How to optimize for BOINC? Work units are all named the same thing

Started by BobRoss, January 09, 2019, 01:47:21 PM

Previous topic - Next topic

BobRoss

I run BOINC and mostly the PrimeGrid sub-project.  The problem is; all the work unit processes are named the same thing.  For instance primegrid_cllr.exe.  When the work unit is done, it fetches another one.  My question is - How can I keep work units on their own cores if they are all named the same and new ones start at random?  Is that even possible?  I can do this manually via task manager by setting affinity but that is a huge pain.

Thank you for any help.

Jeremy Collake

The assignment of unique persistent CPU affinities to multiple instances of the same application is a feature under development. At this time, my only guidance is to wait for this. I do not have an ETA.
Software Engineer. Bitsum LLC.

BobRoss

Thank you for the reply!  I look forward to this feature! 

Cheers,
Bob

Jeremy Collake

Software Engineer. Bitsum LLC.

Jeremy Collake

As of Process Lasso v9.1.0.29 BETA, wildcards and 'chilof: processname' rules are allowed in the Instance Balancer.

That means that to cover *all* Boinc tasks, including those from multiple projects, you would use a rule with a match of:

childof: boinc.exe

See attached screenshot for example config.
Software Engineer. Bitsum LLC.