llama.cpp 发布: b11140

来源摘要

CUDA: enable sparse-fa for dsv4 prefill (again) (#29298) * CUDA: enable sparse-fa for dsv4 prefill (again) * CUDA: unroll the query loop of the sparse mask scan The query loop of flash_attn_mask_to_sparse_indices has a runtime trip count, which keeps the unrolled scan over the values of a lane from issuing its loads together. Template the kernel on ncols1 so the loop is bounded at compile time: batch one decodes compile to straight line code and the scan drops from 46 to 17 us at 49k columns on sparse decode shapes. * CUDA: pick the out of bounds check of the sparse mask scan in host code The query loop of the ncols1 == 8 scan keeps a runtime bound and an early exit, so it does not unroll past its first iteration. Template the kernel on whether the last group of queries is partial, decided on the host from n_queries, and hoist the column bound out of the loop: the loop becomes straight line code and the batched sparse op at 49k context drops from 586 to 244 us. --------- Co-authored-by: Pascal **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11140/llama-b11140-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](https://github.com/ggml-org/llama.cpp/pull/23780) - [macOS Intel (x64)](https://github.com/ggml-org/llama.cpp/releases/download/b11140/llama-b11140-bin-macos-x64.tar.gz) - [iOS XCFramework](https://github.com/ggml-org/llama.cpp/releases/download/b11140/llama-b11140-xcframework.zip) **Linux:** - [Ubuntu x64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11140/llama-b11140-bin-ubuntu-x64.tar.gz) - [Ubuntu arm64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11140/llama-b11140-bin-ubuntu-arm64.tar.gz) - [Ubuntu s3