Sometimes people want to measure performance of the cartesi machine in their computer, to know in advance the speed of execution for some code that might be executed inside a dapp, so that's why I introduced the --bench-cpu command line option, to assist on this.
Furthermore I added a new script in tools/benchmarks/bench-insns.lua to benchmark individually the performance every RISC-V instruction we implement (more than 150+), this script assists me optimizing the cartesi machine, and might be useful for others to consult the overhead of each instruction. Note this script is mostly to give a hint of instructions raw performance, it does not cover worst case scenarios and address translation overhead.
Sometimes people want to measure performance of the cartesi machine in their computer, to know in advance the speed of execution for some code that might be executed inside a dapp, so that's why I introduced the
--bench-cpu
command line option, to assist on this.Furthermore I added a new script in
tools/benchmarks/bench-insns.lua
to benchmark individually the performance every RISC-V instruction we implement (more than 150+), this script assists me optimizing the cartesi machine, and might be useful for others to consult the overhead of each instruction. Note this script is mostly to give a hint of instructions raw performance, it does not cover worst case scenarios and address translation overhead.This is the output of the features this PR adds: