cryptarithm1 (normal) has bimodal runtime
I was investigating performance numbers for cryptarithm1 in #8199 (closed), I noticed that the numbers for this benchmark were curiously bimodal. Here are two representative samples from both sides of the distribution.
../../runstdtest/runstdtest ./cryptarithm1 -o1 cryptarithm1.stdout -o1 cryptarithm1.stdout -ghc-timing ; ../../runstdtest/runstdtest ./cryptarithm1 -o1 cryptarithm1.stdout -o1 cryptarithm1.stdout -ghc-timing ; ../../runstdtest/runstdtest ./cryptarithm1 -o1 cryptarithm1.stdout -o1 cryptarithm1.stdout -ghc-timing ; ../../runstdtest/runstdtest ./cryptarithm1 -o1 cryptarithm1.stdout -o1 cryptarithm1.stdout -ghc-timing ; ../../runstdtest/runstdtest ./cryptarithm1 -o1 cryptarithm1.stdout -o1 cryptarithm1.stdout -ghc-timing ;
real 0m0.676s
user 0m0.660s
sys 0m0.013s
<<ghc: 2051819320 bytes, 3929 GCs (3927 + 2), 0/0 avg/max bytes residency (0 samples), 5221368 bytes GC work, 1M in use, 0.000 INIT (0.000 elapsed), 0.643 MUT (0.643 elapsed), 0.026 GC (0.026 elapsed), 0.026 GC(0) (0.025 elapsed), 0.000 GC(1) (0.000 elapsed), 1 balance :ghc>>
real 0m0.610s
user 0m0.603s
sys 0m0.007s
<<ghc: 2051819320 bytes, 3929 GCs (3927 + 2), 0/0 avg/max bytes residency (0 samples), 5221368 bytes GC work, 1M in use, 0.000 INIT (0.000 elapsed), 0.582 MUT (0.582 elapsed), 0.023 GC (0.023 elapsed), 0.023 GC(0) (0.023 elapsed), 0.000 GC(1) (0.000 elapsed), 1 balance :ghc>>I don't have any further resolution on the problem yet.
Trac metadata
| Trac field | Value |
|---|---|
| Version | 7.9 |
| Type | Bug |
| TypeOfFailure | OtherFailure |
| Priority | normal |
| Resolution | Unresolved |
| Component | NoFib benchmark suite |
| Test case | |
| Differential revisions | |
| BlockedBy | |
| Related | |
| Blocking | |
| CC | |
| Operating system | |
| Architecture |