El Miércoles 15 Jul 2009 11:41:32 Richard Haselgrove escribió:
> Most project Server Status pages, such as
> http://einstein.phys.uwm.edu/server_status.php, explain that they derive
> their FLOP-eqivalence "from the sum of the Recent Average Credit (RAC) for
> all users", rather than the other way round.
>
> I do hope that the weaknesses in BOINC regarding accurate CUDA flop
> recording, highlighted by the recent unfortunate events at AQUA, do not
> lead to a premature announcement which subsequently has to be retracted
> after scientific audit.

And the number will be inflated if FreeHAL's RAC is included, which gives 
quite a few credits for running the sleep() syscall.
_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.

Reply via email to