Skip to content

LLM inference in C/C++

C++ ggml

vulkan: prefer host-visible memory buffers on UMA devices (#22930)

Winston Ma32120c10e33b
diff --git a/ggml/src/ggml-vulkan/ggml-vulkan.cpp b/ggml/src/ggml-vulkan/ggml-vulkan.cpp
index 59582e4f0f..243ab76ee7 100644
--- a/ggml/src/ggml-vulkan/ggml-vulkan.cpp
+++ b/ggml/src/ggml-vulkan/ggml-vulkan.cpp
@@ -3080,8 +3080,10 @@ static vk_buffer ggml_vk_create_buffer_device(vk_device& device, size_t size) {
buf = ggml_vk_create_buffer(device, size, {vk::MemoryPropertyFlagBits::eHostVisible | vk::MemoryPropertyFlagBits::eHostCoherent,
vk::MemoryPropertyFlagBits::eDeviceLocal});
} else if (device->uma) {
- // Fall back to host memory type
- buf = ggml_vk_create_buffer(device, size, {vk::MemoryPropertyFlagBits::eDeviceLocal,
+ // On UMA, prefer host-visible memory so direct tensor borrowing works.
+ // If unavailable, fall back to device-local memory.
+ buf = ggml_vk_create_buffer(device, size, {vk::MemoryPropertyFlagBits::eDeviceLocal | vk::MemoryPropertyFlagBits::eHostVisible | vk::MemoryPropertyFlagBits::eHostCoherent,
+ vk::MemoryPropertyFlagBits::eDeviceLocal,
vk::MemoryPropertyFlagBits::eHostVisible | vk::MemoryPropertyFlagBits::eHostCoherent});
} else if (device->disable_host_visible_vidmem) {
if (device->allow_sysmem_fallback) {