Carson Poole

haystackdb. Rust

safetensors.cpp. Zero Dependency LibTorch Safetensors Loading and Storing in C++

megablock. Mega🅱️lock

quadmul. a fast and customizable CUDA int4 tensor core gemm

gpus. GPU Performance and Cost Cheatsheet

ffvec. C

haystack-embedded. Rust

octomul. Reasonably fast (compared to cublas) and relatively simple int8 tensor core gemm

cuda-utils. CUDA utilties/helpers for simplifying multidimensional indexing

TF2-HRNet. Python

Paul-Graham-Essays.

octoquadmul. Cuda

compress-zip. Rust

markdown-ptx-isa. TypeScript