vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Ensure device compatibility for CLBlast

Browse source 
Some devices path from Vendor, others path from System
This commit is contained in:
JackJollimore 2023-09-01 02:19:47 -03:00 • committed by GitHub
parent 736ff7c103
commit b47ee69572
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 6 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

8
README.md
View file

@ -833,13 +833,17 @@ cd llama.cpp
make LLAMA_CLBLAST=1
```
Use the following to enable GPU:
Use one of the following to enable GPU:
```
export LD_LIBRARY_PATH=/vendor/lib64:$LD_LIBRARY_PATH
```
or
```
export LD_LIBRARY_PATH=/system/vendor/lib64:$LD_LIBRARY_PATH
```
then `./main ... --gpu-layers 1`
(Note: Use `unset LD_LIBRARY_PATH` to re-link executables)
(Note: Use `unset LD_LIBRARY_PATH` to re-link executables.)
### Docker