DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

Tile primitives for speedy kernels

[ICLR2025, ICML2025, NeurIPS2025 Spotlight] Quantized Attention achieves speedup of 2-5x compared to FlashAttention, without losing end-to-end metrics across language, image, and video models.

how to optimize some algorithm in cuda.

[ICML2025] SpargeAttention: A training-free sparse attention that accelerates any model inference.

PyTorch bindings for CUTLASS grouped GEMM.