Hello CarloPiva,
On the web site, select My Grid - Device Manager. (a Device is just a computer with BOINC installed on it.) This should show your 2 registered computers. BOINC allows you to create 3 profiles (named home, work and school). Assign 'home' to one computer and 'work' to the other. Set up each device profile the way you want it. Be sure to select 'SAVE' at the bottom of each profile after you change it, or your changes will not be saved.

If you need more than 3 profiles, it gets tricky. You can set up a local 'Preferences' file on each computer, which will override the website profile, but that can be confusing since you cannot control it from the website. A lot of problems reported on this forum are caused by these invisible 'Preferences' files which members forget about. On each boot, BOINC inserts a line 'reading local preferences file' in the BOINC Messages tab when such a file exists on the computer. This is one of the reasons we almost always ask for a copy of the start of Messages when diagnosing a problem.

Lawrence

Why would client with any number of cores not be able to run with 0.2 or 0.3 buffer size? The client knows exactly how many cores are allowed (BOINC 6.2 introducing a CPU Processors % feature to allow mixing different CPU's, 1,2,3,4,8 etc on one profile), the efficiency they run at and how much wallclock versus CPU time is available per day to include the impact from throttling.... that is, long as you don't change the profile or local settings every day. It takes a week or longer for a change to work through to get the best achievable scheduling and work fetch.

That said, there are requests for more web profiles to include client defined, but it will be quite some time before the developers will get to that at Berkeley, if at all.

Hi astrolab,

Did not say it explicitly but the efficiency values are continuously updated and used as part to determine the amount of work to buffer, so 4 cores would result in getting 4x the amount buffered of a single core. A quad will fetch work more frequently since there are 4x as many jobs of different run times to upload when done. If you wish to reduce the work fetching, the "connect every...." is the value to play with. The Cache/Additional buffer is simply there to ensure you got the X days of work you want, ignoring most build in safeties. Someone who buffers 10 days maximum runs the risk when sudden overlong jobs come along that are tougher than estimated, to not finish within the 12 day deadline.... this is when a client hits High Priority and all work fetch seizes until normality returns. There are several FAQ pages explaining this. Here a filter that highlights the 2 that mention this event.

http://www.worldcommunitygrid.org/forums/wcg/...=&forum=3&rows=20

cheers

NB: Added "Running - High Priority" & "Earliest Deadline First" to the FAQ index linked in my sig. Suggestions for Help and FAQ items always welcome.