Home
last modified time | relevance | path

Searched defs:tSrQ (Results 1 – 2 of 2) sorted by relevance

/external/pytorch/aten/src/ATen/native/transformers/cuda/flash_attn/
Dflash_fwd_kernel.h163 Tensor tSrQ = thr_mma.partition_fragment_A(sQ); // (MMA,MMA_M,MMA_K) in compute_attn_1rowblock() local
609 Tensor tSrQ = thr_mma.partition_fragment_A(sQ); // (MMA,MMA_M,MMA_K) in compute_attn_1rowblock_splitkv() local
Dflash_bwd_kernel.h216 Tensor tSrQ = thr_mma_sdp.partition_fragment_A(sQ); // (MMA,MMA_N,MMA_K) in compute_dq_dk_dv_1colblock() local