Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

allow Boinc to use a variable number of cores #1590

Closed
hpvpp opened this issue Jul 27, 2016 · 5 comments
Closed

allow Boinc to use a variable number of cores #1590

hpvpp opened this issue Jul 27, 2016 · 5 comments

Comments

@hpvpp
Copy link

hpvpp commented Jul 27, 2016

On multi-core cpu's allow the user to specify both the minimum and maximum number of cores available to Boinc.
The idea is to allow Boinc to reduce its cpu-load in a stepwise rather than all-or-nothing manner.

@ChristianBeer
Copy link
Member

I don't understand when to use a stepwise reduction on cpu load? Do you want something that is described in #41?

@hpvpp
Copy link
Author

hpvpp commented Apr 13, 2017

rather, on multi-core systems:
while the value in "suspend when non-Boinc cpu usage is above n %" is exceeded {
reduce the number of cpu's used by Boinc by one
}

@ChristianBeer
Copy link
Member

This is what already happens because science apps have the lowest priority. What the suspend eliminates is disk IO, that the user may need for something else and the need for the OS scheduler to consider low priority processes when scheduling cpu runtime. So I'm not quite sure what the advantages of the stepwise reduction is compared to just setting a higher value and complete suspension.

Asked another way: What's the overhead from having L low-priority and N normal priority processes for C cores running in parallel with L=C and N<=C compared to reducing low-priority processes and have a ratio like L+N=C and N<=C. Normal priority processes means non-BOINC processes.

@hpvpp
Copy link
Author

hpvpp commented Apr 14, 2017

It just seemed to me a more intuitive way of doing things.

@AenBleidd
Copy link
Member

I think more general approach in #41 is better.
If you think that I'm wrong feel free to reopen this ticket.

@AenBleidd AenBleidd closed this as not planned Won't fix, can't repro, duplicate, stale Oct 23, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

3 participants