Hi, I am trying to warm up the cache before simulation and it would be good if I could control what goes into the cache. I know that TrafficGen could replay a memory trace and thus load desired data into the cache, but afterwards I could not run a workload without re-instantiating the system in the Python config script, right? Then all data in the cache would be lost.
It would also help if I could just run two workloads sequentially, without losing cache data between the runs. Checkpointing could restore memory state but it does not restore the cache. Hopefully there could be an easy way to do this. Thanks! Maggie Du
_______________________________________________ gem5-users mailing list -- [email protected] To unsubscribe send an email to [email protected] %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
