llama.cpp b10857 发布：修复 32 位 Vulkan 句柄问题

AI 摘要

llama.cpp 发布 b10857 版本，修复了 Vulkan-Hpp 在 32 位平台上的句柄使用问题，包括为 vk::Buffer 添加 operator<< 以及直接传递 vk::Buffer 给 copyBuffer API。该版本提供了适用于 macOS、Linux、Windows、Android 等多个平台的预编译二进制文件。

正文节选

<details open> Fix Vulkan-Hpp handle usage on 32-bit targets. (#22892) On 32-bit platforms, Vulkan non-dispatchable handles such as VkBuffer are represented as uint64_t, and Vulkan-Hpp disables implicit conversions for type safety. This exposes two issues in ggml-vulkan: 1. vk::Buffer is streamed directly into std::ostream in debug/memory logs. 2. vk::Buffer is cast to VkBuffer before being passed to Vulkan-Hpp CommandBuffer::copyBuffer APIs. Fix these by add the operator<< for vk::Buffer

发布时间：2026-09-08 19:36

抓取时间：2026-09-08 19:56

来源机构：ggml-org