Hi all,

I am investigating unpredictability in systems and am attempting to measure
the effect of different replacement policies on cache timing. I am using
the classic (non-Ruby) cache objects. Currently the caches start empty, so
I am getting identical timing measurements for each execution.

I have tried using the warmup_percentage parameter, but I note that in
stats.txt, when I set it to 100, the cache.tags.warmup_cycle tick is equal
to 0.

My question is, is there a built-in way to warm up the caches by filling
them with random data before I measure the timing of the caches? Or do I
need to make my own warm up program?
Is it better to use Ruby caches in this use-case?

Thanks in advance for your help,

Sam
_______________________________________________
gem5-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
%(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s

Reply via email to