If we do NOT run enforce_schedule at every checkpoint, we have to remove
the test for has the task checkpointed very recently, which means that
normal task switches will lose work.

jm7


                                                                           
             TarotApprentice                                               
             <tarotapprentice@                                             
             yahoo.com>                                                 To 
             Sent by:                  [email protected]          
             boinc_dev-bounces                                          cc 
             @ssl.berkeley.edu                                             
                                                                   Subject 
                                       Re: [boinc_dev] 6.6.20 and work     
             04/28/2009 08:07          scheduling                          
             AM                                                            
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           




> Date: Mon, 27 Apr 2009 15:42:26 -0400
> From: [email protected]
> Subject: Re: [boinc_dev] 6.6.20 and work scheduling
> To: "Josef W. Segur" <[email protected]>
> Cc: BOINC dev <[email protected]>,
[email protected]
> Message-ID:
<of69f1dbee.c26efdf0-on882575a5.006bb27c-852575a5.006c4...@sybase.com>
> Content-Type: text/plain; charset=US-ASCII
>
> We need to run the check for what should be run on a variety of events
and
> at least once every task switch interval.
>
> We need to do schedule enforcement on a different set of events, one of
>which is going to occur very frequently - checkpoint.
>
> jm7

Which begs the question if we DIDN'T do it at every checkpoint (or time)
but rather based upon TSI would things improve? I don't know the answer and
i'm sure nobody else does either, but plenty of people have their
opinion.
Given the differences of opinion regarding this, perhaps we could have a
test version that does the checking based upon events rather than
events + time. The only way we are going to know for sure if it improves
things is to try it and see what it does. My suggestion is we have a test
build that is based only on events and the ability to turn off
checkpointing as being one of those events. We can test it and provide some
facts.
MarkJ


      Enjoy a safer web experience. Upgrade to the new Internet Explorer 8
optimised for Yahoo!7. Get it now.
_______________________________________________
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.

Reply via email to