Okay but if you run another project it may happen again...you don't list what other project you are running but those percentages are based on the daily rac so if the other project gives alot of credits and or Einstein gives just a few credits for the tasks you are running then things like what you are seeing are normal. Boinc could be trying to manage the difference between 100,000 credits per day at one project and 1,000 credits per day as an extreme example. People who run multiple projects rarely see all of them running at the same time at any one moment in time, Boinc just doesn't work like that very often. ie one cpu core running Einstein forever and 3 cpu cores running PrimeGrid forever. Boinc generally runs one project on all cpu cores and then switches to the other project or projects until the daily rac numbers slowly kinda sort come together using your settings over time. Time is the key here in that it can easily take 1, 2 or even 4 weeks or more if you are running a long running project like ClimatePrediction for example, to get the daily rac's at your specs between the different projects. In your case it's 99% daily rac for one project and 1% daily rac for your other project, it's NOT "utilization" as you put it.
The other project is World Community Grid. Einstein is giving me at least 20 times more credit than WCG.
So you are down to runtime then, the Cvoid 19 tasks at WCG take me @3.5 hours while the Africa Rainfall tasks take me @12.5 to 15 hours each. Boinc is having a very hard time trying to reconcile that with your runtime of Einstein at @20+ hours for each task.
The Boinc software is very complicated and with such diverse runtimes it's easy to be on the very wide side of the calculations as far as trying to reconcile such wide apart settings you chose for your resource shares.
Like I said earlier I would run one Project at a time maximizing my output on it then stop it and run the other Project, you can do this for a week, 2 weeks or even a month at a time. This is also why alot of people end up getting multiple pc's...we want to crunch for multiple Projects RIGHT NOW. If you go down that path though be careful as the heat produced can be expensive to offset so it's more than just the electricty ro run the pc's. I currently have 17 pc's running and sometimes I maximize my contribution someplace while other times I spread them out all over the place.
Tom wrote: mikey
)
So you are down to runtime then, the Cvoid 19 tasks at WCG take me @3.5 hours while the Africa Rainfall tasks take me @12.5 to 15 hours each. Boinc is having a very hard time trying to reconcile that with your runtime of Einstein at @20+ hours for each task.
The Boinc software is very complicated and with such diverse runtimes it's easy to be on the very wide side of the calculations as far as trying to reconcile such wide apart settings you chose for your resource shares.
Like I said earlier I would run one Project at a time maximizing my output on it then stop it and run the other Project, you can do this for a week, 2 weeks or even a month at a time. This is also why alot of people end up getting multiple pc's...we want to crunch for multiple Projects RIGHT NOW. If you go down that path though be careful as the heat produced can be expensive to offset so it's more than just the electricty ro run the pc's. I currently have 17 pc's running and sometimes I maximize my contribution someplace while other times I spread them out all over the place.