I have my computers set up with the 'special' BRP7 app and the GW CUDA app. I run the apps at 3x on both computers. When BRP7 and GW run side by side the neutron star app steals nearly all the cycles while GW barely advances. Is there a way to make them run more evenly? The last GW work unit before a long series of BRP7 work units can take a *long* time to finish.
Thank you.
Copyright © 2024 Einstein@Home. All rights reserved.
the special app is highly
)
the special app is highly optimized and will use most of the resources. Running 3x on that app is counterproductive most of the time. I would set this to only 1x.
There really won’t be a way to make them run more evenly unless you use CUDA MPS. But doing that means you won’t be able to run any OpenCL work from other projects.
_________________________________________________________________________