Skip to content

vulkan: reduces lock contention #23376

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
0cc4m merged 2 commits into from
Jun 1, 2026
+5 −3
Merged

vulkan: reduces lock contention #23376

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
81d827d
vulkan: reduces lock contention
winstonma May 20, 2026
af2cc47
replace unique_lock with lock_guard
winstonma May 28, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 5 additions & 3 deletions ggml/src/ggml-vulkan/ggml-vulkan.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -49,6 +49,7 @@ DispatchLoaderDynamic & ggml_vk_default_dispatcher();
#include <map>
#include <set>
#include <unordered_map>
#include <shared_mutex>
#include <mutex>
#include <future>
#include <thread>
Expand Down Expand Up @@ -595,6 +596,7 @@ static constexpr std::initializer_list<std::array<int, 3>> rms_norm_mul_rope_vie

struct vk_device_struct {
std::recursive_mutex mutex;
mutable std::shared_mutex pinned_memory_mutex;

vk::PhysicalDevice physical_device;
vk::PhysicalDeviceProperties properties;
Expand Down Expand Up @@ -6633,7 +6635,7 @@ static void * ggml_vk_host_malloc(vk_device& device, size_t size) {
return nullptr;
}

std::lock_guard<std::recursive_mutex> guard(device->mutex);
std::unique_lock<std::shared_mutex> guard(device->pinned_memory_mutex);
Comment thread
0cc4m marked this conversation as resolved.
Outdated
device->pinned_memory.push_back(std::make_tuple(buf->ptr, size, buf));

return buf->ptr;
Expand All @@ -6644,7 +6646,7 @@ static void ggml_vk_host_free(vk_device& device, void* ptr) {
return;
}
VK_LOG_MEMORY("ggml_vk_host_free(" << ptr << ")");
std::lock_guard<std::recursive_mutex> guard(device->mutex);
std::unique_lock<std::shared_mutex> guard(device->pinned_memory_mutex);

vk_buffer buf;
size_t index;
Expand All @@ -6668,7 +6670,7 @@ static void ggml_vk_host_free(vk_device& device, void* ptr) {
}

static void ggml_vk_host_get(const vk_device& device, const void * ptr, vk_buffer& buf, size_t& buf_offset) {
std::lock_guard<std::recursive_mutex> guard(device->mutex);
std::shared_lock<std::shared_mutex> guard(device->pinned_memory_mutex);
buf = nullptr;
buf_offset = 0;
for (size_t i = 0; i < device->pinned_memory.size(); i++) {
Expand Down
Loading