GPUArrays.jl
GPUArrays.jl copied to clipboard
Adapt to new LinearAlgebra.generic_*mul! interface
This is a companion PR to https://github.com/JuliaLang/julia/pull/52439. There, we avoid constructing MulAddMul objects, which are, however, used here in the most generic multiplication kernel. Along the way, I rearranged the call chain slightly to avoid unpacking already existing MulAddMul objects, just to reconstruct it within the multiplication kernel.