I am using the codebase at http://repo.gem5.org/gem5-stable and was trying to use the prefetcher for a simulation of the miniMD app (https://software.sandia.gov/mantevo/download.html) with different problem sizes. The miss rates for the L2 were looking unrealistic(>70%) and hence i turned on the prefetching. But even with a very small problem size of 5x5x5 and 30 iterations in X86_SE mode, a prefetch degree (stride) of >=2 gives the wrong result for the benchmark.
-Mitchelle. _______________________________________________ gem5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
