3 ms·
It says at the top: > write state of the art kernels Mojo seems to be competing with C++ for writing kernels. PyTorch and Julia are high-level languages where
by jb1991 1y ago
It says at the top:
> write state of the art kernels
Mojo seems to be competing with C++ for writing kernels. PyTorch and Julia are high-level languages where you don't write the kernels.
- Alexander-Barth 1y agoActually in julia you can write kernels with a subset of the julia language: https://cuda.juliagpu.org/stable/tutorials/introduction/#Writing-your-first-GPU-kernel https://cuda.juliagpu.org/stable/tutorials/introduction/#Wri... With KernelAbstractions.jl you can actually target CUDA and ROCm: https://juliagpu.github.io/KernelAbstractions.jl/stable/kernels/ https://juliagpu.github.io/KernelAbstractions.jl/stable/kern... For python (or rather python-like), there is also triton (and probably others): https://pytorch.org/blog/triton-kernel-compilation-stages/ https://pytorch.org/blog/triton-kernel-compilation-stages/
- davidatbu 1y agoChris's claim (at least with regards to Triton) is that it avails 80% of the performance, and they're aiming for closer to 100%.
- jakobnissen 1y agoI think Julia aspires to be performant enough that you can write the kernels in Julia, so Julia is more like Mojo + Python together. Although I have my doubts that Julia is actually willing to make the compromises which would allow Julia to go that low level. I.e. semantic guarantees about allocations and inference, guarantees about certain optimizations, and more.
- pjmlp 1y agoYou can write kernels with Python using CUDA and Open API SDKs in 2025, that is one of the adoption problems regarding Mojo.