Skip to content
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

Jump to bottom

Support forced spreading for multi GPU #4266

Closed
wants to merge 1 commit into from
+18 −4
Closed

Support forced spreading for multi GPU #4266

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
12 changes: 12 additions & 0 deletions server/envconfig/config.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,6 +29,8 @@ var (
NumParallel int
// Set via OLLAMA_RUNNERS_DIR in the environment
RunnersDir string
// Set via OLLAMA_SCHED_SPREAD in the environment
SchedSpread bool
// Set via OLLAMA_TMPDIR in the environment
TmpDir string
)
Expand All @@ -44,6 +46,7 @@ func AsMap() map[string]string {
"OLLAMA_NOPRUNE": fmt.Sprintf("%v", NoPrune),
"OLLAMA_NUM_PARALLEL": fmt.Sprintf("%v", NumParallel),
"OLLAMA_RUNNERS_DIR": fmt.Sprintf("%v", RunnersDir),
"OLLAMA_SCHED_SPREAD": fmt.Sprintf("%v", SchedSpread),
"OLLAMA_TMPDIR": fmt.Sprintf("%v", TmpDir),
}
}
Expand Down Expand Up @@ -137,6 +140,15 @@ func LoadConfig() {
}
}

if spread := clean("OLLAMA_SCHED_SPREAD"); spread != "" {
s, err := strconv.ParseBool(spread)
if err == nil {
SchedSpread = s
} else {
SchedSpread = true
}
}

if noprune := clean("OLLAMA_NOPRUNE"); noprune != "" {
NoPrune = true
}
Expand Down
10 changes: 6 additions & 4 deletions server/sched.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -493,10 +493,12 @@ func pickBestFitGPUs(req *LlmRequest, ggml *llm.GGML, gpus gpu.GpuInfoList) gpu.
sort.Sort(sort.Reverse(gpu.ByFreeMemory(sgl)))

// First attempt to fit the model into a single GPU
for _, g := range sgl {
if ok, estimatedVRAM = llm.PredictServerFit([]gpu.GpuInfo{g}, ggml, req.model.AdapterPaths, req.model.ProjectorPaths, req.opts); ok {
slog.Debug("new model will fit in available VRAM in single GPU, loading", "model", req.model.ModelPath, "gpu", g.ID, "available", g.FreeMemory, "required", format.HumanBytes2(estimatedVRAM))
return []gpu.GpuInfo{g}
if !envconfig.SchedSpread {
for _, g := range sgl {
if ok, estimatedVRAM = llm.PredictServerFit([]gpu.GpuInfo{g}, ggml, req.model.AdapterPaths, req.model.ProjectorPaths, req.opts); ok {
slog.Debug("new model will fit in available VRAM in single GPU, loading", "model", req.model.ModelPath, "gpu", g.ID, "available", g.FreeMemory, "required", format.HumanBytes2(estimatedVRAM))
return []gpu.GpuInfo{g}
}
}
}

Expand Down