Jim Blandy

Do the benchmarks not run directly on macOS? I amended the README to say that I don't really understand why pinning to a single core speeds up thread-brigade. I mean,...

Wow, they're about the same. And, if I may continue to impose, how about one-thread-brigade, to see how much time is due to the I/O alone?

It seems like GLSLang creates temporaries, which the caller initializes before the call and copies out of after the call. The callee only sees pointers to local variables. It seems...

> This is a whole-program analysis, which is what we've considered and rejected previously. Alternatively, we could analyze each function to compute, for each pointer-typed parameter, which globals it might...

With `atan2(x, y)`, you can always just swap x and y to ensure that x is bigger (in absolute value). Then `y/x` is always in `[-1,1]`, single-argument `atan` works well,...

Today's meeting minutes will be attached here soon, but just to summarize: It may make more sense to flatten out some cases than others. (By 'flatten', I mean change `@foo(bar)`...

> Fundamentally, I don't think there's anything wrong with holding things by closure. In Chromium, we can do this pretty literally, with a callback object that holds a reference to...

> If you tried to transfer a getMappedRange-created ArrayBuffer, it would appear to work entirely normally - old ArrayBuffer would get detached - except the underlying data would get copied...

> - If the invalid constexpr expression involves an out-of-bounds access, then it should be a compile-time error. The author is bad and they should feel bad. > - Otherwise...

Here's one approach we might take for the math functions whose domain is not all of `f32`, that provides a hook on which to hang both NaN handling and early...