Is there interest in converting the 15e BOINC siever to the f variant? The ability to sieve below the factor-base-bound means yields are quite a bit higher, which means we could tackle bigger projects on 15e and/or use smaller fb-bounds to reduce memory usage.

The same argument applies for 14e BOINC to 14f, except that nearly every project stretches 14e already so moving to 14f would be more useful. Alternately, we could make 15f its own category for big projects, and move some of the 14e jobs to 15e where they belong while maintaining the option of multiple levels of work for BOINCers.

The compilations that produced 16f should apply equally well to 15f and 14f (hopefully that means BOINC-ifying them would be equally straightforward?).

Stretching 14f into GNFS-180s and SNFS-270s sounds like fun. :)