llama.cpp · b12059
llama : flash_attn_ext_rows - more efficient unified kv for multi-seq (#29510) * llama : use llm_attn_mask and kv_rows for unified attention Assisted-by: dsh:llama.cpp/ggml-org/DeepSeek-V4-Flash-Vision-Exp-GGUF (Mac Studio, M2 Ultra, 192GB) Co-authored-by: Georgi Gerganov * cont : optimize `get_seq_cells` * llama : build the kv_rows lists from per-sequence cell bitmaps Keep a bitmap over the cells for each sequence (the transpose of `seq`), updated in seq_pos_inc/dec, and build the cell lists by walking its set bits. The lists stay in cell order, cells shared by several sequences go to every list, and nothing has to be rebuilt when cells are removed. Also leave out of the lists the cells that the SWA hides from every token of the slice (standard and chunked SWA). * cont : make get_kv_rows_