| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: The Clean Energy Project Forum
Thread: CEP is experiencing a huge PF Delta rate |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 27
|
|
| Author |
|
|
tekennelly
Cruncher Joined: Oct 10, 2005 Post Count: 45 Status: Offline Project Badges: 
|
The CEP has a huge Page Fault Delta rate of 4000+. I am running a quad core Intel with four WCG tasks currently. I have two faah, one hcc1 and lastly one cep. Only the CEP task has a Page Fault Delta greater than zero. The two faah tasks are using 105 MB, hcc1 45 MB and CEP at 32M.
Using the Windows Task Manager I see that CEP likes a Page Fault Delta rate of 4330. Is CEP using memory in an unfriendly way with regard to Intel architecture? Environment: Quad core Intel with XP SP3 and 4 gigs of memory. |
|||
|
||||
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Try HPF2 for comparison before a "huge" declaration
---------------------------------------- Where were you during the beta testing? Ran a few and same as the beta see/saw 1000-2000 Delta, PF's in the millions, quad / vista. Nothing as uncomfortable as what happened with HCC. Looking now in comparison on the XP, after 13 hours 164 million PF with 5-7k Delta, 50 seconds kernel time, but it still is not impacting the system responsiveness in any way. The quorums my devices have been in so far have shown balanced credit near claim if that is telling. Expect the programmers to monitor comments on this new project.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
tekennelly
Cruncher Joined: Oct 10, 2005 Post Count: 45 Status: Offline Project Badges:
|
Hmmm.... I have my profile set up to accept Beta's but rarely get any so lets get beyond that complaint.
My definition of huge came from my comparison to other WCG tasks that are running in my system. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
I did nothing and got 10 betas doing a total of 5 CPU days of work.
----------------------------------------The developers really like to get tests off the widest spread of devices so it's a pity you did not get any. Is your system considered reliable and a fast returner? Large buffers of say over 36-48 hours will work against getting beta test jobs.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
tekennelly
Cruncher Joined: Oct 10, 2005 Post Count: 45 Status: Offline Project Badges:
|
The quad core turns work around in under 12 hours because it has a lowly utilized one core VMWare based Linux server running as well so all 4 cores are usually running WCG work.
I have the additional work buffer set at 1.5 days. |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges: 
|
I have the additional work buffer set at 1.5 days. Most likely that's the reason why you have not got betas. Trying to queue as many as I could I raised my extra work buffer progressively up to 1.0 day. After I passed 1.0 day I could not get any more. Back to page faults I have reported the problem in the CA forum several times and I assume the techs are aware, but they certainly have tasks of higher priority right now. We'll see what they say. As Sekerob says we have seen much worse with HCC and it has finally been solved. I would say that your PF rate is in the current average. I have reported up to 8,000 and I see one WU right now with an average 8,500 and the others between 5 and 7 k. Cheers. Jean. |
||
|
|
KerSamson
Master Cruncher Switzerland Joined: Jan 29, 2007 Post Count: 1684 Status: Offline Project Badges: 
|
Hello Skereob,
----------------------------------------I can confirm that I too did never receive any betas ! Perhaps old beta testers are prefered by distributing beta WUs ? Yves |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Back to page faults I have reported the problem in the CA forum several times and I assume the techs are aware, but they certainly have tasks of higher priority right now. We'll see what they say. As Sekerob says we have seen much worse with HCC and it has finally been solved. I would say that your PF rate is in the current average. I have reported up to 8,000 and I see one WU right now with an average 8,500 and the others between 5 and 7 k. Cheers. Jean. Just to compare to those averages, I started getting some CEP units 2 days ago. I'm now running them across 3 Quads and all 3 show similar Deltas to this screenie. While watching each system over a 5-8 minute span, I never saw any dip below 8k. I'd call 12-15k a bit high, but the units are running so no biggie. They are eating my swap files though considering the current ones I have are scheduled to complete in 24-26 hrs. Thankfully I have lots of ram and large page files...  |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges:
|
The last ones I have got in the quad (from batches E000017 and e000018) have had higher PF rates like yours, varying between 6 and 15 K. Since they were also huge the total of PFs was over the billion and the kernel time exceeded one hour for 25 to 36 hours runtime.
---------------------------------------- Cheers. Jean. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
These units were all from batches E000017 and E000018. I see an E000019 in my queue that looks a more "normal" 5 hour estimated BUT I see some behind that from batches E000021 and E000022 that are estimated at 21 hours and one is estimated at 26 hours.
The nice thing is, they're comparable scoring wise to the Rice project (from a points standpoint) so they seem to be compensating appropriately for the long run times. |
||
|
|
|