All pull requests

Pull requests

• mtmd : add decode limits to mtmd-helper

  mtmdtesting
  #429·dev-nid opened Oct 23, 2026·

  ·

• ggml : TurboVec x86 AVX-512 search, x86 rules and perf review fixes (6/6)

  documentationggmltesting
  #428·dev-nid opened Oct 23, 2026·

  ·

• ggml : TurboVec SMMLA batches over the vm8 layout and the 2-bit vm8 opt-in (5/6)

  documentationggmltesting
  #427·dev-nid opened Oct 23, 2026·

  ·

• ggml : TurboVec SIMD rotation, encode and CRC-32C, worker pool and threaded passes (4/6)

  ggmltesting
  #426·dev-nid opened Oct 23, 2026·

  ·

• QVAC-27327 feat: add Core ML acceleration for EmbeddingGemma 2

  builddocumentationserver
  #425·ishanvohra2 opened Oct 23, 2026·

  ·
  3

• skills : add engine-bench for the shared Tailscale machines

  documentation
  #424·amangupta-tether opened Oct 23, 2026·

  ·
  1

• infra[notask]: pin setup-node to lts/* with check-latest and setup-bun to latest

  devops
  #423·Proletter opened Oct 22, 2026·

  ·
  1

• Rebase b11739

  AMD ZenDNNandroidApple MetalAscend NPUbuildconversionCUDAdevopsdocumentationggmlHexagonIBM zDNNmodelmtmdOpenCLOpenVINOserver/uiserverSYCLtestingvendorVulkanWebGPU
  #421·gagallo7 opened Oct 22, 2026·

  ·
  3

• QVAC-26706: complete Hexagon HTP speech-op port (part 7)

  devopsdocumentationggmlHexagontesting
  #417·mexxik opened Oct 21, 2026·

  ·
  18

• ggml, llama : deferred gated delta net commit for temp-11259, with CUDA and Vulkan (port of #410)

  Apple MetalCUDAdevopsdocumentationexamplesggmlHexagonmodelOpenCLtestingVulkanWebGPU
  #414·amangupta-tether opened Oct 21, 2026·

  ·
  20

• hip : experimental W4A4 Q4_0 matmul with q4_0 activations (DOT8 and WMMA IU4)

  CUDAggmltesting
  #412·amangupta-tether opened Oct 21, 2026·

  ·

• ggml : TurboVec aarch64 permute-dot, staged search, single code copy and audit fixes (3/6)

  devopsdocumentationggmltesting
  #407·dev-nid opened Oct 20, 2026·

  ·

• ggml : TurboVec v8 snapshots and sync, search threads, storage/ids API and swap-remove (2/6)

  documentationggmltesting
  #405·dev-nid opened Oct 20, 2026·

  ·

• ggml : port TurboVec to upstream turbovec 1.1.2 (1/6)

  ggmltesting
  #404·dev-nid opened Oct 20, 2026·

  ·
  2

• xdna : load the backend as a GGML_BACKEND_DL module

  ggml
  #403·gianni-cor opened Oct 20, 2026·

  ·
  2

• ggml, ggml-cuda : keep reserved allocation plans, launch peer-waiting CUDA splits as one graph, long-row TOP_K by radix

  CUDAggmltesting
  #391·pratiknarola-t opened Oct 15, 2026·2/3·

  ·

• opencl : speed up MTP verify and drafting on Adreno (phase 2)

  documentationggmlmodelOpenCLservertesting
  #390·pratiknarola-t opened Oct 15, 2026·

  ·

• common : restore thread-local log levels with backend reload coverage

  devopsggmltesting
  #389·opaninakuffo opened Oct 14, 2026·1 task·

  ·
  11

• llama : support stream gaps in DSA cache

  modeltesting
  #383·gianni-cor opened Oct 13, 2026·2/6·

  ·

• llama : support stream gaps in DSV4 cache

  testing
  #381·gianni-cor opened Oct 13, 2026·6/6·

  ·

• llama : support stream gaps in hybrid index cache

  modeltesting
  #380·gianni-cor opened Oct 13, 2026·5/6·

  ·

• llama : support stream gaps in MSA cache

  modeltesting
  #379·gianni-cor opened Oct 13, 2026·4/6·

  ·

• llama : support stream gaps in DSA-SWA cache

  modeltesting
  #378·gianni-cor opened Oct 13, 2026·3/6·

  ·

• opencl : MTP speculative-decoding measurements of #370 on a Galaxy S25 (Adreno 830) and the adb bench script

  documentationggmltesting
  #377·Zbig9000 opened Oct 13, 2026·

  ·

• Metal: add Core ML ANE matrix multiplication path

  Apple Metalggmltesting
  #373·anvarmohi opened Oct 12, 2026·

  ·
  1