Those benchmarks run the program once from a cold start, right? When people talk about JS nearing c++ performance I think they generally mean cases like the ste

Those benchmarks run the program once from a cold start, right? When people talk about JS nearing c++ performance I think they generally mean cases like the steady-state performance of a game loop, after the JS engine has had a chance to observe the hot code for a little while.