I'm pretty sure this is due to an uninitialized variable in the branch predictor (at least the ARM_SE one). I've got a fix, but i've just been waiting for you to commit stuff before I commit anything.
Ali On Jul 30, 2011, at 10:41 AM, Gabe Black wrote: > Hey everybody. I've started doing some final testing before I put my > x86/O3/FS changes in, and I've found that of the O3 regressions (the > ones I tried running) these two are failing at the tip. > > build/ARM_SE/tests/opt/long/50.vortex/arm/linux/o3-timing > build/X86_SE/tests/opt/long/20.parser/x86/linux/o3-timing > > The last the long regressions ran would have been at most a week ago, > barring technical difficulties, and in the last week only Nilay has > checked in. Could you, Nilay, (or at least someone) verify that these > changed and update the regressions appropriately? I've only run the O3 > so it would be advisable to run the other ones too and see if anything > unexpected changed. The differences are small so there's probably > nothing wrong, the stats just need to be updated. > > Gabe > _______________________________________________ > gem5-dev mailing list > [email protected] > http://m5sim.org/mailman/listinfo/gem5-dev > _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
