixMultiplyDynamic/8/6/6 -0.1261 -0.1233 330 288 329 288
BM_MatrixMatrixMultiplyDynamic/8/6/8 -0.0760 -0.0755 398 368 398 368
BM_MatrixMatrixMultiplyDynamic/8/6/10 -0.0497 -0.0487 505 480 505 480
BM_MatrixMatrixMultiplyDynamic/8/6/12 +0.0090 +0.0091 529 534 529 534
BM_MatrixMatrixMultiplyDynamic/8/6/15 -0.0778 -0.0757 704 649 702 649
BM_MatrixMatrixMultiplyDynamic/8/8/2 -0.1332 -0.1305 150 130 149 130
BM_MatrixMatrixMultiplyDynamic/8/8/4 -0.0903 -0.0878 257 234 257 234
BM_MatrixMatrixMultiplyDynamic/8/8/6 -0.1177 -0.1165 380 335 379 335
BM_MatrixMatrixMultiplyDynamic/8/8/8 -0.0799 -0.0786 509 468 508 468
BM_MatrixMatrixMultiplyDynamic/8/8/10 -0.1098 -0.1057 604 538 601 537
BM_MatrixMatrixMultiplyDynamic/8/8/12 -0.0676 -0.0662 697 650 696 650
BM_MatrixMatrixMultiplyDynamic/8/8/15 -0.0707 -0.0696 856 796 855 796
BM_MatrixMatrixMultiplyDynamic/8/10/2 -0.0429 -0.0417 149 142 149 142
BM_MatrixMatrixMultiplyDynamic/8/10/4 -0.0652 -0.0640 286 267 286 267
BM_MatrixMatrixMultiplyDynamic/8/10/6 -0.0259 -0.0253 402 392 402 392
BM_MatrixMatrixMultiplyDynamic/8/10/8 -0.0131 -0.0129 526 520 526 519
BM_MatrixMatrixMultiplyDynamic/8/10/10 -0.0873 -0.0865 697 636 696 636
BM_MatrixMatrixMultiplyDynamic/8/10/12 -0.0130 -0.0127 799 788 798 788
BM_MatrixMatrixMultiplyDynamic/8/10/15 -0.0441 -0.0438 1018 973 1018 973
BM_MatrixMatrixMultiplyDynamic/8/12/2 -0.0384 -0.0379 172 166 172 165
BM_MatrixMatrixMultiplyDynamic/8/12/4 -0.0321 -0.0307 307 297 307 297
BM_MatrixMatrixMultiplyDynamic/8/12/6 -0.0394 -0.0380 458 440 458 440
BM_MatrixMatrixMultiplyDynamic/8/12/8 -0.0529 -0.0522 608 576 607 576
BM_MatrixMatrixMultiplyDynamic/8/12/10 -0.0566 -0.0560 747 705 747 705
BM_MatrixMatrixMultiplyDynamic/8/12/12 -0.0007 -0.0006 1024 1024 1024 1023
BM_MatrixMatrixMultiplyDynamic/8/12/15 -0.0422 -0.0420 1163 1114 1162 1113
BM_MatrixMatrixMultiplyDynamic/8/15/2 -0.0241 -0.0241 197 192 197 192
BM_MatrixMatrixMultiplyDynamic/8/15/4 -0.0515 -0.0514 352 334 352 333
BM_MatrixMatrixMultiplyDynamic/8/15/6 -0.0192 -0.0194 506 496 506 496
BM_MatrixMatrixMultiplyDynamic/8/15/8 -0.0338 -0.0334 674 651 674 651
BM_MatrixMatrixMultiplyDynamic/8/15/10 -0.0556 -0.0545 922 871 921 871
BM_MatrixMatrixMultiplyDynamic/8/15/12 -0.0473 -0.0462 1052 1002 1050 1001
BM_MatrixMatrixMultiplyDynamic/8/15/15 -0.0522 -0.0521 1484 1406 1483 1406
BM_MatrixTransposeMatrixMultiplyDynamic/2/2/2 -0.1216 -0.1203 31 27 31 27
BM_MatrixTransposeMatrixMultiplyDynamic/2/2/4 -0.3827 -0.3818 49 30 49 30
BM_MatrixTransposeMatrixMultiplyDynamic/2/2/6 -0.3901 -0.3888 67 41 67 41
BM_MatrixTransposeMatrixMultiplyDynamic/2/2/8 -0.5469 -0.5458 86 39 86 39
BM_MatrixTransposeMatrixMultiplyDynamic/2/4/2 -0.2362 -0.2365 47 36 47 36
BM_MatrixTransposeMatrixMultiplyDynamic/2/4/4 -0.4470 -0.4469 77 43 77 43
BM_MatrixTransposeMatrixMultiplyDynamic/2/4/6 -0.4545 -0.4546 111 61 111 61
BM_MatrixTransposeMatrixMultiplyDynamic/2/4/8 -0.6093 -0.6093 142 56 142 56
BM_MatrixTransposeMatrixMultiplyDynamic/2/5/2 -0.2345 -0.2348 54 41 54 41
BM_MatrixTransposeMatrixMultiplyDynamic/2/5/4 -0.5111 -0.5107 99 48 99 48
BM_MatrixTransposeMatrixMultiplyDynamic/2/5/6 -0.4944 -0.4946 136 69 136 69
BM_MatrixTransposeMatrixMultiplyDynamic/2/5/8 -0.6354 -0.6354 171 62 171 62
BM_MatrixTransposeMatrixMultiplyDynamic/2/8/2 -0.3370 -0.3369 82 55 82 55
BM_MatrixTransposeMatrixMultiplyDynamic/2/8/4 -0.5062 -0.5064 139 69 139 69
BM_MatrixTransposeMatrixMultiplyDynamic/2/8/6 -0.5422 -0.5421 208 95 208 95
BM_MatrixTransposeMatrixMultiplyDynamic/2/8/8 -0.6903 -0.6896 294 91 294 91
BM_MatrixTransposeMatrixMultiplyDynamic/2/10/2 -0.1791 -0.1811 99 82 99 81
BM_MatrixTransposeMatrixMultiplyDynamic/2/10/4 -0.5062 -0.5061 177 88 177 88
BM_MatrixTransposeMatrixMultiplyDynamic/2/10/6 -0.5110 -0.5110 256 125 256 125
BM_MatrixTransposeMatrixMultiplyDynamic/2/10/8 -0.6632 -0.6632 334 112 333 112
BM_MatrixTransposeMatrixMultiplyDynamic/2/12/2 -0.3272 -0.3269 120 81 120 81
BM_MatrixTransposeMatrixMultiplyDynamic/2/12/4 -0.5019 -0.5019 212 105 212 105
BM_MatrixTransposeMatrixMultiplyDynamic/2/12/6 -0.5365 -0.5365 303 140 303 140
BM_MatrixTransposeMatrixMultiplyDynamic/2/12/8 -0.6856 -0.6856 398 125 398 125
BM_MatrixTransposeMatrixMultiplyDynamic/2/15/2 -0.2425 -0.2425 143 109 143 109
BM_MatrixTransposeMatrixMultiplyDynamic/2/15/4 -0.5891 -0.5871 298 122 296 122
BM_MatrixTransposeMatrixMultiplyDynamic/2/15/6 -0.5790 -0.5782 414 174 413 174
BM_MatrixTransposeMatrixMultiplyDynamic/2/15/8 -0.7173 -0.7173 509 144 509 144
BM_MatrixTransposeMatrixMultiplyDynamic/4/2/2 +0.0482 +0.0489 37 39 37 39
BM_MatrixTransposeMatrixMultiplyDynamic/4/2/4 -0.2428 -0.2433 60 45 60 45
BM_MatrixTransposeMatrixMultiplyDynamic/4/2/6 -0.2921 -0.2917 84 59 84 59
BM_MatrixTransposeMatrixMultiplyDynamic/4/2/8 -0.4632 -0.4630 105 56 105 56
BM_MatrixTransposeMatrixMultiplyDynamic/4/4/2 -0.0305 -0.0301 60 58 60 58
BM_MatrixTransposeMatrixMultiplyDynamic/4/4/4 -0.3008 -0.3007 101 70 101 70
BM_MatrixTransposeMatrixMultiplyDynamic/4/4/6 -0.3708 -0.3700 149 94 149 94
BM_MatrixTransposeMatrixMultiplyDynamic/4/4/8 -0.5160 -0.5159 190 92 189 92
BM_MatrixTransposeMatrixMultiplyDynamic/4/5/2 -0.0187 -0.0185 71 70 71 70
BM_MatrixTransposeMatrixMultiplyDynamic/4/5/4 -0.3064 -0.3054 130 90 130 90
BM_MatrixTransposeMatrixMultiplyDynamic/4/5/6 -0.3762 -0.3753 184 115 184 115
BM_MatrixTransposeMatrixMultiplyDynamic/4/5/8 -0.5845 -0.5838 240 100 240 100
BM_MatrixTransposeMatrixMultiplyDynamic/4/8/2 +0.0517 +0.0521 109 115 109 115
BM_MatrixTransposeMatrixMultiplyDynamic/4/8/4 -0.2895 -0.2889 194 138 194 138
BM_MatrixTransposeMatrixMultiplyDynamic/4/8/6 -0.3138 -0.3136 279 191 279 191
BM_MatrixTransposeMatrixMultiplyDynamic/4/8/8 -0.5429 -0.5421 381 174 380 174
BM_MatrixTransposeMatrixMultiplyDynamic/4/10/2 +0.0612 +0.0620 135 143 135 143
BM_MatrixTransposeMatrixMultiplyDynamic/4/10/4 -0.3269 -0.3263 259 174 258 174
BM_MatrixTransposeMatrixMultiplyDynamic/4/10/6 -0.2955 -0.2961 343 242 343 241
BM_MatrixTransposeMatrixMultiplyDynamic/4/10/8 -0.5484 -0.5475 466 210 465 210
BM_MatrixTransposeMatrixMultiplyDynamic/4/12/2 +0.1234 +0.1239 157 177 157 177
BM_MatrixTransposeMatrixMultiplyDynamic/4/12/4 -0.3002 -0.2997 291 204 291 204
BM_MatrixTransposeMatrixMultiplyDynamic/4/12/6 -0.3419 -0.3415 415 273 415 273
BM_MatrixTransposeMatrixMultiplyDynamic/4/12/8 -0.5604 -0.5600 545 239 544 239
BM_MatrixTransposeMatrixMultiplyDynamic/4/15/2 +0.1730 +0.1712 185 217 185 216
BM_MatrixTransposeMatrixMultiplyDynamic/4/15/4 -0.3099 -0.3097 346 239 346 239
BM_MatrixTransposeMatrixMultiplyDynamic/4/15/6 -0.3743 -0.3741 521 326 521 326
BM_MatrixTransposeMatrixMultiplyDynamic/4/15/8 -0.5411 -0.5409 654 300 654 300
BM_MatrixTransposeMatrixMultiplyDynamic/6/2/2 +0.2546 +0.2547 41 51 41 51
BM_MatrixTransposeMatrixMultiplyDynamic/6/2/4 -0.0940 -0.0942 66 60 66 60
BM_MatrixTransposeMatrixMultiplyDynamic/6/2/6 -0.1584 -0.1585 95 80 95 80
BM_MatrixTransposeMatrixMultiplyDynamic/6/2/8 -0.4333 -0.4329 130 74 130 74
BM_MatrixTransposeMatrixMultiplyDynamic/6/4/2 +0.1951 +0.1952 69 82 69 82
BM_MatrixTransposeMatrixMultiplyDynamic/6/4/4 -0.2034 -0.2034 123 98 123 98
BM_MatrixTransposeMatrixMultiplyDynamic/6/4/6 -0.1998 -0.1995 178 142 178 142
BM_MatrixTransposeMatrixMultiplyDynamic/6/4/8 -0.4348 -0.4347 233 131 233 131
BM_MatrixTransposeMatrixMultiplyDynamic/6/5/2 +0.5517 +0.5521 82 128 82 128
BM_MatrixTransposeMatrixMultiplyDynamic/6/5/4 -0.1427 -0.1427 145 124 145 124
BM_MatrixTransposeMatrixMultiplyDynamic/6/5/6 -0.1846 -0.1846 220 179 220 179
BM_MatrixTransposeMatrixMultiplyDynamic/6/5/8 -0.4871 -0.4873 290 149 290 148
BM_MatrixTransposeMatrixMultiplyDynamic/6/8/2 +0.3469 +0.3473 131 176 131 176
BM_MatrixTransposeMatrixMultiplyDynamic/6/8/4 -0.1013 -0.1018 248 223 248 222
BM_MatrixTransposeMatrixMultiplyDynamic/6/8/6 -0.1652 -0.1670 354 295 353 294
BM_MatrixTransposeMatrixMultiplyDynamic/6/8/8 -0.4302 -0.4299 447 255 447 255
BM_MatrixTransposeMatrixMultiplyDynamic/6/10/2 +0.4005 +0.4001 157 219 157 219
BM_MatrixTransposeMatrixMultiplyDynamic/6/10/4 -0.0439 -0.0441 295 282 294 281
BM_MatrixTransposeMatrixMultiplyDynamic/6/10/6 -0.1031 -0.1074 420 377 420 375
BM_MatrixTransposeMatrixMultiplyDynamic/6/10/8 -0.3655 -0.3679 556 353 555 351
BM_MatrixTransposeMatrixMultiplyDynamic/6/12/2 +0.5654 +0.5604 183 287 183 286
BM_MatrixTransposeMatrixMultiplyDynamic/6/12/4 -0.1118 -0.1116 353 314 353 314
BM_MatrixTransposeMatrixMultiplyDynamic/6/12/6 -0.2045 -0.2053 517 411 516 410
BM_MatrixTransposeMatrixMultiplyDynamic/6/12/8 -0.4662 -0.4660 657 350 656 350
BM_MatrixTransposeMatrixMultiplyDynamic/6/15/2 +0.3582 +0.3588 242 328 241 328
BM_MatrixTransposeMatrixMultiplyDynamic/6/15/4 -0.1403 -0.1407 428 368 428 367
BM_MatrixTransposeMatrixMultiplyDynamic/6/15/6 -0.2424 -0.2421 632 479 632 479
BM_MatrixTransposeMatrixMultiplyDynamic/6/15/8 -0.5094 -0.5092 835 410 834 409
BM_MatrixTransposeMatrixMultiplyDynamic/8/2/2 +0.3930 +0.3925 47 66 47 66
BM_MatrixTransposeMatrixMultiplyDynamic/8/2/4 +0.0226 +0.0215 78 80 78 80
BM_MatrixTransposeMatrixMultiplyDynamic/8/2/6 -0.0428 -0.0427 105 100 105 100
BM_MatrixTransposeMatrixMultiplyDynamic/8/2/8 -0.3418 -0.3437 141 93 141 92
BM_MatrixTransposeMatrixMultiplyDynamic/8/4/2 +0.3927 +0.3924 85 118 85 118
BM_MatrixTransposeMatrixMultiplyDynamic/8/4/4 -0.0040 -0.0039 137 136 136 136
BM_MatrixTransposeMatrixMultiplyDynamic/8/4/6 -0.0342 -0.0339 204 197 204 197
BM_MatrixTransposeMatrixMultiplyDynamic/8/4/8 -0.3575 -0.3577 267 171 267 171
BM_MatrixTransposeMatrixMultiplyDynamic/8/5/2 +0.7014 +0.7033 99 168 99 168
BM_MatrixTransposeMatrixMultiplyDynamic/8/5/4 +0.0952 +0.0952 163 179 163 179
BM_MatrixTransposeMatrixMultiplyDynamic/8/5/6 -0.1164 -0.1163 246 218 246 218
BM_MatrixTransposeMatrixMultiplyDynamic/8/5/8 -0.3550 -0.3565 322 208 322 207
BM_MatrixTransposeMatrixMultiplyDynamic/8/8/2 +0.7074 +0.7072 140 239 140 239
BM_MatrixTransposeMatrixMultiplyDynamic/8/8/4 +0.1072 +0.1075 276 306 276 306
BM_MatrixTransposeMatrixMultiplyDynamic/8/8/6 -0.0771 -0.0768 398 368 398 367
BM_MatrixTransposeMatrixMultiplyDynamic/8/8/8 -0.3614 -0.3612 524 335 524 335
BM_MatrixTransposeMatrixMultiplyDynamic/8/10/2 +0.6979 +0.7004 184 313 184 312
BM_MatrixTransposeMatrixMultiplyDynamic/8/10/4 +0.0967 +0.0955 326 357 325 356
BM_MatrixTransposeMatrixMultiplyDynamic/8/10/6 -0.0248 -0.0239 486 474 485 474
BM_MatrixTransposeMatrixMultiplyDynamic/8/10/8 -0.3432 -0.3437 603 396 603 396
BM_MatrixTransposeMatrixMultiplyDynamic/8/12/2 +0.5977 +0.5985 210 335 210 335
BM_MatrixTransposeMatrixMultiplyDynamic/8/12/4 -0.0530 -0.0484 405 384 403 384
BM_MatrixTransposeMatrixMultiplyDynamic/8/12/6 -0.0557 -0.0561 574 542 574 541
BM_MatrixTransposeMatrixMultiplyDynamic/8/12/8 -0.2835 -0.2849 730 523 730 522
BM_MatrixTransposeMatrixMultiplyDynamic/8/15/2 +0.7117 +0.7085 254 434 254 433
BM_MatrixTransposeMatrixMultiplyDynamic/8/15/4 +0.0630 +0.0602 458 487 458 486
BM_MatrixTransposeMatrixMultiplyDynamic/8/15/6 -0.0321 -0.0330 662 641 662 640
BM_MatrixTransposeMatrixMultiplyDynamic/8/15/8 -0.3974 -0.3978 911 549 910 548
Change-Id: I09d87c1885be86177b93c8fe0e6f89a75eb2f7a1
2018-02-26 00:17:51 +00:00
d634d6168e
Add small_blas_gemm_benchmark.
...
Change-Id: Ide8c86c77c4adf73a3f5b6be4f23ab3d59705f7b
2018-02-25 14:42:06 -08:00
39ccb79ae6
Implement some C++11 math functions for Jet
...
Change-Id: Ie164aeaac980cd2fe48a86230e1948c9fd75424a
2018-02-25 21:48:48 +00:00
Sameer Agarwal