mirrors/ollama
1
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2025-05-11 18:36:41 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Ignore AMD integrated GPUs

Browse source 
Detect and ignore integrated GPUs reported by rocm.
This commit is contained in:
Daniel Hiltgen 2024-01-25 15:57:32 -08:00
parent 197e420a97
commit 9d7b5d6c91
3 changed files with 35 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

26
gpu/gpu.go
View file

@ -16,6 +16,7 @@ import (
"os"
"path/filepath"
"runtime"
"strconv"
"strings"
"sync"
"unsafe"
@ -147,7 +148,28 @@ func GetGPUInfo() GpuInfo {
if memInfo.err != nil {
slog.Info(fmt.Sprintf("error looking up ROCm GPU memory: %s", C.GoString(memInfo.err)))
C.free(unsafe.Pointer(memInfo.err))
} else if memInfo.igpu_index >= 0 && memInfo.count == 1 {
// Only one GPU detected and it appears to be an integrated GPU - skip it
slog.Info("ROCm unsupported integrated GPU detected")
} else {
if memInfo.igpu_index >= 0 {
// We have multiple GPUs reported, and one of them is an integrated GPU
// so we have to set the env var to bypass it
// If the user has specified their own ROCR_VISIBLE_DEVICES, don't clobber it
val := os.Getenv("ROCR_VISIBLE_DEVICES")
if val == "" {
devices := []string{}
for i := 0; i < int(memInfo.count); i++ {
if i == int(memInfo.igpu_index) {
continue
}
devices = append(devices, strconv.Itoa(i))
}
val = strings.Join(devices, ",")
os.Setenv("ROCR_VISIBLE_DEVICES", val)
}
slog.Info(fmt.Sprintf("ROCm integrated GPU detected - ROCR_VISIBLE_DEVICES=%s", val))
}
resp.Library = "rocm"
var version C.rocm_version_resp_t
C.rocm_get_version(*gpuHandles.rocm, &version)
@ -199,7 +221,9 @@ func CheckVRAM() (int64, error) {
if overhead < gpus*1024*1024*1024 {
overhead = gpus * 1024 * 1024 * 1024
}
return int64(gpuInfo.FreeMemory - overhead), nil
avail := int64(gpuInfo.FreeMemory - overhead)
slog.Debug(fmt.Sprintf("%s detected %d devices with %dM available memory", gpuInfo.Library, gpuInfo.DeviceCount, avail/1024/1024))
return avail, nil
}
return 0, fmt.Errorf("no GPU detected") // TODO - better handling of CPU based memory determiniation