PDA

View Full Version : BOINC: CreditNew flawed (surprise!)



zombie67
12-14-13, 11:38 PM
Over on the SETI forums, some brainiacs have done some investigation into CreditNew, and discovered it is flawed. Who knew?

The math is all way over my head. But if I understand correctly, it is under awarding credits at S@H by 3.3 times for multibeam, and 1.65 times for AP. Not sure the impact for other applications at other projects.

http://setiathome.berkeley.edu/forum_thread.php?id=73306&postid=1444197#1444197

Slicker
12-15-13, 12:16 AM
It doesn't affect Collatz because Collatz doesn't use CreditNew. K.I.S.S. is my motto. Fancy algorithms just lead to more bugs.

zombie67
12-15-13, 01:10 AM
Yeah. I should have been more precise:

Not sure the impact for other applications at other projects...that use CreditNew.

shiva
12-15-13, 09:35 PM
interesting reading, but it really seems that those for and those against can never find any kind of common ground. I would guess that DA will continue to think new credit is a good thing.