4 ms·
The GPU prorgramming language tech landscape is generally pretty low-tech on the compilers side. Tensor cores are not supported. (Culturally of course the big
by fulafel 2y ago
The GPU prorgramming language tech landscape is generally pretty low-tech on the compilers side.
Tensor cores are not supported.
(Culturally of course the big one is the fragmentation and the proprietary nature of everything which is the reason so little gets done on GPUs and the horror of attempting multiplatform software there)
- creata 2y ago> Tensor cores are not supported. Correct me if I'm wrong or misunderstanding you, but doesn't SPIR-V support tensor cores via SPV_KHR_cooperative_matrix?
- fulafel 2y agoI don't know, do share if there's something you can link. At first blush it just sounds like something to allow multiple shader compute elements to work more efficiently together ("cooperate") on a single bigger matrix computation.
- creata 2y agoThat is what it is afaict, but NVIDIA says this in their 2019 post "Machine Learning Acceleration in Vulkan with Cooperative Matrices"[0]: > Additionally, if the GPU includes dedicated hardware for high-speed matrix operations, such as the Tensor Cores on Turing GPUs, then the Cooperative Matrix extension can tap into the power of this acceleration with no application changes. The benchmark graph doesn't look too great, though - around half the "theoretical peak tensor core performance". [0]: https://developer.nvidia.com/blog/machine-learning-acceleration-vulkan-cooperative-matrices/ https://developer.nvidia.com/blog/machine-learning-accelerat...