Hi
Believe there is a fault in the latest wus .
Just had around 80 hit the machine with a runtime of just 50 secs.
This machine normally takes about 16 mins per wu.
When crunched they all fail at 5.08 mins (32%).
Message in Boinc manager:

25/02/2012 10:18:45 PrimeGrid Aborting task pps_sr2sieve_44699365_0: exceeded elapsed time limit 307.921437
25/02/2012 10:18:47 PrimeGrid Computation for task pps_sr2sieve_44699365_0 finished
25/02/2012 10:18:47 PrimeGrid Output file pps_sr2sieve_44699365_0_0 for task pps_sr2sieve_44699365_0 absent

I can confirm this.
WUs that have an estimated RT of 1:04 min (vs ~23:15 for "working" ones) error out after ~7 min.
WU names till now were in the 447xxxxx range

same with me

http://www.primegrid.com/results.php?hostid=195518&offset=0&show_names=0&state=5&appid=

I am working on it. I will stop work generation now to I find the problem.


Lennart

Good, thanx ;-)

Just to make sure:
Are there really working "short" WUs out there now,
or are those part of the "some old" and should be aborted?

Regards
Kai

Had a bunch of them too, roughly after 300 sec they went bad.

Same here.

I'm getting new WU's and they're all still getting Computation Error. I guess there's still bad ones in the pipeline.
____________

Hello Lennart,

maybe you fixed the problem for new tasks,
but the defective ones keep being recycled after erroring out,
if I read this correctly up to 15 times !!!
See http://www.primegrid.com/workunit.php?wuid=253850333.

Is there any way for you to get rid of them for good?
The bloody things keep resetting my GPU clocks to minimum ...

Regards
Kai

All mine are still failing:

PrimeGrid 02-26-12 13:58 Aborting task pps_sr2sieve_44697242_0: exceeded elapsed time limit 478.06 (307529.96G/643.29G)
PrimeGrid 02-26-12 13:58 Computation for task pps_sr2sieve_44697242_0 finished
PrimeGrid 02-26-12 13:58 Output file pps_sr2sieve_44697242_0_0 for task pps_sr2sieve_44697242_0 absent

Switched to another project for now...

I just had one that ran 2 hrs 40 min until it failed on a Computation Error. Not good at all.

I just had a few too, 75% err, looks like there's a few still in the system, reminds me of the Chuckle Brothers, to me, to you, to me... :)
____________
147*2^1392930+1 was my first prime number found, others have followed :)

Me 2.

I have had 4 error units and no good ones in the past 36 hours.

I am still crunching in hope of good ones coming soon.

Should I abort all those ready to run? Will it help to get them out of the system?
____________
Member team AUSTRALIA
My lucky number is 9291*2^1085585+1

I'm getting a pile as well. Ive had the same driver for my GTX 580 for many weeks (290.36) without problem. About 3hr's ago started getting computation errors (-177) one after another. Message log also mentions this...

P6T-PC

46 PrimeGrid 27-02-2012 09:44 PM Restarting task pps_sr2sieve_44117917_2 using pps_sr2sieve version 139
106 PrimeGrid 27-02-2012 10:49 PM Aborting task pps_sr2sieve_44801706_1: exceeded elapsed time limit 777.62 (847722.91G/1090.15G)
107 PrimeGrid 27-02-2012 10:49 PM Computation for task pps_sr2sieve_44801706_1 finished
108 PrimeGrid 27-02-2012 10:49 PM Output file pps_sr2sieve_44801706_1_0 for task pps_sr2sieve_44801706_1 absent

Any Help?
P.S All power saving features disabled.

I have the same problem on my boinc farm (Mix of nvidia and ati). Most of the errors, over 200 at least, are the -177 or timeout. However, both the CUDA and ATI erroring WUs show a few access protection errors attempting to read address 0x10.

- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x69373320 read attempt to address 0x00000010

Engaging BOINC Windows Runtime Debugger...

Temps have been below 76c for last 24 hours according to tthrottle

Those faulty WU's are still coming and coming I have switched over to CW sieve.
____________

Those faulty WU's are still coming

I just went thru and checked all my card drivers, I hate to tell you but they ALL HAVE THE SAME PROBLEM. Here is the drivers, 266.58,285.58,275.00. So I don't know what is going on, I have been running them for quite a long time. ( like 3-9 months) 24x7 and only on prime gird work. Maybe someone ought to start looking some place else. My computers were putting up 3-4 million credits a day with NO errors, It isn't like that now................

Lonnie Christensen
____________

I did the same.

In addition, I aborted all the pps sieve WU's that I had in my queue and set my preferences to not get any new ones until this problem is resolved.

I don't mind processing a few bad WU's that take a few seconds to error out. But when they take an hour or more, that ties up my lonely computer for way too long, and affects other projects that I'm working on.

Flushed all my work units and still having problems.

Hello Admins,

if it is not possible to remove them completely, in order to get rid of the bad ones more quickly, couldn´t you at least reduce the number of retries after erroring out until the misbehaving ones are flushed out?

Look at this one for example, seems to me it will really have to run against the wall 15 times until it stops to annoy someone...

Please advise.

Regards
Kai

Since yesterday I returned only valids from my GT240.

To be sure I centuplicated the rsc_fpops_bound within the pps_sr2sieve workunits in client_state.xml and since then no problems.
____________

Hi guys, can you pelase help.
I just started on PG and I wanted to use my AMD HD7970 for PPS Sieving as it is only project supporting ATI cards, but all my WUs end up with error, this error: exit code -117 (0xffffff8b)
For example this task:
http://www.primegrid.com/result.php?resultid=355461924
It was errored for other too, but with different errors, but after me somebody finished it succesfully :(
Why all my WUs(only for GPU) end up with this error ?
My CPU unit for PPS sieve just got validated succesfully, it was first one and it was good, so how so that two days my GPU is working for no credit .
On other projects (milky and SETI) is my GPU working without any problems, so I think I have everything set up pretty much correctly...
Thank you in advance for any tips you will have, I really want to fix this thing... :(

I wanted to use my AMD HD7970 for PPS Sieving as it is only project supporting ATI cards

I still get 100% failures on PrimeGrid WUs on 7970. WUs fail at around 2% (duration ~28 minutes)

Example - http://www.primegrid.com/workunit.php?wuid=269591227

I'd like to contribute to PrimeGrid (I've been a long contributor of SeventeenOrBust) but because of this I have to dedicate all my GPU time to MW@H

I'm also getting computation errors on a Radeon HD 7750, for example:

http://www.primegrid.com/workunit.php?wuid=269963263

And in Boinc manager, the event log is:

4/22/2012 4:36:32 | PrimeGrid | Computation for task pps_sr2sieve_46767574_0 finished
4/22/2012 4:36:32 | PrimeGrid | Output file pps_sr2sieve_46767574_0_0 for task pps_sr2sieve_46767574_0 absent



This issue seems to be have been going on for some time now, when will this get fixed?

Never mind, it seems to have been a driver issue. I upgraded to Catalyst 12.4 drivers and it runs beautifully.

Incidentally it let met overclock an additional 40 Mhz =P

I wanted to use my AMD HD7970 for PPS Sieving as it is only project supporting ATI cards

Where did you find out that PrimeGrid is the only project supporting Ati cards?
Ever tried Collatz (but not right now),

Incidentally Collatz is the only BOINC project I have found for pre HD3xxx cards.