Hi Dave. I assume you've made some changes to the simulator? Can you tell if this is coming from gem5 or from the benchmark? If it's from gem5, I'd guess you've got a memory leak in your modifications. You can probably see if that's happening by watching it run under top, and if it is track it down with valgrind. We have a file for surpressing false alarms in valgrind (it gets confused by some of the python stuff) in util called valgrind-suppressions, I believe. I think the option to pass it to valgrind is --suppressions.
Gabe On 07/06/11 07:48, David Roberts wrote: > Hi guys, hows it going? > > Quick question - Does the pre-compiled set of SPLASH-2 binaries on the gem5 > site work with the detailed mode CPU model in SE mode? Running just with > Timing mode has no problem. I allocate 1GB of PhysicalMemory space. I run > with the following command line; > > build/ALPHA_SE/m5.opt configs/splash2/run.py --benchmark Cholesky --rootdir > splash2/codes -n 2 --detailed -f 1500MHz --l1size 64kB --l2size 256kB > --l1latency 1ns --l2latency 10ns > > After some time it crashes with: > > terminate called after throwing an instance of 'std::bad_alloc' > what(): St9bad_alloc > Program aborted at cycle 1550644939002 > > Any ideas? > > Cheers > Dave > _______________________________________________ > 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
