Welcome to MilkyWay@home

Posts by MossyRock

1) Message boards : Number crunching : Suddenly, many invalids (Message 70930)
Posted 27 Jun 2021 by MossyRock
Post:
Thanks, Ron. Now I know that others are having the same issues, and that a hardware fault in my system is not likely.
2) Message boards : Number crunching : Suddenly, many invalids (Message 70922)
Posted 26 Jun 2021 by MossyRock
Post:
I suddenly starting getting large amount of validation errors on my WUs.

There is definitely a pattern. All of the validation errors are when my results which ran on one of my CPU cores are compared against those which ran on someone's GPU.

Nearly all of mine that validated ok were when there was apples-to-apples comparison, i.e., a WU which ran on my GPU was validated against another's WU which also ran on a GPU.

An example of one of my invalid WUs: https://milkyway.cs.rpi.edu/milkyway/workunit.php?wuid=121726783

I have read this thread and it doesn't seem to apply to situation, exactly: https://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=4727

Any ideas?

Thanks.
3) Message boards : News : Nbody 1.68 release (Message 67027)
Posted 2 Feb 2018 by MossyRock
Post:
Tom,

Gotcha.

Thanks for letting me know how to find the mis-matches in the "ready to start" state. I just aborted a few.
4) Message boards : News : Nbody 1.68 release (Message 67025)
Posted 1 Feb 2018 by MossyRock
Post:
Sidd,

Most of my v168 runs are blowing up. Do I abort the v168 runs in queue?

Thanks.
5) Message boards : Number crunching : N-Body Simulation (mt) WU (Message 66673)
Posted 27 Sep 2017 by MossyRock
Post:
Thank you, Darrell.
6) Message boards : Number crunching : N-Body Simulation (mt) WU (Message 66670)
Posted 27 Sep 2017 by MossyRock
Post:
I've been away from this project for some time and just returned today.

I was gobsmacked when I saw a WU waiting in the queue that said "8 cores".

This is the first multi-threaded application I've ever seen in BOINC.

So, to test it, I suspended all other jobs and let it run. It was over in less than 10 minutes.

My question is, how does the BOINC manager schedule this when I have multiple projects running? Will it force the other projects and WUs to stop so it can get all 8 cores, or do I have to make tweaks/allowances somewhere to make sure these get run?

Thanks.




©2022 Astroinformatics Group