We don’t properly handle all the arithmetic faults (e.g. divide by 0) in gem5 because correct code normally doesn’t do that. There are two options. (1) xanalcbmk requires faulting handling (2) a bug somewhere else in gem5 is causing a value to be 0 when it otherwise shouldn’t be.
Thanks, Ali On Dec 20, 2013, at 5:42 PM, Yuhang Liu <168liuyuh...@163.com> wrote: > Hi, > > I encounter the error as following when I run spec cpu2006 benchmark > xalancbmk on gem5, I found the error someone encontered before in gem5 > mailing list but it was not solved. Someone can give me some advice? > > info: Increasing stack size by one page. > warn: Prefetch instructions in Alpha do not do anything > warn: sqrtt/su f16,f0: non-standard trapping mode not supported > panic: fault (arith) detected @ PC (0x1204374c0=>0x1204374c4) > @ cycle 20385000 > [invoke:build/ALPHA/sim/faults.cc, line 48] > Memory Usage: 8518940 KBytes > Program aborted at cycle 20385000 > ./xalancbmk_test.runscript: line 1: 32746 Aborted > > > Thanks in advance, > John > > > _______________________________________________ > gem5-users mailing list > gem5-users@gem5.org > http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
_______________________________________________ gem5-users mailing list gem5-users@gem5.org http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
