Commit graph

  • e6fa9052b3 Add NPU offload results: offload_op, direct-compute, 10GB, Q8_0 4.3x prompt speedup main Jimmy Devine 2026-05-02 14:17:27 +0200
  • 627236a505 Update with full NPU analysis and benchmarks Jimmy Devine 2026-05-02 12:42:42 +0200
  • 18970e3258 Initial commit: Q6A Hexagon v68 + llama.cpp guide Jimmy Devine 2026-05-02 10:28:51 +0200