Annoyingly enough, doing the runs we need to gather compiler DNA results in crashes/hangs of various flavors on both ia32 and ppc32-aix.
I think we probably can't ignore this because it really shouldn't be that hard to run 20 iterations of all the SPECjvm98 benchmarks in one JVM instance.
David Grove added a comment - 06/May/08 05:00 PM Annoyingly enough, doing the runs we need to gather compiler DNA results in crashes/hangs of various flavors on both ia32 and ppc32-aix.
I think we probably can't ignore this because it really shouldn't be that hard to run 20 iterations of all the SPECjvm98 benchmarks in one JVM instance.
Oh dear, I thought I'd fixed the issues with RVM-352 and RVM-355. I ran 20 iterations successfully when I closed RVM-355 so maybe this is a break within the time frame of 2 weeks?
Ian Rogers added a comment - 07/May/08 04:51 AM Oh dear, I thought I'd fixed the issues with RVM-352 and RVM-355. I ran 20 iterations successfully when I closed RVM-355 so maybe this is a break within the time frame of 2 weeks?
David Grove added a comment - 09/May/08 08:18 AM The partial data I've been able to gather is close enough to the current data we can live without updating it for 2.9.3
Improvement
Closed
Critical
Gather new compiler DNA
I think we probably can't ignore this because it really shouldn't be that hard to run 20 iterations of all the SPECjvm98 benchmarks in one JVM instance.