Hi cosmicboss,

Click here https://secure.worldcommunitygrid.org/ms/devi...iguration.do?name=Default , check if CEP2 is selected in "Available Projects", then scroll down to "Project Specific Settings" and select "Number of work units per host for The Clean Energy Project - Phase 2?" to 1 from dropdown menu. Then click "Save" button.

This will not do exactly what you want, but since you "know almost nothing about programing", it might help.

If you want to try app_config which will ensure you will always run exactly 1 CEP2 WU, post your Operating System and we will give you all instructions you will need to set it w/o a knowledge of programming.

Cheers and

Where's Branjo? (It's dead easy IMO, no programming skills needed, but like to see his solution to see if another road leads to Rome ;-)

Ingredients:

1) Set device profile CEP2 to 2 or higher [not unlimited]. Done!
2) Set minimum work buffer to less than total estimated runtime hours of the buffered CEP2 work (Say 2 = 12 hours), so buffer maybe 0.3 days.
3) Set app_config.xml cep2 (sampled earlier in thread by deltavee) to max_concurrent of 1
4) Put file in wcg project folder
5) Restart BOINC or read in config via BOINC Manager menu.

This ensures there's always one CEP2 running and one waiting to start. If you don't care what else runs alongside, you only need this bit in the app_config.xml:

<app_config>
<app>
<name>cep2</name>
<max_concurrent>1</max_concurrent>
</app>
<app>
</app_config>

It's better, because if you want to control the running count of all sciences, you run risk of idling cores if one or the other has no work available.

The principle of BOINC is to always run in FIFO, so with one CEP2 waiting it will become the oldest, thus the first to start, if no CEP2 is running. Long as CEP2 is restricted in device profile to 2-16 it will always be sent first, meaning if one completes and reports, the next new task issued to the client will be a CEP2 [if available of course]. In practice, there will nearly always be 2. In effect got 2 devices running like that, pretty much returning ~48 hours of results worth per day [average].

Edit: <max_concurrent> has to be 1 of course for CEP2.

What I gave you is what I said it would do "This ensures there's always one CEP2 running". (Can't say it any clearer).

Of course you need to pick in the device profile where you set in the custom section, down below, you want 2 CEP2, whatever sciences at the top. Ticking 'All' would be the easiest.

If you do that, not a second of computing time is missed [when BOINC is allowed to run].

One proviso: If you run projects outside of WCG, you will have to learn programming and understand the logic of how BOINC works and allocates computing time.

Good luck.

Where does one find a list of the current WCG (sub-) project names, as used in the <name> fields within app_config.xml?
Shouldn't WCG keep such a list somewhere in its documentation or help sections?

BTW, the parameters available (but not their available values) in the BOINC client configuration files (cc_config.xml, app_config.xml) are documented at https://boinc.berkeley.edu/wiki/Client_configuration

Here they are for the current projects. I don't recall seeing a list:

cep2
mcm1
ugm1
oet1
fahb
hst1