All pull requests

Pull requests

• RDNA4 phase 4: FlashAttention on gfx12 WMMA (bf16, FP8 QK^T)

  #397·morousg opened Oct 18, 2026Member·

  ·
  7

• RDNA4 phase 3: scalar FlashAttention, FP8/int8 KV, decode and softmax on HIP

  #396·morousg opened Oct 18, 2026Member·

  ·
  5

• RDNA4 phase 2: CUDA tests for FP8 KV, flash decode and the mma kernel (CUDA-side, unverified locally)

  #395·morousg opened Oct 18, 2026Member·

  ·
  5

• RDNA4 phase 1: make HIP_ARCH work, attention wave-size helpers

  #394·morousg opened Oct 18, 2026Member·

  ·
  2

• testing ubuntu 26.04 wsl cuda

  #389·albertandaluz opened Oct 13, 2026Member·

  ·
  1

  1

• Clarify FKL skills for pipeline composition and Operation/DPP authoring

  #366·albertandaluz with CopilotAI opened Sep 8, 2026·

  ·
  1

  11

  

• Decouple thread fusion from the Read/Write Operation authoring contract

  #324·johnnynunez opened Jul 18, 2026Contributor·

  ·
  

• First-class fp16/bf16/fp8/fp4 element types for both backends

  #321·johnnynunez opened Jul 17, 2026Contributor·

  ·
  4