Code available at: https://github.com/facebookresearch/ads_model_kernel_library In this post, we present the design of TLX Block Attention — a Triton kernel targeting NVIDIA Blackwell GPUs that exploits compile-time knowledge of a block-diagonal…
TLX Block Attention: A Warp-Specialized Blackwell Kernel for Fixed-Block Sparse Self-Attention
calendar_today
May 26, 2026
person
Jake Siso, Dev (Devashish) Shankar, Jackie (Jiaqi) Xu, Jacky Zhou, Darren Liu, Han Xu, Yasmine Badr, Dan Chanpuriya, Hongtao Yu, Max Leung
domain
pytorch