Quote:
A few years ago there was a serious effort to adapt the FAH uniprocessor client to the BOINC environment. This project did not succeed but it did discover quite a few differences in philosophy that were difficult to reconcile. At this point, I don't see any reason why a new project is needed that would force a reconciliation of those differences. Both FAH and BOINC seem quite capable of running without assistance from each other.
FAH scheduling is much more demanding than the BOINC infrastructure is designed to support. In fact, the newest FAH clients increase the scheduling demands by using the full resources of a Quad CPU, a GPU, or a PS3 to return work more quickly rather than allowing the allocation of a fraction of those resources. That simply means that you need to choose which you want to support.
|
In short, there will doubtful ever be an F@H client for BOINC because the deadlines are too strict for F@H. Plus if one runs multiple projects, as I do, the F@H will unlikely have enough time to complete a work unit.
I could always go back to set BOINC on one core and F@H on the other... but not fair as my BOINC runs a dozen projects to F@H's ... one.