Hi all,
I am trying to use gem5 to simulate a many-core system, focusing on the uncore
design. I have managed to build a detailed cache coherent system model based on
the ruby system and referred to CHI model, it is a 4-core cluster. However, I
found it extremely slow to run real software on a multicore configuration,
typically it can take over 50 mins to boot Linux with a 4-core config, it is
even slower than my rtl simulation.
In my next step, I plan to extend my design to a 64-core cache-coherent shared
memory system, and I want to take advantage of the core models, ruby cache and
interconnect models of gem5, but I am facing a serious simulation performance
problem. I have looked into the dist-gem5, but I didn't find it suitable for my
usage(all cores share memory), and I am not sure if the SST+gem5 scheme can
help with my problem.
Has anyone encountered this problem? Any help would be appreciated!
Thank you!
Best regards, zexin Fu
_______________________________________________
gem5-users mailing list -- gem5-users@gem5.org
To unsubscribe send an email to gem5-users-le...@gem5.org
%(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
