Message boards :
Number crunching :
Not Playing Nicely With Others
Message board moderation
Author | Message |
---|---|
Send message Joined: 4 Aug 11 Posts: 1 Credit: 13,046,918 RAC: 0 |
I'm running two projects, Milkyway@Home and VirtualLHC@home. I'd like LHC to run all the time, and Milkyway to run on the rest of computer (I have an 8-core box, 16-core with hyperthreading, and an NVIDIA 980 GPU, on 64-bit Windows 10). I set the resource share for VirtualLHC to 300 and MilkyWay to 100, but no matter what, if there are ANY work units for MilkyWay, VirtualLHC never runs. I have read on forums "just let it go for a while, BOINC will balance it out after they both run." I just got back from a 10 day vacation and LHC didn't run once. I let MilkyWay drain, and LHC kicked in. After a day of running solo, I opened MW back up, and LHC is once again Waiting to Run. Any advice on how to force MilkyWay to let LHC have a couple of cores all the time? If I can't get it to play nicely I'll have to exit MW. |
Send message Joined: 24 Jan 11 Posts: 708 Credit: 543,855,222 RAC: 126,905 |
Your problem is the very short turn-a-round time of MW tasks. A couple of solutions. Use the venue mechanism to assign MW to a different venue than your LHC tasks. Set a very low cache setting for MW ..... like on the order of 0.01 days. The other mechanism is to set the project resource share to 99 for LHC and 1 for MW. |
©2024 Astroinformatics Group