Quote:
Originally posted by wblipp
It's an interesting design question whether P-1 (and possibly P+1) factoring should be done separately, like sieving, or should be integrated into the SoB client to be performed automatically as part of every test.
One advantage of integration is that the testing will always be delayed until the last possible moment, so there is every opportunity for a sieving factor to arrive first.
One advantage of separate tests is that stage 2 of P-1 and P+1 tests work best with large amounts of memory, so we could dedicate big memory machines to this testing.
Another interesting design issue will be determining our optimization points. Our optimization points should be different from GIMPS because they always need a double check and we don't really know how much double checking we will be doing.
i think it will be best to start it out kinda like sieving and then make it a server project. probably seperate but i'm not sure. we'll see.