CUDA & C++ 20 Compatibility, DLPack, Float8 & 3D Mesh Speedups#349

Merged

ashvardanian merged 8 commits into

mainfrom

Apr 20, 2026

Commits on Apr 15, 2026

Commits on Apr 16, 2026

Improve: SSD trace-identity fold across all mesh backends + Genoa/NEONFHM kernels
ashvardanian
committed
Make: Enable symbol exports for `nk_shared` Emscripten builds
ashvardanian
committed

Commits on Apr 18, 2026

Commits on Apr 19, 2026

Improve: FP8 pairwise distance kernels via Giesen trick + F16 widen path
ashvardanian
committed

Commits on Apr 20, 2026

Improve: FP8 GEMM throughput on Skylake/Haswell + Granite Rapids E5M2 kernel
ashvardanian
committed