twill.pdf
compilers.stanford.edu · 6,726 words · saved by 1 readers
N/A
Optimal Software Pipelining and Warp Specialization for Tensor Core GPUs Rupanshu Soi∗† Rohan Yadav∗† Fredrik Kjolstad Alex Aiken† Stanford University Stanford University Stanford University Stanford University Maryam Mehri Dehnavi Michael Garland Michael Bauer…
saved by
related reading
- How to Optimize a CUDA Matmul Kernel for cuBLAS-like Performance: a Worklogsiboehm.com
- Inside NVIDIA GPUs: Anatomy of high performance matmul kernels - Aleksa Gordićaleksagordic.com
- Robust Code Generation of Efficient Tensor Computation Using 𝔽_2arxiv.org
- AI Chip Architecturesjacobpeake.com
- Tiny TPUtinytpu.com
- Modern GPU Programming For MLSys — Modern GPU Programming For MLSysmlc.ai
- How to Think About GPUs | How To Scale Your Modeljax-ml.github.io
- AMD GPUs go brrr · Hazy Researchhazyresearch.stanford.edu
- Understanding and expressing scalable concurrencyaturon.github.io
- We reverse-engineered Flash Attention 4modal.com
- Grouped GEMM for Imbalanced Experts on Blackwell: A WIP Workloggauravjain.bearblog.dev
- Microsoft Word - CACM_roofline_complete.docusers.cs.duke.edu