CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-04 09:32:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

vulkan: add environment variable GGML_VK_PREFER_HOST_MEMORY to avoid VRAM allocation (#11592)

Browse Source
This commit is contained in:
Wagner Bruna
2025-02-10 03:08:22 -03:00
committed by GitHub
parent 19d3c8293b
commit b044a0fe3c
1 changed files with 7 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
ggml/src/ggml-vulkan/ggml-vulkan.cpp
View File

@@ -167,6 +167,7 @@ struct vk_device_struct {
uint32_t subgroup_size; uint32_t subgroup_size;
uint32_t shader_core_count; uint32_t shader_core_count;
bool uma; bool uma;
bool prefer_host_memory;
bool float_controls_rte_fp16; bool float_controls_rte_fp16;
bool subgroup_size_control; bool subgroup_size_control;
@@ -1294,7 +1295,9 @@ static vk_buffer ggml_vk_create_buffer_check(vk_device& device, size_t size, vk:
static vk_buffer ggml_vk_create_buffer_device(vk_device& device, size_t size) { static vk_buffer ggml_vk_create_buffer_device(vk_device& device, size_t size) {
vk_buffer buf; vk_buffer buf;
try { try {
if (device->uma) { if (device->prefer_host_memory) {
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 // Fall back to host memory type
buf = ggml_vk_create_buffer(device, size, vk::MemoryPropertyFlagBits::eDeviceLocal, vk::MemoryPropertyFlagBits::eHostVisible | vk::MemoryPropertyFlagBits::eHostCoherent); buf = ggml_vk_create_buffer(device, size, vk::MemoryPropertyFlagBits::eDeviceLocal, vk::MemoryPropertyFlagBits::eHostVisible | vk::MemoryPropertyFlagBits::eHostCoherent);
} else { } else {
@@ -2199,6 +2202,9 @@ static vk_device ggml_vk_get_device(size_t idx) {
device->physical_device = physical_devices[dev_num]; device->physical_device = physical_devices[dev_num];
const std::vector<vk::ExtensionProperties> ext_props = device->physical_device.enumerateDeviceExtensionProperties(); const std::vector<vk::ExtensionProperties> ext_props = device->physical_device.enumerateDeviceExtensionProperties();
const char* GGML_VK_PREFER_HOST_MEMORY = getenv("GGML_VK_PREFER_HOST_MEMORY");
device->prefer_host_memory = GGML_VK_PREFER_HOST_MEMORY != nullptr;
bool fp16_storage = false; bool fp16_storage = false;
bool fp16_compute = false; bool fp16_compute = false;
bool maintenance4_support = false; bool maintenance4_support = false;