Bing Xu

Before fully investigation, I think we need to change： - [ ] base.h - [ ] stream_gpu-inl.h - [ ] tensor_gpu-inl.h - [ ] reduce.cuh - [ ] tensor_gpu-inl.cuh

I think we'd also make change to cl memory/ buffer stuff, which I need to do more investigation. On Mon, Nov 2, 2015 at 11:03 Tianqi Chen [email protected] wrote: >...

I just pushed a quick fix for CUDA 6.5. I don't have CUDA 7.0 installed so please help us check again. Thanks.

Yes Python runtime is just a simple wrapper of C++ runtime. You can use the generated model with C API directly. This note is helpful: https://github.com/facebookincubator/AITemplate/tree/main/static And the Model class...

cc @mikeiovine

Many examples are not working with T4/V100, such as deterctron2 and stable diffusion, this is why we directly blocked V100 and T4. Another reason is that CUTLASS focus is shifted...

@HamidShojanazeri Thanks for suggestion. Given our team size and our workloads on supporting internal production needs, we don't have bandwidth to enable V100/T4. If community/NVIDIA is going to help on...

I don’t have V100 access, will try to find one and make the list. On Mon, Oct 24, 2022 at 18:45 Ehsan Azar ***@***.***> wrote: > @antinucleon Is there a...

@terrychenism maybe we can prioritize it and make a v0.11/v0.12 release.

AITemplate is coming from Meta production needs, we don't have T4/V100 so in our first release we didn't consider about this. We will help to pass the voice to NVIDIA...