[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$f370lup3wsetm":3},{"slug":4,"category":5,"publishedAt":6,"titleZh":7,"titleEn":8,"summaryZh":9,"summaryEn":10,"models":11,"hardwares":12,"bodyZh":13,"bodyEn":14,"sourceName":15,"sourceUrl":16,"updatedAt":6},"llamacpp-b12548-metal-private-buffer-range-2026-11-20","RUNTIME","2026-11-19T23:07:55.736Z","llama.cpp修复Metal私有缓冲区范围：Intel独显Mac与特殊环境变量路径需留意","llama.cpp fixes Metal private-buffer ranges on discrete-GPU Macs and opt-in paths","11月19日合并的#32602进入b12548，修正NSMakeRange第二参数的语义。影响私有缓冲区路径，不代表所有Apple Silicon部署都会提速；上游明确这一分支未纳入CI。","Merged November 19 and shipped in b12548, #32602 fixes the length argument to NSMakeRange on the private-buffer path. It is not a general Apple Silicon speedup; upstream says this path is outside CI.",[],[],"llama.cpp于11月19日合并Metal修复#32602，窗口内发布的b12548已列入这项变更。问题不是模型参数量或量化档位，而是底层缓冲区范围的表达：NSMakeRange的第二个参数表示长度，不是结束位置。把两者混用，即使代码可以编译，也不能说明所描述的内存区间正确。\n\n1. 先确认是否走到受影响路径。贡献者指出，这是private buffers分支，涉及Intel Mac上的独立显卡，或设置了GGML_METAL_SHARED_BUFFERS_DISABLE=1的环境。不能将“Metal有修复”直接翻译为所有M系列Mac都存在同一种问题，更不能按芯片名称判断某次运行一定命中了它。排错要同时记录设备、构建和环境变量。\n\n2. 合并和发版是本窗口内的新事实，不是新的全机型测试。PR给出了API语义和影响范围，并明确这些路径不跑CI。本站未独立复现，也没有可用于发布速度结论的同协议前后对照。因此本文不报提速百分比，不推断输出质量变化，也不把补丁合并当成所有模型与系统组合都已验收。\n\n3. 升级应验证目标任务，不只看进程能启动。建议保留原二进制和实际启动配置，在同一权重、量化、上下文及后端设置下对照加载日志和任务输出；使用特殊环境变量时，把该变量一并带入复现材料。若默认共享缓冲区路径正常，不应为测试修复而把新的绕行配置直接搬进生产。\n\n这次修复值得关注的原因是影响范围清楚，而不是数字夸张：它解决一个具体的范围参数问题，没有给出通用性能承诺。引擎许可和模型许可仍需分别核对。后续若出现独立复现，应把系统版本、构建号、触发条件与失败原文一起留存，再判断影响是否扩大。\n\n来源：https:\u002F\u002Fgithub.com\u002Fggml-org\u002Fllama.cpp\u002Fpull\u002F32602\n发行：https:\u002F\u002Fgithub.com\u002Fggml-org\u002Fllama.cpp\u002Freleases\u002Ftag\u002Fb12548","llama.cpp merged Metal fix #32602 on November 19 and included it in b12548 during this reporting window. The issue is not parameter count or quantization: the second argument of NSMakeRange is a length, not an end position. Code compiling successfully does not prove that it describes the intended memory range.\n\n1. Establish whether the affected path is used. The contributor identifies private buffers, affecting discrete GPUs on Intel Macs or environments with GGML_METAL_SHARED_BUFFERS_DISABLE=1. A Metal fix does not imply that every Apple Silicon deployment has the same defect. Record the device, build and environment variables together.\n\n2. Merge and release are new events, not a new all-device validation campaign. The PR states that these paths are not covered by CI. AraoAI has not independently reproduced the issue, and no matched performance comparison is available here. We report no speedup percentage or quality improvement and do not equate a merged patch with validation of every model and system combination.\n\n3. Validate the target workload after upgrading. Keep the old binary and launch configuration, then compare load logs and task outputs using the same weights, quantization, context and backend settings. Include non-default environment variables in the reproduction. Do not move an experimental private-buffer configuration into production merely to exercise this fix.\n\nThis is a narrowly scoped range-argument correction, not a general performance promise. Engine and model licenses remain separate. Future independent reports should retain the OS version, build, trigger and original failure before broadening the impact assessment.\n\nSource: https:\u002F\u002Fgithub.com\u002Fggml-org\u002Fllama.cpp\u002Fpull\u002F32602\nRelease: https:\u002F\u002Fgithub.com\u002Fggml-org\u002Fllama.cpp\u002Freleases\u002Ftag\u002Fb12548","llama.cpp GitHub PR \u002F release","https:\u002F\u002Fgithub.com\u002Fggml-org\u002Fllama.cpp\u002Fpull\u002F32602"]