Issues

• [Feature Request] Support Ascend NPU for sparse attention implementation

  mit-han-lab/Block-Sparse-Attention#18 · Filter by author kertenn opened on Feb 13, 2026

• compute_120

  mit-han-lab/Block-Sparse-Attention#15 · Filter by author gunerhui opened on Dec 31, 2025
  2

• Any Plan to Support FlashAttention3？

  mit-han-lab/Block-Sparse-Attention#13 · Filter by author xiaolong-li1 opened on Nov 27, 2025
  1

• RTX Blackwell (5090) Compatibility?

  mit-han-lab/Block-Sparse-Attention#9 · Filter by author zelenooki opened on Oct 18, 2025
  4

• Why x-attention only support batch size = 1 when it used block-sparse-attention repo?

  mit-han-lab/Block-Sparse-Attention#7 · Filter by author ZetangForward opened on May 6, 2025
  2

• Why block size have to be 128?

  mit-han-lab/Block-Sparse-Attention#6 · Filter by author suhmily10 opened on Mar 12, 2025
  7

• Compile Issue

  mit-han-lab/Block-Sparse-Attention#5 · Filter by author huangyuxiang03 opened on Nov 29, 2024
  11

• Install problem

  mit-han-lab/Block-Sparse-Attention#4 · Filter by author phyllispeng123 opened on Nov 19, 2024
  4

• Errors during install

  mit-han-lab/Block-Sparse-Attention#3 · Filter by author twistedcubic opened on Nov 4, 2024
  3

• h100 support

  mit-han-lab/Block-Sparse-Attention#2 · Filter by author Lucas-TY opened on Oct 22, 2024
  4

• Comparison against FlexAttention?

  mit-han-lab/Block-Sparse-Attention#1 · Filter by author jzhang38 opened on Oct 16, 2024
  1