5 Pull requests merged by 1 person

• refactor: simplify mha block tiling logic

  #488 merged Jul 18, 2025

• feat: added KV multi-stages support for attn sm120

  #489 merged Jul 17, 2025

• refactor: move kernel code into different folders

  #487 merged Jul 17, 2025

• feat: [2/n] added warp specialization kernel for sm120 fmha

  #486 merged Jul 17, 2025

• feat: [1/n] added sm120 fmha using collective async copy

  #483 merged Jul 17, 2025