It does not appear to be work unit specific. We will probably release another batch of similar work units (same paths through the code just different values). When the next beta will start is unknown at this point. Probably next week would be my guess (of course that guess is unofficial and my opinion) :)

Thanks,
-Uplinger

Copy that, 5 by 5.

But for the meantime, I'm missing the fun here: not a single beta-FAAH CPU-WU to date landed on my poor Ubu12.10 machine...
;
; andzgridPost#842
;

Just in case anyone is interested, I've just seen a WU with replication 8: one no-reply and SIX invalids. I don't think crunchers will be happy if this is a common situation in production. FYI, here is a copy of the status dialog:

Result Name App Version Number Status Sent Time Time Due /
Return Time CPU Time / Elapsed Time (hours) Claimed/ Granted BOINC Credit
BETA_ faah38220_ ZINC32945408_ xPR_ wC6_ 11_ 1ref9_ 00_ 8-- 703 Valid 04/02/13 01:49:37 04/02/13 05:17:32 3.40 78.7 / 78.7
BETA_ faah38220_ ZINC32945408_ xPR_ wC6_ 11_ 1ref9_ 00_ 7-- 703 Invalid 03/02/13 11:47:40 04/02/13 00:59:06 3.53 103.0 / 39.3
BETA_ faah38220_ ZINC32945408_ xPR_ wC6_ 11_ 1ref9_ 00_ 6-- 703 Invalid 02/02/13 17:27:33 03/02/13 11:47:34 4.40 45.0 / 39.3
BETA_ faah38220_ ZINC32945408_ xPR_ wC6_ 11_ 1ref9_ 00_ 5-- 703 Invalid 01/02/13 16:33:19 02/02/13 17:27:21 8.01 76.6 / 39.3
BETA_ faah38220_ ZINC32945408_ xPR_ wC6_ 11_ 1ref9_ 00_ 4-- 703 Invalid 01/02/13 02:02:19 01/02/13 16:33:14 13.68 82.6 / 39.3
BETA_ faah38220_ ZINC32945408_ xPR_ wC6_ 11_ 1ref9_ 00_ 3-- - No Reply 29/01/13 20:04:34 01/02/13 00:35:15 0.00 0.0 / 0.0
BETA_ faah38220_ ZINC32945408_ xPR_ wC6_ 11_ 1ref9_ 00_ 2-- 703 Valid 28/01/13 20:55:07 29/01/13 20:04:25 3.48 78.7 / 78.7
BETA_ faah38220_ ZINC32945408_ xPR_ wC6_ 11_ 1ref9_ 00_ 1-- 703 Invalid 26/01/13 00:09:56 26/01/13 06:17:55 2.49 75.3 / 39.3
BETA_ faah38220_ ZINC32945408_ xPR_ wC6_ 11_ 1ref9_ 00_ 0-- 703 Invalid 26/01/13 00:09:39 28/01/13 20:54:55 4.59 91.8 / 39.3

Rob,

You misunderstand me. I know that 5 is the usual maximum replication, so I am pointing out that something unusual is happening here. Furthermore, and since others have noted an increased number of invalids (which was, in itself, not to be unexpected in this beta as Keith had already flagged to us), I'm saying that WERE this new code to be released into production, so that this increased number of invalids BECAME normal, then I would have expected some complaints from the crunchers.

I'm simply saying that it seems to me that this code is probably not ready for production in its current state. More work is needed to understand why these invalids are occurring.

I'm guessing that the validator has been set to allow for more replications either just because this is a beta, or because it's a way to help the techs work out what might be causing the "divergence" that Keith talked of. We don't know if it's some random choices in the program code or if there is some difference in behaviour either of different CPUs or of different code paths caused by compiler optimisations aimed at different CPUs. It would be great to get to understand more about what's going on, but I won't cry if we don't.

I MIGHT cry if the code is released into production as it now stands.

Just my 2p-worth.

So, what's the latest news about the unfolding Invalids story in this beta-FAAH test?
;
; andzgridPost#890
;