CFD/hypre
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

new benchmark results on tioga (#727)

Browse Source 
Improved benchmark performance on tioga.
This commit is contained in:
Ruipeng Li 2022-09-01 08:25:03 -07:00 committed by GitHub
parent f48a5ce0e3
commit 6bd19c272f
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 106 additions and 106 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

108
src/test/TEST_bench/benchmark_ij.perf.saved.tioga
View File

@ -1,81 +1,81 @@
# Output file: benchmark_ij.out.1 # Output file: benchmark_ij.out.1
PCG Setup wall clock time = 0.359271 seconds PCG Setup wall clock time = 0.194717 seconds
PCG Solve wall clock time = 0.504610 seconds PCG Solve wall clock time = 0.484421 seconds
# Output file: benchmark_ij.out.2 # Output file: benchmark_ij.out.2
PCG Setup wall clock time = 0.312138 seconds PCG Setup wall clock time = 0.181343 seconds
PCG Solve wall clock time = 0.416958 seconds PCG Solve wall clock time = 0.378145 seconds
# Output file: benchmark_ij.out.3 # Output file: benchmark_ij.out.3
PCG Setup wall clock time = 0.448648 seconds PCG Setup wall clock time = 0.227025 seconds
PCG Solve wall clock time = 1.005090 seconds PCG Solve wall clock time = 0.938486 seconds
# Output file: benchmark_ij.out.4 # Output file: benchmark_ij.out.4
PCG Setup wall clock time = 0.477832 seconds PCG Setup wall clock time = 0.258373 seconds
PCG Solve wall clock time = 0.215731 seconds PCG Solve wall clock time = 0.204451 seconds
# Output file: benchmark_ij.out.5 # Output file: benchmark_ij.out.5
PCG Setup wall clock time = 0.433146 seconds PCG Setup wall clock time = 0.278769 seconds
PCG Solve wall clock time = 0.182600 seconds PCG Solve wall clock time = 0.181327 seconds
# Output file: benchmark_ij.out.6 # Output file: benchmark_ij.out.6
PCG Setup wall clock time = 0.419559 seconds PCG Setup wall clock time = 0.219248 seconds
PCG Solve wall clock time = 0.929381 seconds PCG Solve wall clock time = 0.822896 seconds
# Output file: benchmark_ij.out.7 # Output file: benchmark_ij.out.7
PCG Setup wall clock time = 0.498759 seconds PCG Setup wall clock time = 0.313280 seconds
PCG Solve wall clock time = 0.217208 seconds PCG Solve wall clock time = 0.206185 seconds
# Output file: benchmark_ij.out.8 # Output file: benchmark_ij.out.8
PCG Setup wall clock time = 0.336192 seconds PCG Setup wall clock time = 0.177377 seconds
PCG Solve wall clock time = 0.492642 seconds PCG Solve wall clock time = 0.473177 seconds
# Output file: benchmark_ij.out.9 # Output file: benchmark_ij.out.9
PCG Setup wall clock time = 0.317082 seconds PCG Setup wall clock time = 0.169903 seconds
PCG Solve wall clock time = 0.408582 seconds PCG Solve wall clock time = 0.383708 seconds
# Output file: benchmark_ij.out.10 # Output file: benchmark_ij.out.10
PCG Setup wall clock time = 0.415598 seconds PCG Setup wall clock time = 0.209733 seconds
PCG Solve wall clock time = 0.956484 seconds PCG Solve wall clock time = 0.888684 seconds
# Output file: benchmark_ij.out.11 # Output file: benchmark_ij.out.11
PCG Setup wall clock time = 0.464038 seconds PCG Setup wall clock time = 0.246183 seconds
PCG Solve wall clock time = 0.228662 seconds PCG Solve wall clock time = 0.204804 seconds
# Output file: benchmark_ij.out.12 # Output file: benchmark_ij.out.12
PCG Setup wall clock time = 0.400749 seconds PCG Setup wall clock time = 0.266770 seconds
PCG Solve wall clock time = 0.179209 seconds PCG Solve wall clock time = 0.174594 seconds
# Output file: benchmark_ij.out.13 # Output file: benchmark_ij.out.13
PCG Setup wall clock time = 0.459325 seconds PCG Setup wall clock time = 0.204547 seconds
PCG Solve wall clock time = 0.950444 seconds PCG Solve wall clock time = 0.822958 seconds
# Output file: benchmark_ij.out.14 # Output file: benchmark_ij.out.14
PCG Setup wall clock time = 0.969487 seconds PCG Setup wall clock time = 0.975108 seconds
PCG Solve wall clock time = 0.908802 seconds PCG Solve wall clock time = 0.908121 seconds
# Output file: benchmark_ij.out.15 # Output file: benchmark_ij.out.15
PCG Setup wall clock time = 0.883830 seconds PCG Setup wall clock time = 0.881669 seconds
PCG Solve wall clock time = 0.983369 seconds PCG Solve wall clock time = 0.981050 seconds
# Output file: benchmark_ij.out.16 # Output file: benchmark_ij.out.16
PCG Setup wall clock time = 0.901717 seconds PCG Setup wall clock time = 0.903953 seconds
PCG Solve wall clock time = 0.867845 seconds PCG Solve wall clock time = 0.866769 seconds
# Output file: benchmark_ij.out.17 # Output file: benchmark_ij.out.17
PCG Setup wall clock time = 0.466032 seconds PCG Setup wall clock time = 0.265580 seconds
PCG Solve wall clock time = 0.188925 seconds PCG Solve wall clock time = 0.184383 seconds
# Output file: benchmark_ij.out.18 # Output file: benchmark_ij.out.18
PCG Setup wall clock time = 0.514625 seconds PCG Setup wall clock time = 0.262278 seconds
PCG Solve wall clock time = 0.164335 seconds PCG Solve wall clock time = 0.152684 seconds
# Output file: benchmark_ij.out.19 # Output file: benchmark_ij.out.19
PCG Setup wall clock time = 0.510175 seconds PCG Setup wall clock time = 0.304177 seconds
PCG Solve wall clock time = 0.466027 seconds PCG Solve wall clock time = 0.451599 seconds
# Output file: benchmark_ij.out.20 # Output file: benchmark_ij.out.20
PCG Setup wall clock time = 0.303927 seconds PCG Setup wall clock time = 0.209579 seconds
PCG Solve wall clock time = 0.249723 seconds PCG Solve wall clock time = 0.246713 seconds
# Output file: benchmark_ij.out.21 # Output file: benchmark_ij.out.21
PCG Setup wall clock time = 0.477654 seconds PCG Setup wall clock time = 0.270481 seconds
PCG Solve wall clock time = 0.243726 seconds PCG Solve wall clock time = 0.237917 seconds
# Output file: benchmark_ij.out.22 # Output file: benchmark_ij.out.22
PCG Setup wall clock time = 0.510603 seconds PCG Setup wall clock time = 0.273456 seconds
PCG Solve wall clock time = 0.187041 seconds PCG Solve wall clock time = 0.170919 seconds
# Output file: benchmark_ij.out.23 # Output file: benchmark_ij.out.23
PCG Setup wall clock time = 0.538188 seconds PCG Setup wall clock time = 0.268130 seconds
PCG Solve wall clock time = 0.192323 seconds PCG Solve wall clock time = 0.175079 seconds
# Output file: benchmark_ij.out.24 # Output file: benchmark_ij.out.24
PCG Setup wall clock time = 0.484426 seconds PCG Setup wall clock time = 0.244931 seconds
PCG Solve wall clock time = 0.203396 seconds PCG Solve wall clock time = 0.188812 seconds
# Output file: benchmark_ij.out.25 # Output file: benchmark_ij.out.25
PCG Setup wall clock time = 0.504628 seconds PCG Setup wall clock time = 0.236667 seconds
PCG Solve wall clock time = 0.211952 seconds PCG Solve wall clock time = 0.205809 seconds
# Output file: benchmark_ij.out.26 # Output file: benchmark_ij.out.26
PCG Setup wall clock time = 0.479777 seconds PCG Setup wall clock time = 0.289616 seconds
PCG Solve wall clock time = 0.542081 seconds PCG Solve wall clock time = 0.536969 seconds
# Output file: benchmark_ij.out.27 # Output file: benchmark_ij.out.27
PCG Setup wall clock time = 0.539575 seconds PCG Setup wall clock time = 0.310268 seconds
PCG Solve wall clock time = 0.392673 seconds PCG Solve wall clock time = 0.384598 seconds

72
src/test/TEST_bench/benchmark_spgemm.perf.saved.tioga
View File

@ -1,72 +1,72 @@
# Output file: benchmark_spgemm.out.1 # Output file: benchmark_spgemm.out.1
Device Parcsr Matrix-by-Matrix wall clock time = 0.019733 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.019889 seconds
# Output file: benchmark_spgemm.out.2 # Output file: benchmark_spgemm.out.2
Device Parcsr Matrix-by-Matrix wall clock time = 0.077451 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.076819 seconds
# Output file: benchmark_spgemm.out.3 # Output file: benchmark_spgemm.out.3
Device Parcsr Matrix-by-Matrix wall clock time = 0.008892 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.008884 seconds
# Output file: benchmark_spgemm.out.4 # Output file: benchmark_spgemm.out.4
Device Parcsr Matrix-by-Matrix wall clock time = 0.009887 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.010011 seconds
# Output file: benchmark_spgemm.out.5 # Output file: benchmark_spgemm.out.5
Device Parcsr Matrix-by-Matrix wall clock time = 0.008617 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.008854 seconds
# Output file: benchmark_spgemm.out.6 # Output file: benchmark_spgemm.out.6
Device Parcsr Matrix-by-Matrix wall clock time = 0.082292 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.081874 seconds
# Output file: benchmark_spgemm.out.7 # Output file: benchmark_spgemm.out.7
Device Parcsr Matrix-by-Matrix wall clock time = 0.004192 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.004065 seconds
# Output file: benchmark_spgemm.out.8 # Output file: benchmark_spgemm.out.8
Device Parcsr Matrix-by-Matrix wall clock time = 0.006853 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.006895 seconds
# Output file: benchmark_spgemm.out.9 # Output file: benchmark_spgemm.out.9
Device Parcsr Matrix-by-Matrix wall clock time = 0.039470 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.039068 seconds
# Output file: benchmark_spgemm.out.10 # Output file: benchmark_spgemm.out.10
Device Parcsr Matrix-by-Matrix wall clock time = 0.115897 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.116622 seconds
# Output file: benchmark_spgemm.out.11 # Output file: benchmark_spgemm.out.11
Device Parcsr Matrix-by-Matrix wall clock time = 0.016865 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.016892 seconds
# Output file: benchmark_spgemm.out.12 # Output file: benchmark_spgemm.out.12
Device Parcsr Matrix-by-Matrix wall clock time = 0.020121 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.019927 seconds
# Output file: benchmark_spgemm.out.13 # Output file: benchmark_spgemm.out.13
Device Parcsr Matrix-by-Matrix wall clock time = 0.009073 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.009113 seconds
# Output file: benchmark_spgemm.out.14 # Output file: benchmark_spgemm.out.14
Device Parcsr Matrix-by-Matrix wall clock time = 0.131827 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.132173 seconds
# Output file: benchmark_spgemm.out.15 # Output file: benchmark_spgemm.out.15
Device Parcsr Matrix-by-Matrix wall clock time = 0.002368 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.002346 seconds
# Output file: benchmark_spgemm.out.16 # Output file: benchmark_spgemm.out.16
Device Parcsr Matrix-by-Matrix wall clock time = 0.008839 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.008833 seconds
# Output file: benchmark_spgemm.out.17 # Output file: benchmark_spgemm.out.17
Device Parcsr Matrix-by-Matrix wall clock time = 0.002929 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.002913 seconds
# Output file: benchmark_spgemm.out.18 # Output file: benchmark_spgemm.out.18
Device Parcsr Matrix-by-Matrix wall clock time = 0.005855 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.005856 seconds
# Output file: benchmark_spgemm.out.19 # Output file: benchmark_spgemm.out.19
Device Parcsr Matrix-by-Matrix wall clock time = 0.004649 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.002930 seconds
# Output file: benchmark_spgemm.out.20 # Output file: benchmark_spgemm.out.20
Device Parcsr Matrix-by-Matrix wall clock time = 0.002893 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.003172 seconds
# Output file: benchmark_spgemm.out.21 # Output file: benchmark_spgemm.out.21
Device Parcsr Matrix-by-Matrix wall clock time = 0.033856 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.034395 seconds
# Output file: benchmark_spgemm.out.22 # Output file: benchmark_spgemm.out.22
Device Parcsr Matrix-by-Matrix wall clock time = 0.115937 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.116523 seconds
# Output file: benchmark_spgemm.out.23 # Output file: benchmark_spgemm.out.23
Device Parcsr Matrix-by-Matrix wall clock time = 0.014854 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.014950 seconds
# Output file: benchmark_spgemm.out.24 # Output file: benchmark_spgemm.out.24
Device Parcsr Matrix-by-Matrix wall clock time = 0.021270 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.017896 seconds
# Output file: benchmark_spgemm.out.25 # Output file: benchmark_spgemm.out.25
Device Parcsr Matrix-by-Matrix wall clock time = 0.059886 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.060106 seconds
# Output file: benchmark_spgemm.out.26 # Output file: benchmark_spgemm.out.26
Device Parcsr Matrix-by-Matrix wall clock time = 0.588026 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.588868 seconds
# Output file: benchmark_spgemm.out.27 # Output file: benchmark_spgemm.out.27
Device Parcsr Matrix-by-Matrix wall clock time = 0.026862 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.024888 seconds
# Output file: benchmark_spgemm.out.28 # Output file: benchmark_spgemm.out.28
Device Parcsr Matrix-by-Matrix wall clock time = 0.032868 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.030994 seconds
# Output file: benchmark_spgemm.out.29 # Output file: benchmark_spgemm.out.29
Device Parcsr Matrix-by-Matrix wall clock time = 0.065964 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.060870 seconds
# Output file: benchmark_spgemm.out.30 # Output file: benchmark_spgemm.out.30
Device Parcsr Matrix-by-Matrix wall clock time = 0.081763 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.062836 seconds
# Output file: benchmark_spgemm.out.31 # Output file: benchmark_spgemm.out.31
Device Parcsr Matrix-by-Matrix wall clock time = 0.022984 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.022955 seconds
# Output file: benchmark_spgemm.out.32 # Output file: benchmark_spgemm.out.32
Device Parcsr Matrix-by-Matrix wall clock time = 0.021927 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.021522 seconds
# Output file: benchmark_spgemm.out.33 # Output file: benchmark_spgemm.out.33
Device Parcsr Matrix-by-Matrix wall clock time = 0.067044 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.069183 seconds
# Output file: benchmark_spgemm.out.34 # Output file: benchmark_spgemm.out.34
Device Parcsr Matrix-by-Matrix wall clock time = 0.071896 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.071740 seconds
# Output file: benchmark_spgemm.out.35 # Output file: benchmark_spgemm.out.35
Device Parcsr Matrix-by-Matrix wall clock time = 0.029163 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.026649 seconds
# Output file: benchmark_spgemm.out.36 # Output file: benchmark_spgemm.out.36
Device Parcsr Matrix-by-Matrix wall clock time = 0.028319 seconds Device Parcsr Matrix-by-Matrix wall clock time = 0.024854 seconds

32
src/test/TEST_bench/benchmark_struct.perf.saved.tioga
View File

@ -1,24 +1,24 @@
# Output file: benchmark_struct.out.1 # Output file: benchmark_struct.out.1
PCG Setup wall clock time = 0.901832 seconds PCG Setup wall clock time = 0.120527 seconds
PCG Solve wall clock time = 0.700184 seconds PCG Solve wall clock time = 0.524466 seconds
# Output file: benchmark_struct.out.2 # Output file: benchmark_struct.out.2
PCG Setup wall clock time = 1.046741 seconds PCG Setup wall clock time = 0.153928 seconds
PCG Solve wall clock time = 0.867303 seconds PCG Solve wall clock time = 0.703765 seconds
# Output file: benchmark_struct.out.3 # Output file: benchmark_struct.out.3
PCG Setup wall clock time = 8.931370 seconds PCG Setup wall clock time = 1.099122 seconds
PCG Solve wall clock time = 6.930335 seconds PCG Solve wall clock time = 5.576967 seconds
# Output file: benchmark_struct.out.4 # Output file: benchmark_struct.out.4
PCG Setup wall clock time = 12.306432 seconds PCG Setup wall clock time = 1.768987 seconds
PCG Solve wall clock time = 11.384939 seconds PCG Solve wall clock time = 9.280957 seconds
# Output file: benchmark_struct.out.5 # Output file: benchmark_struct.out.5
PCG Setup wall clock time = 0.174218 seconds PCG Setup wall clock time = 0.013666 seconds
PCG Solve wall clock time = 0.190781 seconds PCG Solve wall clock time = 0.106329 seconds
# Output file: benchmark_struct.out.6 # Output file: benchmark_struct.out.6
PCG Setup wall clock time = 0.200355 seconds PCG Setup wall clock time = 0.020694 seconds
PCG Solve wall clock time = 0.377829 seconds PCG Solve wall clock time = 0.259598 seconds
# Output file: benchmark_struct.out.7 # Output file: benchmark_struct.out.7
PCG Setup wall clock time = 0.250921 seconds PCG Setup wall clock time = 0.056668 seconds
PCG Solve wall clock time = 0.444261 seconds PCG Solve wall clock time = 0.355209 seconds
# Output file: benchmark_struct.out.8 # Output file: benchmark_struct.out.8
PCG Setup wall clock time = 0.262308 seconds PCG Setup wall clock time = 0.074631 seconds
PCG Solve wall clock time = 0.687628 seconds PCG Solve wall clock time = 0.548870 seconds