Issues

• [Feature]: Integrating Ultra low latency TopK for into flash-infer-rocm for DeepSeek V3.2

  AMD-Ecosystem/flashinfer#217 · Filter by author yiakwy-xpu-ml-framework-team opened on Apr 27, 2026
  1

• [Feature]: Add gfx1151 support

  AMD-Ecosystem/flashinfer#134 · Filter by author neuhaus opened on Jan 29, 2026
  2

• [Feature]: Implement a Batch Attention Kernel

  AMD-Ecosystem/flashinfer#101 · Filter by author diptorupd opened on Dec 12, 2025

• Make it possible to support different HEAD_DIM in prefill kernels outside of those specified in dispatch.cuh::DISPATCH_HEAD_DIM

  enhancement
  AMD-Ecosystem/flashinfer#85 · Filter by author diptorupd opened on Dec 3, 2025

• Batch decode FP8 tests fail on mi300a

  bug
  Bug

  AMD-Ecosystem/flashinfer#84 · Filter by author diptorupd opened on Dec 1, 2025

• Enable FP8 WFMA support for CDNA3

  enhancement
  Feature

  AMD-Ecosystem/flashinfer#82 · Filter by author diptorupd opened on Dec 1, 2025

• Support manylinux wheel builds for amd-flashinfer

  infrastructure
  Feature

  AMD-Ecosystem/flashinfer#81 · Filter by author diptorupd opened on Dec 1, 2025

• Redesign of attention/generic

  questionrefactoring
  Task

  AMD-Ecosystem/flashinfer#41 · Filter by author diptorupd opened on Nov 11, 2025
  

• Does rocm/flashinfer have something similar to flashinfer.green_ctx?

  enhancementquestion
  Feature

  AMD-Ecosystem/flashinfer#29 · Filter by author monopodium opened on Oct 22, 2025
  1