I am running with this kernel and configuration:
./build/ALPHA/gem5.opt configs/example/fs.py \
--num-cpus=16 \
--cpu-type=DerivO3CPU \
--caches --l2cache \
--l2_size=512kB \
--l2_assoc=16 \
--num-l2caches=1 \
--mem-size=1GB \
--maxtime=2 \
--cpu-clock='1GHz' \
--take-checkpoints=0,100000000000 \
--script=run_scripts/dedup_16c_simsmall.rcS \
--kernel=/home/fur/gem5/full_system_images_ALPHA/binaries/vmlinux_2.6.27-gcc_4.3.4 \
--checkpoint-at-end \
--max-checkpoints=2
Should I change the kernel file (vmlinux_2.6.27-gcc_4.3.4) for each workload (i.e. bodytrack, canneal, dedup, etc.) or it's the same for all workloads and, so, I should change script file only?
For each running, this gives me the same hits and miss numbers for all workloads.