Maybe Korey can correct me if I'm wrong, but I think max_insts_all_threads says that each thread within an SMT processor must hit N insts, but doesn't apply across multiple cpus. There was a thread in the last month or so with advice on how to make all cpus go to a certain limit, see link.
http://www.mail-archive.com/[email protected]/msg03094.html <http://www.mail-archive.com/[email protected]/msg03094.html>Lisa 2010/4/14 健勇 张 <[email protected]> > Hi,everyone, > I have a question on the definition “system.cpu[i].max_insts_all_threads”. > > I ran 4core simulation on ALPHA SE mood,each core had 1 thread workload. I > set the system.cpu[i].max_insts_all_threads to 100 milion, after > termination ‘normally’, I checked the m5stats.txt for actual num of > instructions ran on each core: > > /**************************************************************/ > ….. > sim_insts 284804592 # Number of instructions simulated > ….. > system.cpu0.committedInsts 39885377 # Number of Instructions Simulated > ….. > system.cpu1.committedInsts 86194454 # Number of Instructions > Simulated > ….. > system.cpu2.committedInsts 58724755 # Number of Instructions > Simulated > ….. > system.cpu3.committedInsts 100000006 # Number of Instructions > Simulated > ….. > > /*********************************************************************/ > > Well, it seems that only cpu3 reached 100 million instruction counts. > And I think the above statistic suggested that the termination of cpu3 > cause other cpu terminated(may be the thread in cpu3 was run faster). > > My question is , if I want each thread run until reaching the maximum > instruction count I set, how should I do ? > > Thanks in advance! > > York > > > > > > _______________________________________________ > m5-users mailing list > [email protected] > http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
_______________________________________________ m5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
