All pull requests

Pull requests

• Back out "Frequency-adaptive learning rate for rowwise_rmsprop_ar: codegen"

  cla signedmeta-exported
  #6524·spcyppt opened Oct 20, 2026Contributor·

  ·
  1

• Add opt-in INT8 storage to the owner-local direct enrichment cache (#6522)

  cla signedmeta-exported
  #6522·EddyLXJ opened Oct 20, 2026Contributor·

  ·
  1

• Serialize composite DRAM and SSD restore (#6521)

  cla signedmeta-exported
  #6521·q10 opened Oct 20, 2026Contributor·

  ·
  1

• Resolve SSD cache associativity per device (#6519)

  cla signedmeta-exported
  #6519·q10 opened Oct 20, 2026Contributor·

  ·
  1

• Preserve global IDs in whole-row SSD storage (#6518)

  cla signedmeta-exported
  #6518·q10 opened Oct 20, 2026Contributor·

  ·
  1

• Preserve DRAM whole-row checkpoint IDs (#6517)

  cla signedmeta-exported
  #6517·q10 opened Oct 20, 2026Contributor·

  ·
  1

• Fix relocation overflow breaking ms_depth/metastereo_bundle builds

  cla signedmeta-exported
  #6516·q10 opened Oct 19, 2026Contributor·

  ·
  1

• Complete wave32 TBE codegen

  cla signedmeta-exported
  #6515·q10 opened Oct 19, 2026Contributor·

  ·
  1

• Add INT8 owner-local direct cache ops (#6514)

  cla signedmeta-exported
  #6514·EddyLXJ opened Oct 19, 2026Contributor·

  ·
  1

• Run merge_pooled_embeddings cross-device copies with the source device current

  cla signedmeta-exported
  #6513·kmonian opened Oct 19, 2026·

  ·
  1

• Require strongly typed inputs in the public CPU FP8 embedding API

  cla signed
  #6512·cyyever opened Oct 18, 2026Contributor·

  ·
  1

• Remove unused FP8 rowwise scratch allocation

  cla signed
  #6511·cyyever opened Oct 18, 2026Contributor·

  ·
  1

• Remove unused FP8 rowwise host template parameter

  cla signed
  #6510·cyyever opened Oct 18, 2026Contributor·

  ·
  1

• Specialize GPU N-bit quantization kernels for common bit widths

  cla signed
  #6509·cyyever opened Oct 18, 2026Contributor·

  ·
  1

• Vectorize N-bit dequantization with front scale and bias on AVX2

  cla signed
  #6507·cyyever opened Oct 18, 2026Contributor·

  ·
  1

• Enable constexpr convolution and scalar quantization helpers (#6485)

  cla signedmeta-exported
  #6504·q10 opened Oct 17, 2026Contributor·

  ·
  1

• Add laser_key_suffix to EnrichmentPolicy for versioned Laser tiers

  cla signedmeta-exported
  #6499·wilsonyqm-debug opened Oct 17, 2026·

  ·

• Add NEON kernels for fp32 <-> fp16 conversion

  cla signed
  #6498·sremes opened Oct 17, 2026·

  ·
  1

• Validate CK FP8 inputs using the native GPU dtype (#6497)

  cla signedmeta-exported
  #6497·q10 opened Oct 17, 2026Contributor·

  ·
  1

• Add MI350/MI450 TBE and GIS benchmark harness

  cla signedmeta-exported
  #6496·q10 opened Oct 17, 2026Contributor·

  ·
  1

• Frequency-adaptive learning rate for rowwise_rmsprop_ar: TBE frontend (#6492)

  cla signedmeta-exported
  #6492·hdmeta opened Oct 16, 2026Contributor·

  ·
  1

• Replace legacy null CUDA stream constants

  cla signedmeta-exported
  #6479·q10 opened Oct 16, 2026Contributor·

  ·
  1

• Avoid eager empty tensors in INT-N-bit inference dispatch (#6473)

  cla signedmeta-exported
  #6473·mrmiywj opened Oct 15, 2026Contributor·

  ·
  1

• Re-enable TBE backward none and SGD writeback tests on ROCm (#6472)

  ciflow/rocmcla signedmeta-exportedmodule: rocm
  #6472·q10 opened Oct 15, 2026Contributor·

  ·
  2

• Suppress type errors for Pyre upgrade (landing group 36)

  cla signedmeta-exported
  #6471·maggiemoss opened Oct 15, 2026Contributor·

  ·
  1