KernelAbstractions.jl
KernelAbstractions.jl copied to clipboard
JuliaGPU
Heterogeneous programming in Julia
Alternative to #399 I am still thinking if a tiled iteraion like approach wouldn't be better. Explicitly only support edge iteration.
From a short conversation with @vchuravy I learned that there is currently no way of selecting between different GPUs by the same vendor on a single host using KernelAbstractions. It...
It'd be extremely useful to have a feature for offsetting the kernel index, either at launch time or setup time. Not sure what the best design or interface would be....
I want to build a standalone module that can run on any supported GPU. How do I detect what packages need to be loaded so that I can have a...
I noticed DiffEqGPU doing a plain `error()` in GPU code: https://github.com/SciML/DiffEqGPU.jl/blob/dddcb594ce054c0677bc1b18fdabca2fc0c2eaa9/src/perform_step/gpu_tsit5_perform_step.jl#L152 That's of course not great, and leads to inscrutable errors: ``` ERROR: a exception was thrown during kernel execution....
Hi! I just tried to get the backend of a CuArray as follows, which causes a stack overflow. Am I using it correctly? ```julia using CUDA using KernelAbstractions get_backend(cu(rand(3))) ERROR:...
In discussion with @maleadt, we came up with the idea to solve user code like: ``` kernel(...) @sync for i in 1:4 @spawn begin kernel(...) # can race with previous...
consider this example from a Jax discussion: Source code ```julia function run_julia(height, width) y = range(-1.0f0, 0.0f0; length = height) # need Float32 because Jax defaults to it x =...
Hi! I'm trying to use RuntimeGeneratedFunctions (via MTK) on CUDA GPUs and I'm encountering the following issue MWE ```julia using RuntimeGeneratedFunctions RuntimeGeneratedFunctions.init(@__MODULE__) function foo() expression = :((du, u, p, t)...
