The atomic and timing CPUs count instruction in slightly different ways. Ali
On Jul 25, 2011, at 2:36 PM, xxx wrote: > Hi, > I run a benchmark (openjpeg) on ARM_SE, with two kinds of configurations. > > The first configuration is as the following: > ./build/ARM_SE/gem5.opt configs/example/se.py --cmd > ./benchmark/image_to_j2k --options="-i ./benchmark/Bretagne1.bmp -o > ./benchmark/Bretagne1.j2k" > > The second configuration is as the following: > ./build/ARM_SE/gem5.opt configs/example/se.py -d -t --caches --cmd > ./benchmark/image_to_j2k --options="-i ./benchmark/Bretagne1.bmp -o > ./benchmark/Bretagne1.j2k" > > I have two questions about the result in stats.txt. > 1. The number of instruction under the first configuration is 664046521, and > the number of instruction under the second configuration > is 653963923. I run the same benchmark, why the instruction number is > different? > > 2. Under both of the configurations, I find that the number of idle cycle is > is 0. I don't know why the number of idle instruction is 0. Do I miss some > configuration options? > > > Thanks. > Dan > > > > _______________________________________________ > gem5-users mailing list > [email protected] > http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users _______________________________________________ gem5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
