Welcome to MilkyWay@home

Posts by wolfman1360

21) Questions and Answers : Unix/Linux : Attempting to run CPU tasks but get the following. Not requesting tasks: don't need (CPU: ; NVIDIA GPU: job cache full) (Message 68637)
Posted 2 May 2019 by wolfman1360
For the past 2 years I've been pretty much exclusively with WCG. I've decided to branch out a bit.

I've got an interesting problem. Perhaps boinc is designed to work this way and I've never realized it.

I'm very new to running multiple projects and I'm still trying to figure out what I'd like my primary to be. I really like this project for a variety of reasons - I love that it is one of the few that I know of that will use multiple CPU cores per WU for some applications (I honestly forget the specific ones), the WUs are short and have decent deadlines, and there are a variety of different applications for just about everything out there.
I currently have a prime grid GPU task in progress that has a 10 day estimate. However, I have no other CPU tasks running or in the cue from any project.

Does the work cache limit apply to both CPU and GPU combined? I would think that if the CPU has no pending work available I could still get tasks for it? I'm really not a fan of creating a 10 day cache.

What am I doing wrong here? I don't want to abort this 10 day task but I really don't want this machine sitting here idle for 10 days on the processor. That's a lot of wasted CPU. Granted, I'm sure there's a GPU a lot more powerful than mine that can handle this prime grid task, but I'd just like to know if there is a setting I can change. I've tried suspending it with no luck.

I should note that I have this machine set to the home profile which has CPU and GPU crunching enabled. I did update after hitting apply and have tried updating since. My default profile only allows crunching on the GPU.

Any help appreciated!

Previous 20

©2022 Astroinformatics Group