R@H Errors Anyone?

[Sean D]

Well, I decided to check on my main cruncher, and this is what I got:

Computation Errors

Has anyone had this before? Did I have too much work queued?

Well this sucks. That first one took up over 6 hours of crunching time.

 

[JC]

Are you overclocking? My 165 returned lots of errors when I neared its limits.

 

[Sean D]

Yes, but I am not pushing it hard. I haven't increased the overclock since I first put it together. I actually downclocked it a few weeks ago. It is a 144@2.6ghz. Temps seem the same as always. It has always ran the same speed since I started R@H, and I've never had problems previously. Maybe it was just a bad batch, idk.

 

[Wolfsraider]

I am not, all mine eem to be running good, no errors like that.

I wish I could be of more help.

Mike

 

[BespinReactorShaft]

:Q That's a LOT of compu errors. I only get the occasional one or two maybe a couple of times a week, which I assume is just a peculiarity of R@H.

edit: My rigs are all at stock speeds, fyi.

 

[Sean D]

Ugh, just downloaded a new batch of WUs and am getting the same errors. They all seem to be hanging at 1% Initializing, but none will initialize. I wish I had time to work on this now, but I am going out. I exited the client for now, and later I will try to set my comp to stock speeds, and see if that makes a difference.

 

[Insidious]

Sean,

I really don't think that many errors would come from 'bad work units'

It looks like somehow your Boinc Install got borked or else there is a system issue at work here.

If it were me, I'd first try a re-install of boinc/Rosetta (detatch the project first)

If those results don't change, I really do think it's time for a round of troubleshooting.

Any way I can help, just PM or post... I'll help in any way I can.

-Sid

 

[Sean D]

Thanks for the advice Sid. I downclocked my CPU back to 2.25ghz, and I'm gonna let it run while I'm out. If that works, then thats good, but its bad too. My system should be stable at 2.6ghz I don't know why all of a sudden it wouldn't be, I haven't changed anything recently. Oh well, if I have to live with a slower comp, so be it. I'll let you know how it goes tonight.
Thanks again,

Sean

 

[Malak]

I've gotten a couple computation errors myself. One got it in the first minute, no wasted time. I do not overclock, but both times I was playing a game at the time, so maybe it was just stress on the processor. Seems odd, never seen it happen before.

 

[networkman]

Originally posted by: Malak
I've gotten a couple computation errors myself. One got it in the first minute, no wasted time. I do not overclock, but both times I was playing a game at the time....

Eek! :shocked: Blasphemer! Using CPU cycles for something other than DC?!? :Q :Q

 

[Malak]

Originally posted by: networkman

Originally posted by: Malak
I've gotten a couple computation errors myself. One got it in the first minute, no wasted time. I do not overclock, but both times I was playing a game at the time....

Eek! :shocked: Blasphemer! Using CPU cycles for something other than DC?!? :Q :Q

I can't help it man! NFSMW is crazy sweet!

 

[Insidious]

We game over Rosey on a daily basis here.

Prime95... naw. My best stability tester is a couple teenagers on a rainy Saturday!

:beer:

-Sid

 

[Wolfsraider]

LOL so true so true :thumbsup:

 

[BlackMountainCow]

Sean, that many errors can't be due to bad WUs. I'm most positive that it's your machine.

Did the room temperature change?
Did somebody hit the PC and the heatsink got unseated a bit?
Too much dust in your PC?
Did you ever test the OC with Prime95/3DMaek/other burn-in software?
Is your memory memtest stable?

I know, many questions, but in this case, it's not BOINC or Rosetta, at least in my opinion.

 

[Sean D]

Well, after downclocking the CPU further, and cleaning the dust out of my comp, it looks like all is well. It just sucks I have to run my comp at a slower speed. I don't understand it, however. When I first OC'ed, it was Prime95 stable, and Memtest86+ stable at 2.7ghz. I've got it at 2.5ghz now, and it looks like things are running smoothly. Thanks for the help guys. :beer:

 

[mondobyte]

I've just run into a rash of Rosetta WU that are crashing ...

most of the computers are entirely stock that are crashing ...

Only 2 different errors ... but lots of them.

mondo

 

[Insidious]

SeanD, Great to hear you got them settled down. If it's any consolation, I am running two X2s and they are both at 2.5GHz.... best I could do and have total stability.

Mondo, lots of craching? ouch. If you figure that one out, be sure to let us know. I've been incident free on Rosey for quite a while now.. I'll watch mine closer just in case.

-Sid

edit: are you running multiple projects on each computer? which ones?

 

[petrusbroder]

Rosetta is running on 5 computers 24/7, sharing with other BOINC-projects, have had 35 WUs in the last 36 hours (or less ... ) and no errors.

It seems - from other threads too - that Rosetta is very sensitive for OC-ing: I had a Athlon-XP-1900+ running at a stable 22% (1950 MHz) overclock, all seemed stable (Prime98, burn-in programs, memtest etc.), no problems with seti classic, seti-BOINC, Folding@home, einstein, CPDN, LHC, SIMAP, Predictor, DPAD, or any other program except Rosetta@home. When I decreased the OC to 15% all went well, except an occasional 1% error. When I ran the comp @ stock-speed no problems at all for a long time.

On a PIII-comp all is well when running a 866MHz PIII @ 1080MHz even with Rosetta; so I am asking myself if the OC-ing-problem is one of the "final speed" when OC-ing :Q , because when I OC-ed a P4 2.8GHz and a Athlon 64 3000+ I got problems with Rosetta almost at once... all other programs ran perfectly stable. :Q

 

[Insidious]

I suspect Rosey isn't handling 'common' read errors well.
Perhaps an error (yes they happen all the time in PCs, especially overclocked ones, as they run) that would be re-read in most apps. is hozing Rosey.

=Sid