matt/ollama37
1
0
Fork 0
mirror of https://github.com/dogkeeper886/ollama37.git synced 2025-12-11 08:17:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

nvidia libs have inconsistent ordering (#7473)

Browse Source 
The runtime and management libraries may not always have
identical ordering, so use the device UUID to correlate instead of ID.
This commit is contained in:
Daniel Hiltgen
2024-11-02 16:35:41 -07:00
committed by GitHub
parent b8d5036e33
commit 29ab9fa7d7
3 changed files with 13 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
discover/gpu.go
View File

@@ -316,7 +316,9 @@ func GetGPUInfo() GpuInfoList {
// query the management library as well so we can record any skew between the two
// which represents overhead on the GPU we must set aside on subsequent updates
if cHandles.nvml != nil {
C.nvml_get_free(*cHandles.nvml, C.int(gpuInfo.index), &memInfo.free, &memInfo.total, &memInfo.used)
uuid := C.CString(gpuInfo.ID)
defer C.free(unsafe.Pointer(uuid))
C.nvml_get_free(*cHandles.nvml, uuid, &memInfo.free, &memInfo.total, &memInfo.used)
if memInfo.err != nil {
slog.Warn("error looking up nvidia GPU memory", "error", C.GoString(memInfo.err))
C.free(unsafe.Pointer(memInfo.err))
@@ -417,7 +419,9 @@ func GetGPUInfo() GpuInfoList {
}
for i, gpu := range cudaGPUs {
if cHandles.nvml != nil {
C.nvml_get_free(*cHandles.nvml, C.int(gpu.index), &memInfo.free, &memInfo.total, &memInfo.used)
uuid := C.CString(gpu.ID)
defer C.free(unsafe.Pointer(uuid))
C.nvml_get_free(*cHandles.nvml, uuid, &memInfo.free, &memInfo.total, &memInfo.used)
} else if cHandles.cudart != nil {
C.cudart_bootstrap(*cHandles.cudart, C.int(gpu.index), &memInfo)
} else if cHandles.nvcuda != nil {