Hi, Remember that in full-system runs a lot of non-linear behaviours come into play. Even small perturbations in timing can suddenly lead to a change in scheduling, an interrupt being delivered in the middle of a mutex operation, etc. Changing CPU models, changing frequency, or even just delaying the start of an experiment (by inserting call to sleep) can easily change the instruction count by many 10’s of percents. Check what is happening in your runs, check again, and preferably run more than one experiment so that you can get some statistical rigor.
Andreas From: gem5-users <[email protected]<mailto:[email protected]>> on behalf of Ayaz Akram <[email protected]<mailto:[email protected]>> Reply-To: gem5 users mailing list <[email protected]<mailto:[email protected]>> Date: Tuesday, 23 February 2016 at 06:20 To: gem5 users mailing list <[email protected]<mailto:[email protected]>> Subject: [gem5-users] BBV generation through gem5 I am generating BBV file using gem5 for a benchmark in full system mode. Surprisingly the number of simulated instructions when gem5 is run for BBV generation is very different when I run the same benchmark on gem5 in detailed mode. I am running benchmark in both cases from a checkpoint and the instruction difference that I see right now is approx. 20 % (more in BBV generation run). Did anyone has a similar experience ? IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
_______________________________________________ gem5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
