Questions and Answers :
Windows :
BOINC resource allocation
Message board moderation
Author | Message |
---|---|
![]() ![]() ![]() Send message Joined: 26 Dec 03 Posts: 66 Credit: 997,346 RAC: 1 ![]() |
Hello all, I have recently returned and been out of the loop for awhile, my question is regarding resource allocation. For instance in my case I set BOINC up for the following project priorities; 1. SETI 2. ClimatePrediction 3. Rosetta@Home 4. Einstein@Home Now after having allocated the resources appropriately on a newly installed BOINC cruncher it downloaded large quantities of Einstein@Home which if you refer to the sequence above I have set to lowest priority settings. is this a normal way BOINC handles the resource allocations? Ni newly returned, not newly knighted! Ni |
![]() ![]() ![]() Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 ![]() ![]() |
BOINC allocates resources according to a percentage of 100. So if you want Seti to have the highest priority you should set it to say 60%. For example: Seti = 60 Climate Prediction = 20 Rosetta = 15 Einstein = 5 Total = 100 Project share time is based on Recent Estimated Credit. So BOINC tries to run the projects so that all the projects produce about equal amount of credit over time. This would work well if all projects stuck to the BOINC Credit New algorithm, but many projects go their own way and assign credit on their own scales. So REC does not work as well as it theoretically should. Einstein is one of the projects that sets their own credit per task and at a fairly high rate for the flops used to calculate the task. BOINC used to calculate credit assigned based on the number of Cobblestones it took to crunch the task but now Credit New has no relationship to flops anymore in reality. One way to keep from getting too much work for a project it to reduce your number of days of work in your cache to very small values. Like 0.5 days of work and 0.01 days of additional work. This will keep the number of seconds of requested work small and prevent projects like Einstein from sending you too much work and preventing other projects from crunching. Einstein is probably the worst offender. At least on my hosts. I just set NNT (no new tasks) on the project until I am running low on their work and then I request work one cycle and then reset NNT. I still normally get more work than I want so then I abort 50% of the work received right away. That way I can manage how much work I get from them. You can also set up venues for a host to set very small cache sizes but that only works if the machine is exclusive to a project and no other projects run on the machine concurrently. The settings for all projects are global unless you set a venue for a specific machine to either Home, Office or School for exclusive preferences just for that venue. Seti@Home classic workunits:20,676 CPU time:74,226 hours ![]() ![]() A proud member of the OFA (Old Farts Association) |
![]() ![]() ![]() Send message Joined: 26 Dec 03 Posts: 66 Credit: 997,346 RAC: 1 ![]() |
Thank you kindly sir, appreciate that feedback. I had forgotten that it allocated resources in the 100 percent range, for some reason or another I had it set to 200 Seti, 100 CPDN, 80 Rosetta, 40 Einstein. Not really a good setup considering that it works only within the 100 percentile range. |
rob smith ![]() ![]() ![]() Send message Joined: 7 Mar 03 Posts: 22652 Credit: 416,307,556 RAC: 380 ![]() ![]() |
You can use the full 0-1000 range for each value, but make sure you use either % or full range for all the projects - don't mix or it will get very confusing. Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
©2025 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.