LLM training in simple, raw C/CUDA

📚LeetCUDA: Modern CUDA Learn Notes with PyTorch for Beginners🐑, 200+ CUDA Kernels, Tensor Cores, HGEMM, FA-2 MMA.🎉

DeepEP: an efficient expert-parallel communication library

[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.

Mirage Persistent Kernel: Compiling LLMs into a MegaKernel

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

Several optimization methods of half-precision general matrix multiplication (HGEMM) using tensor core with WMMA API and MMA PTX instruction.

CUDA Matrix Multiplication Optimization

🤖FFPA: Extend FlashAttention-2 with Split-D, ~O(1) SRAM complexity for large headdim, 1.8x~3x↑🎉 vs SDPA EA.