List, suppose you have configured (caches parameters and memory size) gem5 to model a real X86 CPU. running a benchmark gives us some statistics. Running this on the real CPU will also gives us some stats. Is it possible/logical/acceptable (in computer architecture research community) to compare these two results? If I want to show (in my paper, for example) that i'm actually simulating a system with similar properties as the real CPU, how can I judge it? Suppose I want to add a new feature to my simulated CPU later to show that this feature could improve a performance measure.
Best Seyed Hossein
_______________________________________________ gem5-users mailing list gem5-users@gem5.org http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
