cudnn-frontend
cuDNN Frontend is NVIDIA's modern, open-source entry point to the cuDNN library and a growing collection of high-performance open-source kernels.
- 1k
- Cuda
- Apache-2.0
- 3 h ago
cuDNN Frontend is NVIDIA's modern, open-source entry point to the cuDNN library and a growing collection of high-performance open-source kernels.
Implementation and examples from Trajectory Optimization with Optimization-Based Dynamics https://arxiv.org/abs/2109.04928
Cross-platform FlashAttention-2 Triton implementation for Turing+ GPUs with custom configuration mode
From-scratch, heavily-annotated CUDA inference runtime for Qwen2.5-Coder-7B on H100 (sm_90). Custom INT4 packer, fused GEMV, paged KV, split-KV attention, CUDA graph decode — every hot path commented for the why. Educational, not a llama.cpp replacement.
Faster attention kernels for serving TML's Inkling model on vLLM. 2.7x over the shipping path on H100, and the only implementation that runs on A100.
Code from "How useful is quantilization for mitigating specification-gaming?"
A single executable notebook showing how to use turbopairformer, developed by Lambda, Stevens Institute of Technology, and OpenFold — standalone CUDA kernels for the two most expensive operations in an AlphaFold3-style pair stack.
Portable, evidence-first AI-agent skill for lawful reverse engineering with REA CLI/MCP workflows.