Twan Koolen
Twan Koolen
Yeah, this is the kind of thing that we're giving up by not basing off of Cassette. With Cassette, you could perform the optimizations directly on the code of e.g....
Retiring the package sounds good to me.
Thank you!
Yeah, strange. I guess IJulia is using a different build script underneath though, maybe because it still supports 0.6. I see ```bash echo "%JL_BUILD_SCRIPT%" "using Pkg; Pkg.build()" set PYTHON="" set...
Yeah it is. Maybe disable 32 bit here for now, I'll fix that later.
Sorry if this is a little unrelated to the question in the issue description, but on the topic of deprecations, I was wondering whether it would be a good idea...
Presumably a duplicate of #193, but good to have another data point.
> Do you have benchmarks for small vectors and matrices? It's important (to me at least) that we don't slow down any sums or norms or dot products of 2-vectors...
I think the way `_vecnorm` is currently implemented, compile time does not scale linearly, but it could of course.
Ah, I guess you're right. I thought that https://github.com/JuliaArrays/StaticArrays.jl/blob/7ddb8e4a02656d0816a53a6d819e1384923132e5/src/linalg.jl#L287-L289 would allocate a new `Vector` containing `:+` and its (growing number of) arguments at each iteration (as the `args` field of...