| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Computing for Sustainable Water Forum
Thread: Computing for Sustainable Water 64 bit Applications Released |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 31
|
|
| Author |
|
|
uplinger
Former World Community Grid Tech Joined: May 23, 2005 Post Count: 3952 Status: Offline Project Badges: 
|
Greetings,
We have released the 64 bit versions of the Computing for Sustainable Water. The Beta test went very well and you will notice the 64 bit version of CFSW as 6.12. The 32 bit version will stay at 6.11 as no changes were made to that application. Thanks, -Uplinger |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Just noticed the first 2 arrived 5 minutes ago. Pushing them to run now as gamma test ;)
----------------------------------------Edit: The graphics window of 6.12 is a bit too small. Running 6.11 and 6.12 on Windows 7. Imho, no need to change. [Edit 1 times, last edit by Former Member at Jun 15, 2012 4:50:39 PM] |
||
|
|
Dataman
Ace Cruncher Joined: Nov 16, 2004 Post Count: 4865 Status: Offline Project Badges: 
|
Thanks uplinger; more good news.
----------------------------------------  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Great... leaving the question asked in the Beta test thread: Will the host be tested to see if the 64 bit is faster than the 32 bit (can't test this here since the 64 bit is faster), then remain on 32 bits if slower? It was said in the Beta test OP, it [the 64 bit issue] could run as much as 15% less speedy than the old 6.11. Not seen any beta test comments of running longer, but then that was a small sample size and many probably doing the zzzz through it [never knew]
--//-- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Just curious: In midst of some 6.12 jobs I got a repair job for 6.11. So have repair jobs generally more priority than the 64/32 bit differentiation?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello uplinger
Reference: uplinger [Jun 15, 2012 4:04:00 PM] post If there were no changes to the CFSW 32-bit version of today at production, then what was the recent CFSW 32-bit beta version all about? If those CFSW 32-bit beta versions were for 64-bit machines (alongside CFSW 64-bit beta versions for 64-bit machines), then what was the relevance of having 32-bit machines be qualified in-, else not disqualified from-, downloading the CFSW 32-bit part of the mixed 32/64-bit CFSW beta test of recent? Wouldn't it have been more logical that only 64-bit machines participated in the recent CFSW beta test? I seek only to understand and I appreciate 64-bit versions of apps being developed and pushed to realization. ; |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I have no trouble doing *educated* guesses from prior information: It was about cross-validation of a known set of results... which found they were so close in mixed quorum of 6.11 [32 bit], and 6.12 [64 bit] of the fast release too, that it eliminated the worry of having to do app matching [see post knreed], such as during the transition/upgrade. The mixed quorums validated properly [32 bit machine meeting 64 bit machine]. Per same knreed, there will though be homogeneous groups of 32 and 64 bit IIRC. One of the advantages to that is also credit pairing. Since we've moved on into production [and happily see my Linux now slightly outperforming W7], the *will* transitioned now to *is*. :D
------------------------------------------//-- P.S. It is important to note that a 32 bit can execute faster [on a 64 bit machine, than a 64 bit compile]. That bit of testing I'm looking forward to hear on. Clients would get assessed by giving a set of each, after which the server would lock in a client to run the 32 bit, would that consistently run faster. If it were fringe... 1-2% I'd though not bother with that further complicating mechanism. [Edit 2 times, last edit by Former Member at Jun 16, 2012 10:45:39 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
To add a wonder, my 64 bit Linux which tested faster with the 64 bit science app is still receiving loads of 10 day deadline 6.11. The last 6.12 received was 12 hours ago [23 in queue on this quad]. There's no quota message, since work is received, but could that be the cause? The BOINCtasks history does not show any of the 6.12 actually processed, so putting a set ahead to see if this forces the hand... the 6-8% faster has strong appeal on me [both on W7 and Lnx]. :D
------------------------------------------//-- edit: And have frozen fetching for a little, as to ascertain that the next call will be for 6.12 backfill (if it will). [Edit 1 times, last edit by Former Member at Jun 16, 2012 11:15:57 AM] |
||
|
|
genes
Advanced Cruncher USA Joined: Jan 28, 2006 Post Count: 132 Status: Offline Project Badges: 
|
With these new faster CFSW versions (both 64- and 32-bit), is this not causing a bit of server overload? Would it be better to modify the now much more efficient app to include more work?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Long as we are not exceeding the [accidentally] tested 1.5-1.7 million daily, we're good [see post kneed on this matter]. Whilst, the GPU tasks for HCC will get the 5 images or more included, to mitigate the upcoming issue. Of course, when the push comes to shove, that [more jobs in one task], will come to the fore again... either that or as was done in past, to slow down the priority [yes it was], as not to let the whole fall over due the one.
--//-- |
||
|
|
|