Hi! i installed Boinc client on my friend's notebook. The problem is that he doesn't keep the power on 24/7 and 6 tasks were preloaded at once. So it is likely that some of the tasks in queue will not be computed before deadline.
Is it possible to do so that task will be preloaded exactly before another one is finished and computer is ready for new computation?
Thanks.

Thanks for this answer (and question!). Short, quick and easy.

---

10:13 of Team Ninja
unofficial Team Recruitment Officer ... making a fool of myself for the team's good!

Check out our forums. Join us.

In your BOINC preferences set the "Connect about every..." to 0 and the "Additional work buffer" to 0.01. That doesn't guarantee you will get 1 task at a time but you probably will.

If you want 100% guarantee of only 1 task at a time then create a file named cc_config.xml in the BOINC data directory and add the following lines to it:

<cc_config>
<options>
<fetch_minimal_work>1</fetch_minimal_work>
</options>
</cc_config>

This second method won't download a new task until after the current task is completed and uploaded but you lose only a few seconds of crunching time compared to downloading the new task just before the current task finishes.

If you use Windows Notebook to create the cc_config.xml file be sure to save it with the .xml extension as Notebook will try to give it the .txt extension.

You must use BOINC 6.11 or greater to use the <fetch_minimal_work> option in cc_config.xml.

Thanks for answer, that helped. However there is another problem. Notebook where boinc is installed has 2 cores, so it is best to set fetch_minimal_work to 2 instead of 1. First i created cc_config.xml with fetch_minimal_work equal to 1. Now i changed it to 2, but still only one task is being computed. i rebooted twice and pushed update button of my project several times - the same problem, only one task is computed.

<fetch_minimal_work> is a binary switch! It can be either on (1) or off (0).

What about setting "Computer is connected to the Internet about every 0 days" and "Maintain enough work for an additional 0 days" in your computing preferences and switching <fetch_minimal_work> off?


Gruß,
Gundolf