Welcome to MilkyWay@home

sending out some more workunits

Message boards : News : sending out some more workunits
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · Next

AuthorMessage
Chris S
Avatar

Send message
Joined: 20 Sep 08
Posts: 1391
Credit: 203,563,566
RAC: 0
Message 47442 - Posted: 10 Apr 2011, 10:49:00 UTC

Perhaps an announcement on the home page that 32 bit machines need a CC update to 11.3 to work with 0.57 workunits would be a good idea? Don't forget also that some machines are "set & forget" so it could be many many weeks before everybody upgrades drivers.

Don't drink water, that's the stuff that rusts pipes
ID: 47442 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
popo666

Send message
Joined: 18 Dec 09
Posts: 9
Credit: 236,002,715
RAC: 0
Message 47454 - Posted: 10 Apr 2011, 11:36:14 UTC
Last modified: 10 Apr 2011, 11:36:34 UTC

Milkyway@home 0.57 ATI app is behaving like this on my server (2,7ghz dualcore AMD, ATI 5870 and GTX260) : sometimes it runs smoothly, sometimes it does not use my 5870 at all (progress is 1% after 5 minutes, GPU load is 0%), after reboot same WU crounches well.
Second thing is finishing of the task. When progress is 100%, some WU stay in this state for more than 30 seconds.
No errors so far, all WU are validated, with no errors and I have 16 task cached.
ID: 47454 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Chris S
Avatar

Send message
Joined: 20 Sep 08
Posts: 1391
Credit: 203,563,566
RAC: 0
Message 47458 - Posted: 10 Apr 2011, 11:51:15 UTC

I'm also running 16 cached workunits, and I also get the workunit running on for 30-40 seconds after it says 100% finished. I think that these niggles will get sorted in the next few days. When you bear in mind that we have just had a fairly major restart here, this was only to be expected.

What with Seti's recent troubles and other Boinc projects as well, it looks like 2011 is going to be an interesting test of Boinc loyalties :-))

Don't drink water, that's the stuff that rusts pipes
ID: 47458 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Michael

Send message
Joined: 14 Nov 10
Posts: 13
Credit: 3,616,915
RAC: 0
Message 47467 - Posted: 10 Apr 2011, 12:14:22 UTC - in response to Message 47458.  
Last modified: 10 Apr 2011, 12:14:36 UTC

What with Seti's recent troubles and other Boinc projects as well, it looks like 2011 is going to be an interesting test of Boinc loyalties :-))


I guess, the efficiency and the importance of the project should be the point of interest ... and not the loyalty
ID: 47467 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Chris S
Avatar

Send message
Joined: 20 Sep 08
Posts: 1391
Credit: 203,563,566
RAC: 0
Message 47473 - Posted: 10 Apr 2011, 12:50:01 UTC

I'm not sure that I can entirely agree with you but no matter. I happen to think that if you support and believe in something, then you should continue that in the bad times as well as in the good.

Sadly I expect we are going to see the usual credit chasers flitting around from project to project, as always happens in these cases.

Don't drink water, that's the stuff that rusts pipes
ID: 47473 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile banditwolf
Avatar

Send message
Joined: 12 Nov 07
Posts: 2425
Credit: 524,164
RAC: 0
Message 47477 - Posted: 10 Apr 2011, 13:11:49 UTC

How long should I expect a cpu task to take on a p4? So far 68 min @ 4%. At this rate it looks to be a long task.
Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.
ID: 47477 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Zydor
Avatar

Send message
Joined: 24 Feb 09
Posts: 620
Credit: 100,587,625
RAC: 0
Message 47486 - Posted: 10 Apr 2011, 13:37:42 UTC - in response to Message 47473.  
Last modified: 10 Apr 2011, 13:39:02 UTC

... Sadly I expect we are going to see the usual credit chasers flitting around from project to project, as always happens in these cases.


True, always happen.

Problem is more complex on this one though. The massive hit on CPUs during GPU crunching takes them out for any CPU crunching. There will always be a CPU hit during GPU chrunching - usually very small - but as things stand, the hit on CPUs with the current GPU version is too much if an individual is running CPU WUs for other projects for other reasons.

The CPU issue, is the biggest right now in terms of keeping individuals here at MW crunching GPU.

Regards
Zy
ID: 47486 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Beyond
Avatar

Send message
Joined: 15 Jul 08
Posts: 383
Credit: 729,293,740
RAC: 0
Message 47489 - Posted: 10 Apr 2011, 14:08:50 UTC - in response to Message 47486.  

The CPU issue, is the biggest right now in terms of keeping individuals here at MW crunching GPU. Regards, Zy

It's the biggest issue for me, along with the related effect it's having on other GPUs in the system running other apps.
ID: 47489 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Sunny129
Avatar

Send message
Joined: 25 Jan 11
Posts: 271
Credit: 346,072,284
RAC: 0
Message 47492 - Posted: 10 Apr 2011, 14:57:06 UTC - in response to Message 47402.  

i'm gonna stick w/ MW@H and just leave my back-up S@H work suspended. then when the double points bonus is over and done with (in a few days time i would think), i'll jump back on S@H to clear the cache. fortunately i can do that b/c the deadline for all my S@H tasks is about a month away. not sure if you can afford to keep your Collatz tasks suspended for that long though...


Well it looks like we still have a few bugs to fix so double credit days might be a little bit off.

thanks for the heads up Travis. to be completely honest though, the science and the cause takes precedence over the credit earned...for me anyways. i know there are lots of folks out there like me, and i know there are lots of folks out there who are just chasing PPD. but if i can increase my PPD without sacrificing the science/cause of another project, then its an added bonus. given that it takes no time for my 5870 GPU to crunch through a S@H MB or AP task, and given that these tasks don't need to be returned for weeks, i can afford to keep them suspended for the time being...but i'll probably switch back and forth between S@H and MW@H until you make an official announcement about the double-credit period.

thanks again,
Eric
ID: 47492 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Beyond
Avatar

Send message
Joined: 15 Jul 08
Posts: 383
Credit: 729,293,740
RAC: 0
Message 47493 - Posted: 10 Apr 2011, 14:58:12 UTC
Last modified: 10 Apr 2011, 15:02:35 UTC

It looks like at least some of the de_separation_13_3s_free WUs are failing to validate. It's kind of hard to tell for sure though unless you keep the WUs in the database longer.
ID: 47493 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Sunny129
Avatar

Send message
Joined: 25 Jan 11
Posts: 271
Credit: 346,072,284
RAC: 0
Message 47494 - Posted: 10 Apr 2011, 15:03:32 UTC - in response to Message 47432.  

I only have 5870/5970 & 6970 and the only wu's failing are those 17's for me.

unfortunately i cannot confirm this, as i have no evidence that i've gotten any 17's since the servers went back up last night. i've only seen 10's and 13's so far.
ID: 47494 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Simplex0
Avatar

Send message
Joined: 11 Nov 07
Posts: 232
Credit: 178,229,009
RAC: 0
Message 47496 - Posted: 10 Apr 2011, 15:07:20 UTC - in response to Message 47493.  
Last modified: 10 Apr 2011, 15:08:09 UTC

It looks like the de_separation_13_3s_free WUs are failing to validate. It's kind of hard to tell for sure though unless you keep the WUs in the database longer.


It seams to depend on the card, driver, OS.
In my case I have 1 PC with ATI HD 4870, Catalyst 11:2, Vista 64 and on that computer all the de_separation_13_3s_free_1_... wu's fails but on my other PC ATI HD 5870, Catalyst 11:2, Vista 64, there is no errors at all.
ID: 47496 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
cncguru
Avatar

Send message
Joined: 11 Jun 10
Posts: 329
Credit: 1,166,222,661
RAC: 0
Message 47498 - Posted: 10 Apr 2011, 15:16:17 UTC
Last modified: 10 Apr 2011, 15:17:29 UTC

I haven't seen any 17s wu's for a long time and I aborted maybe like 10 of them at the very beginning so donno what's up with those but I can say that I am running very smoothly with win7x64/5870/5970/6970 and the cal 1.4.900(10-12_vista64_win7_64_dd_ccc_ocl)drivers.
Other than the 50% CPU load everything is running great especially the way the server is acting.
I see that alot of people were not so lucky and I hope they get all the issues sorted and fixed quickly.
ID: 47498 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile banditwolf
Avatar

Send message
Joined: 12 Nov 07
Posts: 2425
Credit: 524,164
RAC: 0
Message 47500 - Posted: 10 Apr 2011, 15:37:36 UTC - in response to Message 47477.  

How long should I expect a cpu task to take on a p4? So far 68 min @ 4%. At this rate it looks to be a long task.


200 min @ 11.75%. I think I'm done running the tasks for now till its sorted out.
Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.
ID: 47500 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile dskagcommunity
Avatar

Send message
Joined: 26 Feb 11
Posts: 170
Credit: 205,557,553
RAC: 0
Message 47507 - Posted: 10 Apr 2011, 17:01:15 UTC

ok i figured out that my 3850 can compute all "free" Tasts and all fix10 or fix20 tasks have computing errors. Not that bad, im happy that i can compute anything for now *G*
DSKAG Austria Research Team: http://www.research.dskag.at



ID: 47507 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Beyond
Avatar

Send message
Joined: 15 Jul 08
Posts: 383
Credit: 729,293,740
RAC: 0
Message 47510 - Posted: 10 Apr 2011, 17:43:25 UTC

de_separation_10_3s_fix20_1_289101_1302452704
Too many errors (may have bug)
http://milkyway.cs.rpi.edu/milkyway/workunit.php?wuid=301101
ID: 47510 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
gregre

Send message
Joined: 3 Aug 10
Posts: 1
Credit: 30,020,916
RAC: 0
Message 47512 - Posted: 10 Apr 2011, 18:08:25 UTC
Last modified: 10 Apr 2011, 18:31:50 UTC

scores for Wu looks strange:



on HD 4850

oh, and i forgot, WU should not use CPU right?
now CPU is in use...
ID: 47512 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Zydor
Avatar

Send message
Joined: 24 Feb 09
Posts: 620
Credit: 100,587,625
RAC: 0
Message 47513 - Posted: 10 Apr 2011, 18:11:18 UTC - in response to Message 47512.  

In what way ?

Regards
Zy
ID: 47513 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile banditwolf
Avatar

Send message
Joined: 12 Nov 07
Posts: 2425
Credit: 524,164
RAC: 0
Message 47516 - Posted: 10 Apr 2011, 20:12:20 UTC - in response to Message 47513.  

In what way ?

Regards
Zy

I'm guessing he has it set to only use GPUs and to not use CPUs.
Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.
ID: 47516 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
KWH*
Avatar

Send message
Joined: 24 Aug 10
Posts: 181
Credit: 83,101,286
RAC: 0
Message 47517 - Posted: 10 Apr 2011, 20:31:09 UTC

MW running fine on one box(Vista) with single 5870 although higher than usual CPU usage.
MW won't run on another(XP) with 2-5830's without a lot of errors.
Both are using 11.3 drivers
ID: 47517 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Previous · 1 · 2 · 3 · 4 · Next

Message boards : News : sending out some more workunits

©2024 Astroinformatics Group