Superminaren/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-09-04 02:37:27 +02:00
Code Issues Packages Projects Releases Wiki Activity

metal : fix memory query under low-memory conditions (#27701)

Browse Source 
* metal: Fix memory query under low-memory conditions

* Simply variable name

Co-authored-by: Georgi Gerganov <[email protected]>

* Write it even shorter

Co-authored-by: Niklas Wenzel <[email protected]>

---------

Co-authored-by: Georgi Gerganov <[email protected]>
Co-authored-by: Niklas Wenzel <[email protected]>
This commit is contained in:
Mads Marquart
2026-09-02 20:09:46 +02:00
committed by GitHub
co-authored by Georgi Gerganov Niklas Wenzel
parent 03dbcc53e1
commit 8c0b9cd04a
1 changed files with 4 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-metal/ggml-metal-device.m
+4 -2
View File
@@ -1471,8 +1471,10 @@ void ggml_metal_device_event_synchronize(ggml_metal_device_t dev, ggml_metal_eve
void ggml_metal_device_get_memory(ggml_metal_device_t dev, size_t * free, size_t * total) {
if (@available(macOS 10.12, iOS 16.0, *)) {
*total = dev->mtl_device.recommendedMaxWorkingSetSize;
*free = *total - dev->mtl_device.currentAllocatedSize;
*total = dev->mtl_device.recommendedMaxWorkingSetSize;
size_t cur = dev->mtl_device.currentAllocatedSize;
// it's possible to allocate more than `recommendedMaxWorkingSetSize`
*free = *total > cur ? *total - cur : 0;
} else {
*free = 0;
*total = 0;