What effect do the web configuration settings have on the actual client? Say I go to rosetta@home and change the computing settings online and update the project, does it change anything? Does it only change the setting for that specific project (and if so how do the actual client settings come into play)?

The "Computing preferences" are global preferences (as opposed to the "Community preferences" and project-specific preferences) and are propagated to all projects you've added.

However, the local preferences override the online ones. So, to have the online settings take effect, you'll have to clear the local ones (if you've ever set them).

Whether you have local preferences set, you can see in the startup messages in your BOINC event log (or in the presence of the global_prefs_override.xml file in your BOINC data directory).

Gruß,
Gundolf

---

Computer sind nicht alles im Leben. (Kleiner Scherz)

There are at the website device profiles/locations, project specific and global settings. If for instance you change the CPU processors allowed from 100% to 50%, that applies to all hosts associated to the profile/location that have Rosetta attached **. Frankly, would not know if other devices that have the same member CPID and are associated to same profile, but not to rosetta [but do have other common projects] would adopt that profile's change. My guess is it would not. Settings do not transfer from one project website to another FAIK, so clients not attached to Rosetta would not get told i.e. a client just abides by the last project that had a change for the projects it's attached to.

--//--

** until you change the same setting at another project associated to same clients+profile. That would supersede [bar absence of the local preferences as mentioned above.]

Frankly, would not know if other devices that have the same member CPID and are associated to same profile, but not to rosetta [but do have other common projects] would adopt that profile's change.

From my observations, those setting propagate just like CPID does. I suppose limitations of CPID propagation apply here too.

p.s. Those really interested (anyone?) would just look in the code. I'm off.

---

I'm counting for science,
points just make me sick.