AISpot

llama.cpp b12077 支援 GLM-5.3,修復多裝置 MLA 拆頭

llama.cpp releases版本更新地端推論開源

llama.cpp 建置 b12077 加入 GLM-5.3(glm5-next)支援,可在 meta 後端跑 MLA 推理。由於 MLA 快取只有一個 KV head,多裝置時每張卡分到的頭數取決於 GQA 比例;HIP 上 D=512 的 tile FA kernel 要求該比例為偶數,三卡下 16 個頭被拆成 5/5/6 會直接中止,現已改為按對拆頭。

原標題:b12077
閱讀原文

評分45 / 45(平均 45,門檻 60)
狀態未入選

原文

meta: support GLM-5.3 next sm tensor ( #30499 ) meta: support GLM-5.3 next sm tensor glm5-next : cont q_absorbed for -sm tensor with non-unified KV meta : keep GGML_OP_ARANGE in ggml.h order glm5-next : split MLA heads in pairs for -sm tensor The MLA cache has a single KV head, so the heads on a device are its GQA ratio. The HIP tile FA kernel for D=512 needs an even GQA ratio; with 3 devices the 16 heads of the test model were split 5/5/6 and it aborted. meta : keep the split for views inside one slab of the split axis With a non-unified KV cache, build_attn_mha splits q into streams with a view. For a permuted q, e.g. the absorbed query of glm5-next, that view aborted with "view of permuted tensor not implemented". The split is now kept if every other dim of the source and the view stays inside one slab of the split axis. The per-device strides are only rescaled above it. From tetherto#353 . Co-authored-by: Pratik Narola [email protected] glm5-next : drop the q_absorbed cont The meta backend can now split the per-stream view of the permuted q, so the cont is not needed for -sm tensor with a non-unified KV cache. With the cont, the q of flash attention was not contiguous, and when FA ran on the CPU (WebGPU) the meta get_tensor aborted. Co-authored-by: Pratik Narola [email protected] Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/63228377 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu 26.04 s390x (CPU) Ubuntu 26.04 s390x (zDNN) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (CUDA 12) - CUDA 12.8 libraries Ubuntu x64 (CUDA 13) - CUDA 13.4 libraries Ubuntu arm64 (CUDA 13) - CUDA 13.4 libraries Ubuntu x64 (ROCm 10.0) Ubuntu x64 (OpenVINO) Ubuntu x64 (SYCL FP32) Ubuntu x64 (SYCL FP16) Linux arm64 (Snapdragon: CPU, Adreno GPU, Hexagon NPU) - setup guide Android: Android arm64 (CPU) Android arm64 (Snapdragon: CPU, Adreno GPU, Hexagon NPU) - setup guide Windows: Windows x64 (CPU) Windows arm64 (CPU) Windows arm64 (OpenCL Adreno) Windows x64 (CUDA 12) - CUDA 12.4 DLLs Windows x64 (CUDA 13) - CUDA 13.4 DLLs Windows arm64 (CUDA 13) - CUDA 13.4 DLLs Windows x64 (Vulkan) Windows arm64 (Vulkan) Windows x64 (OpenVINO) Windows x64 (SYCL) Windows x64 (ROCm 10.0) openEuler: DISABLED openEuler x86 (310p) openEuler x86 (910b, ACL Graph) openEuler aarch64 (310p) openEuler aarch64 (910b, ACL Graph) UI: UI

相關報導

llama.cpp releasesAI 評分43

llama.cpp b11515:SYCL 加入 Q5_K 重排 MMVQ 與融合 GLU

llama.cpp 發布 b11515,主要變更是 SYCL 後端為 Q5_K 量化補上重排佈局的 MMVQ 核心,並加入融合 GLU(#29375),面向 Intel GPU 的量化推理。建置矩陣照舊覆蓋 macOS/iOS、Linux、Windows、Android 與 openEuler,包含 CUDA 12/13、ROCm 10.0、Vulkan、OpenVINO、SYCL FP32/FP16 等後端;

llama.cpp releases● 精選AI 評分68

llama.cpp SYCL 為 GLM MLA 預填充接入 MKL,pp8192 提速 2.99 倍

llama.cpp 的 SYCL 後端放行了 GLM-4.7 Flash 的 MLA 形狀走 MKL flash attention,此前該形狀因 K/V 寬度不一致、head dim 上限 512 被排程器拒絕,預填充只能回退到更慢的 TILE kernel。在 Intel Arc Pro B70 上,pp8192 從 432.80 提升到 1292.29 tok/s(2.99 倍,+198.6%),tg256 在噪聲範圍內基本不變(45.67 對 45.65)。

llama.cpp releasesAI 評分45

llama.cpp Vulkan 後端支援 ggml src1 精度並調整累加器精度選擇

llama.cpp 建置 b11854 為 Vulkan 後端加入 ggml src1 精度支援,並更新累加器精度選擇邏輯。改動通過縮放處理 fp32 溢位,並修復 copy-to-fp16 啟用中的溢位問題;同時補充 ggml 精度查詢輔助函式、在 mul_mat 中增加檢查與測試,並統一命名為 acc_prec。此次建置仍覆蓋 macOS、Linux、Windows、Android 的 Vulkan、CUDA、ROCm、SYCL、OpenVINO 等後端二進位制。

llama.cpp releases● 精選AI 評分78

llama.cpp 合併 GLM5-Next MTP 支援,可作 draft 模型

llama.cpp 合併 PR #29928,為 GLM5-Next 加入多 token 預測(MTP)的 NextN 計算圖,使其能作為投機解碼的 draft 模型使用。改動包括支援只含 NextN 層或只含主幹張量的拆分 GGUF 檔案載入,並把 MTP 圖按輸出行裁剪,無輸出行的 4-token catch-up 核心耗時從 6.9 ms 降到 2.9 ms,貪心輸出雜湊與 draft 接受率不變。