I am not saying to not to run GPU tasks in parallel with either EDF or non
EDF CPU tasks. That portion should remain unchanged as the GPU tasks do
only eat a very small % of a CPU. I am talking about MULTI CPU tasks
always taking precedence over single CPU tasks if the multi CPU task wants
to be in the schedule at all.
jm7
Raistmer
<[email protected]
> To
Sent by: "David Anderson"
<boinc_dev-bounce <[email protected]>,
[email protected] <[email protected]>
u> cc
BOINC Developers Mailing List
<[email protected]>, John
01/15/2010 09:28 McLeod <[email protected]>
AM Subject
Re: [boinc_dev] job scheduling
continued
>
> My point is that a non EDF task should never move ahead of an EDF task -
> even if it means leaving some resources idle. Deadlines are important to
> both the users and to the projects. A missed deadline means that the
user
> has possibly wasted a bunch of CPU time doing something that will be
> thrown
> out, and it means that some other user will possibly have to waste a
bunch
> of CPU time doing replacement work that could have been avoided. The
> project gets its work done later (important to some but not all
projects).
>
BOINC should be very good in correct estimation of deadline miss then,
cause
currently it can easely promote fresh downloaded Einstein task on my host
(einstein project has very low project share in attempt to use it as
"backup
project", conception still completely missed by BOINC) that has ~6 hours to
complete and ~14 days deadline. IF in such situation GPU-related SETI tasks
will be stopped too (they eat small fraction of CPU only, but still eat it)
my GPU will be left idle, its performance almost 1/2 of whole host
performance. For other users GPU performance can be much more than 50%. In
such situation true and real computational resource will be wasted just to
fulfill BOINC's illusions about that some CPU time _could_ be wasted
(einstein's task could miss deadline from BOINC's point of view).
It's absolutely unacceptable situation IMO.
_______________________________________________
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.
_______________________________________________
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.
