A user should have the ability to tell camb_boinc2docker how many cores it can use, using all four cores over-taxes my computers.

I don't want to restrict what applications Cosmology@Home can run, just lessen the burden on my computers I volunteer your applications can use.

LHC@Home has that preference, Cosmology@Home should as well.

I second that request - was just looking around on my settings page for something like that but none exists. Therefore I run this project infrequently as it is a huge resource hog (as is Climate Prediction).

---

Please see 'How can I limit the number of CPUs used?' in the FAQ section.

Please see 'How can I limit the number of CPUs used?' in the FAQ section.

That doesn't address the issue, really.

Cosmology@Home needs to make it an EASY TO USE configuration setting in their preference section, not a computer hack that many people won't use because of the complexity of the file editing.

Try turning off the camb_boinc2docker tasks for a while and see if it is better with just the legacy tasks. Virtualbox uses resources at a higher priority than the way Boinc does and it causes problems for some people.

A lot of people at LHC use the method from the FAQ. It came from Crystal Pellet, an active user in the LHC forums.

Your other option is to try turning back the processor usage in the Boinc preferences section but that will affect all your projects.

If you need help with the file, you can post here or at LHC and someone will help. It is really just pretty much a copy and paste in a text file, save as a *.xml file in the correct directory and restart boinc.

Hello,

after changing the app_config.xml name and app_name in this lines
<name>camb</name>
<app_name>camb</app_name>

from camb_boinc2docker to camb
Boinc-Manager(7.8.3) accept this app_config.xml.
But there is no work downloading with docker.
Preferences for C@H are set to get work from both.
This is the configurated app_config.xml for my Computer:

<app_config>
<app>
<name>camb</name>
<max_concurrent>2</max_concurrent>
</app>
<app_version>
<app_name>camb</app_name>
<plan_class>vbox64_mt</plan_class>
<avg_ncpus>2</avg_ncpus>
</app_version>
</app_config>

Edit: now Boinc-manager say plan_class is fault:
<plan_class>vbox64_mt</plan_class>
After deleting this line all is well.

Most projects that have mt or vbox apps do not even have an example app_config in a FAQ. I have come here in the past just to copy the app_config setup for other projects. I have linked to it on other project forums as well. It's no hack. Paste it in a text file and rename to app_config.xml. The event log will mention the correct app name/plan class to enter in the file.

Thanks mmonnin,
this app_config.xml:
<app_config>
<app>
<name>camb</name>
<max_concurrent>2</max_concurrent>
</app>
<app_version>
<app_name>camb</app_name>
<avg_ncpus>2</avg_ncpus>
</app_version>
</app_config>

works for camb_legacy 2.17. 2 CPU's are shown using in Boinc 7.8.3 for each task.
The duration time is shorter than with one CPU.
Boinc_docker work isn't downloading anymore since the beginning.