Matthew Nicely

icon company NVIDIA icon location Huntsville, AL icon location DL Compiler Product Manager [email protected]

Results 113 comments of Matthew Nicely

Python Wrapper ?

@Millow What seems to be your issue? I've been using https://github.com/rayhou0710/pycudasift for some time now with no issues.

`cub::ThreadLoadAsync` and friends, abstractions for asynchronous data movement

@brycelelbach If it is possible to pass a pointer within a SMEM array, will `cub::ThreadLoadAsync` be smart enough to convert the instructions to `LDGSTS`?

CUDA Support

Has there been anymore progress on this front?

CUDA Support

Axel, Thanks for the feedback. I'm trying to write a notebook using OpenMP, OpenACC, BLAS, cuBLAS, and CUDA. Just started this morning but I was able to get cuBLAS to...

[QST] Request useful script shown in GTC

Added to 2.11

[QST] Datatype conversion for tensor on GPU

Closing due to inactivity. Please reopen if needed.

[QST] How to efficiently implement 16bit convolution

Closing due to inactivity. Please reopen if needed.

Cutlass

Hi! Very excited to CUTLASS benefit your framework. Based on the dates, I'm assuming you're using v2.5? I wanted to let you know that we have added a plethora of...

Are there any examples of mixed input Gemm (A type fp16, B type fp8)?

@liligwu has you resolved your issue?

[QST] Support for Sparse Tensor Operations in CuTe

Structured sparsity through CuTe is planned for a later release.