Posts by CodeRedDewd

1) Message boards : BOINC client : Partial download problems (Message 74568)
Posted 4 Dec 2016 by CodeRedDewd
Post:
With GPUGRID and another, I'm having difficulty downloading project files. Files start downloading, get a few % then stop. Eventually they time out and try again getting a few more %. Eventually they say - trying again in 5 hours or so it happens so many times. I have to keep pressing Retry Now every time to keep things moving. Lots of others are having the same problem with GPUGRID. Another one was also giving me problems as well, World Community Grid, but not as badly. All of the uploads go well every time. To help me get the large files, I put the transfer timeout to 5 seconds and wait to hit Retry Now over and over until I get enough work for a while. What are my options? Is there any way to prevent the download retry from becoming so long after repeated retries and eventually backing off? If I could just get a timeout then a retry immediately after, things would work. Maybe have a limit of 30 retries or something. That's usually enough times to get most files I'm having trouble with. Any suggestions?
2) Message boards : Questions and problems : How to assgn each GPU core to dedicated projects??? (Message 33715)
Posted 9 Jul 2010 by CodeRedDewd
Post:
The problem was that my OC on my processor was too high. I'm not sure why all of a sudden after months it's causing problems, but that's just how it is. No further problems.
3) Message boards : Questions and problems : How to assgn each GPU core to dedicated projects??? (Message 33709)
Posted 8 Jul 2010 by CodeRedDewd
Post:
I have a GTX 295 with 2 cores and an GTX 280. When the 295 runs Collatz Conjecture on both cores, it will get a computation error on the core not completed when the other completes, every time. I want to assign one of the 295 cores to some other project like GPUGRID permanently, and the other to Collatz permanently, so they never run the same project. Is this possible? I've had a 9800gx2 which had the same problem as the 295. Not sure what's going on, but when the cores are always on different projects neither ever get comput errors. What can I do?



Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.