#cuda
3
2026-10-28
johnfink8/sfnx
B
CUDA engine that splits every layer of a production quality MoE LLM across three PCIe V100 cards delivering 382 tok/s on a used $1,800 box.
2026-10-12
kindlingai/gr8
C
GLM-5.3 (full) on DGX Spark (GB10): g6, a from-scratch CUDA inference engine
2026-11-09
lucialuka/OpenAssembler
unrated
Open source Python/CUDA C/C++ to SASS/CUDA Binary compiler (previously closed source by nvidia), more efficient (3-12% speedup) General Matrix Multiplication (GEMM)