> Note that some of the examples can run even faster with Shed Skin by providing --nowrap/--nobounds options, which disable wrap-around/bounds-checking respectively. I'm not sure if PyPy has any options to make it run faster, at the cost of certain features (in the distant past there was talk of RPython - does that still exist?).
My understanding: RPython was never meant for that sort of thing. Its purpose is to make the bootstrapping process possible; it isn't really a separate thing, it just represents the features of Python that they don't use in the core implementation, so that the core implementation can translate itself to C. Much like the limitations on what Shedskin can accept for compilation. Once the implementation is bootstrapped, it's unclear whether restricting end users to RPython could help squeeze out any more JIT performance; it doesn't appear that they're interested in trying.
zahlman•1mo ago
My understanding: RPython was never meant for that sort of thing. Its purpose is to make the bootstrapping process possible; it isn't really a separate thing, it just represents the features of Python that they don't use in the core implementation, so that the core implementation can translate itself to C. Much like the limitations on what Shedskin can accept for compilation. Once the implementation is bootstrapped, it's unclear whether restricting end users to RPython could help squeeze out any more JIT performance; it doesn't appear that they're interested in trying.