NVIDIA/cutlass
View on GitHub ↗ ★ 10,684 C++
CUDA Templates and Python DSLs for High-Performance Linear Algebra
19 mentions on Hacker News by 16 people: 5 stories, 14 comments · first 2017-12-14 · latest 2026-01-02
20082026
-
I'm telling your it works now. It's just not called `tcgen05`. Put this in nsight compute: https://github.com/NVIDIA/cutlass/blob/main/examples/79_blac... (I said 83, it's 79). If you want to know what NVIDIA really thinks, watch this…
comment by benreesman · 2026-01-02 · in Dell's version of the DGX Spark fixes pain points -
…of cuda.) This ticket is a good starting place to see the chain of issues around the ongoing work: https://github.com/NVIDIA/cutlass/pull/2037
comment by spoaceman7777 · 2025-07-11 · in FP8 is ~100 tflops faster when the kernel name has "cutlass" in it -
Github version: https://github.com/NVIDIA/cutlass I wonder if we search the comments if we can find something referencing this.
comment by gpm · 2025-07-11 · in FP8 is ~100 tflops faster when the kernel name has "cutlass" in it -
It is a law of "AI" that every project must have a name collision with Nvidia: https://github.com/NVIDIA/cutlass
comment by Y_Y · 2025-06-29 · in Show HN: A different kind of AI Video generation -
CuTe DSL examples from the MIT Dao-AILab for writing high performance cuda kernels using Python (see https://github.com/NVIDIA/cutlass for more background info).
comment by musebox35 · 2025-06-23 · in QuACK: A Quirky Assortment of Cute Kernels -
…e.g. the closed-source cuBLAS. Keen observers will notice that Strassen is not used in CUTLASS. [0] https://github.com/NVIDIA/cutlass
comment by david-gpu · 2024-12-12 · in The GPU is not always faster -
…Also checked and apparently Nvidia Cutlass now supports generic convolutions: https://github.com/NVIDIA/cutlass
comment by mratsim · 2024-07-20 · in Memory and ILP handling in 2D convolutions -
…( https://people.maths.ox.ac.uk/gilesm/cuda/ ). Also explore the tutorial section of cutlass ( https://github.com/NVIDIA/cutlass/blob/main/media/docs/cute/... ) if you want to learn more about high performance gemm. OpenAI triton is…
comment by yzh · 2023-08-09 · in Optimization Techniques for GPU Programming [pdf] -
…and managed to get back fp32 accuracy with 3 rounds of the things Hey are you referring to 3xTF32 ( https://github.com/NVIDIA/cutlass/tree/master/examples/28_am... )? IMO this is a perfect example where proper abstraction could save…
comment by junrushao1994 · 2023-03-23 · in AI’s compute fragmentation: what matrix multiplication teaches us -
…far without interchanging the innermost loop over k to the outermost loop, as is done in CUTLASS ( https://github.com/NVIDIA/cutlass ). As you can see in this blog post the code ends up with a lot of compile-time constants (e.g. BLOCKSIZE,…
comment by gpuhacker · 2023-01-05 · in How to Optimize a CUDA Matmul Kernel for CuBLAS-Like Performance: A Worklog -
Yup, in a semi-related field, NVIDIA has 3xTF32 for cases needing higher precision: https://github.com/NVIDIA/cutlass/discussions/361
comment by my123 · 2022-03-22 · in Nvidia Hopper GPU Architecture and H100 Accelerator -
…is _only_ 2X faster than the SIMT path. I'll leave the prefix sum for other devs who know more :D https://github.com/NVIDIA/cutlass/blob/master/examples/27_am... //part of nod.ai/shark team
comment by magic_at_enimai · 2022-02-23 · in PyTorch on Apple M1 MAX GPUs with SHARK – faster than TensorFlow-Metal -
…sent to Tensor Cores by the scheduler. The cutlass library is the open source version of cublas: https://github.com/NVIDIA/cutlass
comment by rubatuga · 2021-05-02 · in AMD Reports Q1 2021 Earnings -
…and it's not about C vs C++ vs PTX. Or let's hear about what Nvidia says about optimizing GEMM: https://github.com/NVIDIA/cutlass/blob/master/media/docs/eff... , it's all about memory locality and tiling. Or maybe Stanford, the US…
comment by mratsim · 2020-04-22 · in A possible new back end for Rust