Dear all, I have followed my question and more info to debug it. As I found at last time, this panic is caused by changing the cache line size. After I searched the gem5 web, I found a relative reason: "This is only used for two cases in readIpr: when a write-only register is accessed, and when the index doesn't match any IPR. In those cases it returns an UnimplementedOpcodeFault."[http://gem5.org/Alpha_Dependencies]
But I have no idea what's the relationship between this illustration and my change of cache line size. Can you tell me which part of codes I should read? Thanks! Chao Zhang Peking University ----- "Chao Zhang" <[email protected]> ： > Dear all, > > I'm running parsec in FS of stable version, but faced with a fault > when I just change the cacheline_size to 32 from the default, 64. > The fault is shown as following. It works well with my configuration > when the size is 64. > Anyone has met this problem? > > **** REAL SIMULATION **** > info: Entering event queue @ 0. Starting simulation... > warn: Prefetch instructions in Alpha do not do anything > 2947604000: system.terminal: attach terminal 0 > warn: Prefetch instructions in Alpha do not do anything > panic: attempt to execute unknown instruction (inst 0x800e4c00, opcode > 0x20) > @ cycle 2240745367500 > [execute:build/ALPHA/arch/alpha/generated/atomic_simple_cpu_exec.cc, > line 87] > Memory Usage: 320212 KBytes > Program aborted at cycle 2240745367500 > > My cmd used: > > /home/chao/Research/gem5_stable/build/ALPHA/gem5.opt > --outdir=/home/chao/Research/gem5_stable/PF_parsec > --stats-file=stats_32_parsec.txt > /home/chao/Research/gem5_stable/configs/example/fs.py > --cpu-type=atomic --caches -n 1 --script=configs/parsec/test.rcS > --cacheline_size=32 > > Chao Zhang > Peking University _______________________________________________ gem5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
