tcsenpai/ollama
1
0
Fork 0
mirror of https://github.com/tcsenpai/ollama.git synced 2025-06-07 11:45:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

Remove mmap related output calc logic

Browse Source
This commit is contained in:
Daniel Hiltgen 2024-06-13 09:59:36 -07:00
parent 6f351bf586
commit 17df6520c8
1 changed files with 1 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
llm/memory.go
View File

@ -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)