3 ms·
Anecdotally, the Julia approach certainly seems... nicer via e.g. KernelAbstractions etc. I don't know if the performance/flexibility is quite there vs. native
by empyrrhicist 3y ago
Anecdotally, the Julia approach certainly seems... nicer via e.g. KernelAbstractions etc. I don't know if the performance/flexibility is quite there vs. native CUDA, but going from zero to GPU kernel programming in Julia is as close to painless as I've ever seen, especially given how modular everything is (for example, you can use OffsetArrays.jl directly in GPU kernel code).