I found the problem. I need to run some more tests, and then I'll commit
the patch.
--
Nilay
On Thu, 4 Sep 2014, Nilay Vaish wrote:
Thanks for the information. I'll take a look today.
--
Nilay
On Thu, 4 Sep 2014, Andreas Hansson via gem5-dev wrote:
Hi all,
I suspect the last changes to Ruby has put a notable dent in the
performance, around 30-35% in fact. Based on the host_seconds from the
last few days regression runs, the specific runs that use Ruby (five
instances of 50.memtest for ALPA, and one instance of X86 10.linux-boot)
have slowed down markedly, and the change is consistent. I don’t now why
at this point, but it definitely happened as a result of the last set of
changes to Ruby pushed at the beginning of the week.
I don’t use Ruby myself so I am not too worried, but I figured someone out
there might want to do some profiling.
Andreas
_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev
