Skip to content

GitLab

Commit 17df6520 authored by Daniel Hiltgen's avatar Daniel Hiltgen
Browse files

Remove mmap related output calc logic

parent 6f351bf5
Show whitespace changes
Inline Side-by-side
Showing with 1 addition and 2 deletions
llm/memory.go
View file @ 17df6520
...@@ -197,8 +197,7 @@ func EstimateGPULayers(gpus []gpu.GpuInfo, ggml *GGML, projectors []string, opts ...@@ -197,8 +197,7 @@ func EstimateGPULayers(gpus []gpu.GpuInfo, ggml *GGML, projectors []string, opts
} }
// Determine if we need to consider output then find where it fits // Determine if we need to consider output then find where it fits
if ((gpus[0].Library == "metal" && opts.UseMMap) || (gpus[0].Library != "metal" || !opts.UseMMap)) && if memoryLayerOutput > 0 && (opts.NumGPU < 0 || layerCount < opts.NumGPU) {
memoryLayerOutput > 0 && (opts.NumGPU < 0 || layerCount < opts.NumGPU) {
for j := len(gpusWithSpace); j > 0; j-- { for j := len(gpusWithSpace); j > 0; j-- {
g := gpusWithSpace[layerCount%j] g := gpusWithSpace[layerCount%j]
used := gpuAllocations[g.i] + max(graphPartialOffload, graphFullOffload) used := gpuAllocations[g.i] + max(graphPartialOffload, graphFullOffload)
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment