BOINC Linux GPU
As of version 7.0.40 BOINC has the ability to split GPU tasks across multiple work units.
Preliminary Change Log 7.0.39 –> 7.0.40:
- limit the number of concurrent jobs of a given app (e.g. for WCG apps that are I/O-intensive)
- Specify the CPU and GPU usage parameters of GPU versions of a given app.
BOINC development versions are found here: http://boinc.berkeley.edu/download_all.php
Create a new file named
app_config.xml in the project directory to make use of the new options.
Details are found under ClientAppConfig of the BOINC wiki.
On Ubuntu the project directory is found:
<app_config> <app> <name>hcc1</name> <max_concurrent>4</max_concurrent> <gpu_versions> <gpu_usage>.25</gpu_usage> <cpu_usage>1</cpu_usage> </gpu_versions> </app> </app_config>
name: Short name of the project. Project short names are found in the
max_concurrent: This specifies the maximum number of tasks the app can run at a time. Set this to the number of total HCC1 tasks your computer will be doing at a time.
gpu_usage: This specifies the % use of the GPU per task. Set this according to how many tasks you want to run per GPU.
Tasks per GPU = gpu_usage
cpu_usage: This specifies the % use of CPU threads per GPU task. The default value is 1, but if you want to run more GPU tasks than you have CPU cores edit it accordingly.
Total CPU Threads / Total GPU Task = cpu_usage