CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-02 09:12:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

Disable CUDA host buffers on integrated GPUs (#16308)

Browse Source
This commit is contained in:
ai-fonsi
2025-10-08 20:21:46 +02:00
committed by GitHub
parent d2ee056e1d
commit 9d0882840e
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/src/ggml-cuda/ggml-cuda.cu
View File

@@ -231,7 +231,7 @@ static ggml_cuda_device_info ggml_cuda_init() {
info.default_tensor_split[id] = total_vram; info.default_tensor_split[id] = total_vram;
total_vram += prop.totalGlobalMem; total_vram += prop.totalGlobalMem;
info.devices[id].integrated = prop.integrated; info.devices[id].integrated = false; // Temporarily disabled due to issues with corrupted output (e.g. #15034)
info.devices[id].nsm = prop.multiProcessorCount; info.devices[id].nsm = prop.multiProcessorCount;
info.devices[id].smpb = prop.sharedMemPerBlock; info.devices[id].smpb = prop.sharedMemPerBlock;
info.devices[id].warp_size = prop.warpSize; info.devices[id].warp_size = prop.warpSize;