Hi,
I'm curious if there is a more recent version of the July 2016 paper "Renewed case for RISC : Avoiding ISA bloat ..." showing a comparison between RISCV and ARMv8 of dynamic instruction counts for SPEC2017 workloads (Table II in that paper) with latest ISA extensions.
Can someone share a recipe for SPEC2017 with GCC version, ISA extensions, etc. ? I see significant bloat on x264 in particular, which I am unable to explain.
thanks,
/mehul