Hi everyone, I am trying to use the elastic trace solution proposed by ARM.
1. I create a checkpoint with the command line: ./build/ARM/gem5.opt -d m5out/traces configs/example/fs.py --machine-type VExpress_EMM --kernel ~/Big_LITTLE/arch/binaries/linux-arm-gem5/vmlinux --disk-image ~/Big_LITTLE/arch/disks/linux-aarch32-ael.img --caches --mem-size 256MB -n 1 --script ~/Big_LITTLE/script/rodinia/backprop.sh ----- backprop.sh ----- #!/bin/sh /sbin/m5 checkpoint cd /benchmark/bin/ echo "Reset stats and start $(basename $0 .rcS)" /sbin/m5 resetstats echo "Job launched" ./backprop 65535 1 echo "Job done" /sbin/m5 exit ----------------------- 2. I collect the traces with: ./build/ARM/gem5.opt -d m5out/traces configs/example/fs.py --machine-type VExpress_EMM --kernel ~/Big_LITTLE/arch/binaries/linux-arm-gem5/vmlinux --disk-image ~/Big_LITTLE/arch/disks/linux-aarch32-ael.img --caches --mem-size 256MB --cpu-type arm_detailed --checkpoint-dir m5out/traces -r 1 -n 1 --script ~/Big_LITTLE/script/rodinia/backprop.sh --elastic-trace-en --mem-type SimpleMemory --inst-trace-file fetchtrace.proto.gz --data-trace-file deptrace.proto.gz But when I launch this command gem5 panic: ... panic: Packet queue system.cpu.dcache.cpu_side-CpuSidePort has grown beyond 100 packets @ tick 2797832682500 [schedSendTiming:build/ARM/mem/packet_queue.cc, line 120] Memory Usage: 716364 KBytes Program aborted at tick 2797832682500 ... I have seen in the code that the 100 packets limit is put "to ensure that the buffer is not growing beyond reasonable limits". So I am wondering : is there any error with my command, or should I modify the code by my own to exceed this limit? Thanks +-------------------------------+ | Louisa Bessad | | PhD student - LIRMM - Sysmic | | Bâtiment 4 Bureau H2.2 | +-------------------------------+ _______________________________________________ gem5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
