Posts by Kunlun Li

Support Blackwell with FFA_FA4 Backend

• 07 February 2026
Yunpeng Huang , Yufeng Yang , Jerry Chen , Yujia Liu , Zewei Tao , Kunlun Li
China
English
MagiAttention
Blackwell Flex-Flash-Attention Flash-Attention HSTU Function Representation

Before the release of MagiAttention-v1.1.0, MagiAttention had supported only the Hopper GPUs, since the attention kernel backend Flex-Flash-Attention (FFA) is built upon open-sourced Flash-Attention 3 (FA3) [Shah et al., 2024], tailored for SM90 compute capability.

Read more ...

---