2021-10-20 03:42:17 <veltas      > MrMobius: albeit looking at numpy I don't see many intrinsics lol
2021-10-20 03:43:52 <veltas      > I just picked numpy out of thin air, I assumed it was properly optimised but I guess maybe not
2021-10-20 08:35:48 <MrMobius    > veltas, by intrinsics do you mean telling it to use vector operations? im sure you know gcc and similar will use SIMD all by itself if it can determine from the code that it's appropriate although it can't always tell
2021-10-20 08:37:38 <MrMobius    > but in any case, "optimization" is not just vectorizing. a lot of programs don't need that
2021-10-20 17:29:27 <veltas      > MrMobius: It usually can't do it because it requires higher alignment, there are some extensions that make it more automatic but I don't know how reliable that is
2021-10-20 17:29:55 <veltas      > Higher alignment but also it's just not a very developed part of the compiler
2021-10-20 17:30:04 <veltas      > It's very good at certain types of optimisations, not these ones
2021-10-20 21:51:43 <remexre     > huh, looks like I might be back in forth-land soon
2021-10-20 21:53:04 <remexre     > turns out wasm is a fairly painful compile target for a GC'd language, so I think I might do a Forth instead (probably a noninteractive one, because it's terribly inefficient to write new code at runtime still)