baalajimaestro/ollama
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

update cuda overhead to 20% to fix crashes when switching between models and large context sizes

Browse source
This commit is contained in:
Jeffrey Morgan 2024-01-09 00:05:23 -05:00
parent f6cb0a553c
commit 37708931fb
1 changed files with 2 additions and 8 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
gpu/gpu.go
View file

@ -131,14 +131,8 @@ func getCPUMem() (memInfo, error) {
func CheckVRAM() (int64, error) {
gpuInfo := GetGPUInfo()
if gpuInfo.FreeMemory > 0 && (gpuInfo.Library == "cuda" || gpuInfo.Library == "rocm") {
// leave 15% or 400MiB of VRAM free for overhead
overhead := gpuInfo.FreeMemory * 3 / 20
minOverhead := uint64(400 * 1024 * 1024)
if overhead < minOverhead {
overhead = minOverhead
}
return int64(gpuInfo.FreeMemory - overhead), nil
// leave 20% of VRAM free for overhead
return int64(gpuInfo.FreeMemory * 4 / 5), nil
}
return 0, fmt.Errorf("no GPU detected") // TODO - better handling of CPU based memory determiniation