One dead DC computer from the fleet

[dajeepster]

ok... here's my spec.
i7 920 @ 3.6Ghz (wc)
MSI X58 Platinum
6gigs OCZ DDR3 1600
WD 640 black
Vista ultimate 64
two GTX260s (not in sli... was just a cruncher)
PCP&C 750 (plenty of power... so i think anyways)

Well... I think i've lost a gtx260 to gpugrid.
I didn't realize it until i noticed my gpugrid score was dropping... i went and started checking my fleet and found one had been locked up since the 16th... do'h... so i rebooted... and it locked up again.... many times wouldn't do anything.. just BSOD or locks up.
I removed the overclock to see if that was it... nah.. still same.
thought maybe memory... ran memtest 86+ for 10 hours straight.. did 16 pass and passed all no errors.
did the drive check from ubuntu disk... said drive was ok.

so... i've tried re-installing vista...no go... tried win7.. no go... tried server2008R2... no go... tried ubuntu ... no go

ubuntu gave me an error message (i'll post it later when i'm home... i'm on campus right now and doing this from memory) as to why it wouldn't install.

I did a search on the web and appearently someone else was having the same issue and it turned out to be his video card.
I'm starting to believe this, since my cards have been crunching nonstop for a very long time now... and i always keep my fans cranked to 80%.

so... when i get more time, i'm going to swap out one of the video cards and do various configurations to see what fixes the problem.

and i'll keep people updated

~dajeep

p.s.... and I have a nice shiny new gtx480 sitting on my desk ready to be installed... I'm going to have to wait a little while for that just incase my power supply is starting to go bad (which could be a good possibility since it's starting to give a low hum when the computer is turned on)

 

[Assimilator1]

Check the PSU voltages with a meter to be sure.

I was gonna say swap out the vid cards but you are anyway
GL on tracing the fault :thumbsup:

 

[dajeepster]

Check the PSU voltages with a meter to be sure.

I was gonna say swap out the vid cards but you are anyway
GL on tracing the fault :thumbsup:

I pulled out the second video card video card and I'm attempting to reload windows now... that's as far as I have time for today. I'll check again before I head to work.

I just bought batteries for my DVM too.. I'll check voltages tomorrow morning when i get off of work.... it's a nice meter too.. fluke 87

 

[dajeepster]

ok... reloading failed.. I'll attempt again tomorrow

 

[GLeeM]

since my cards have been crunching nonstop for a very long time now... and i always keep my fans cranked to 80%.

Sorry to hear about the problems

Is it bad to have fans cranked to 80%?

I've had mine at 96% for a while, maybe I better back off on the OC and fan speed?

 

[Markfw]

I run all my video cards @ stock, and 55% fan, but they are all double wide. No deaths so far. Temps never go over 75c on the vid, never over 60c on the cpu.

Sorry to hear of your loss.

 

[dajeepster]

ok.. the secondary card I put in another machine and it appears to be running fine. the computer it's going in is a P55 with an i7 860... it already has a gtx280 (wc) in it, and now has the gtx 260 in it too.

The primary card I'll stick in different machine tomorrow and let it run.

right now, it's looking like it could be the power supply... i'll check voltages tomorrow before I pull out the primary card.
I had planned on buying a new power supply on friday so this just might work out for the best... OCZZ850M is on sale at MC for $129 AR...

 

[biodoc]

Glad to hear the video card is ok.

Did you ever test the voltages on the PSU?

 

[dajeepster]

Glad to hear the video card is ok.

Did you ever test the voltages on the PSU?

not yet... that's when i get off of work this morning... but it looks like it's going to have to wait til this afternoon when i'm done with classes.

 

[Assimilator1]

I pulled out the second video card video card and I'm attempting to reload windows now... that's as far as I have time for today. I'll check again before I head to work.

I just bought batteries for my DVM too.. I'll check voltages tomorrow morning when i get off of work.... it's a nice meter too.. fluke 87

Nice .

Don't forget to check the voltages under load......err if you can get that far, if not & the PSU is the problem then you should see low readings anyway .
Max Tolerance for the 12,5 & 3.3v lines is 5%, although being close to that isn't good either. I expect you'll know all this anyway with the large fleet you maintain so I'll shut up now .

 

[Rudy Toody]

My three rigs held a séance last night and made contact on the otherside with your dead rig. It is the PSU.

 

[dajeepster]

primary video card checks out ok... thinking it's definately the power supply

 

[Assimilator1]

Lol Rudy

 

[dajeepster]

My three rigs held a séance last night and made contact on the otherside with your dead rig. It is the PSU.

your three rigs need to find a new line of work... cause it wasn't the psu... new psu and still same problem.

i've got it narrowed down to either cpu, mb, and possibly the harddrive... i haven't fully ruled out the memory yet even though memtest said the memory was ok. I'm currently running on one stick of memory and I have a new hd in the system (500Gb wd black drive). It's loading now... I'll just have to wait for it to finish... the bios screen showed normal status for the other drive and it wasn't clicking or anything...
once i get this up and running, i'll run a surface scan on the other drive to see how it is doing.
does anyone know of a good diagnostic utility for wd hard drives that will run under win7 with an x58 chipset.... wd current stuff is sorely lacking and it seems like they haven't updated it in years.

*edit* wd has updated... for thier newer drives at least

 

[Rudy Toody]

Check this link for a good hd repair and recovery program.

 

[petrusbroder]

I know Spinrite very well - it has served me well since v. 3.0 ... I agree with Rudy Toody!

 

[dajeepster]

well... seems the drive allocation table was corrupted... or something along the line.

I did a full format and then wrote 1's to the entire drive (or was it zeroes?)... after that was done, I was able to install windows with no problems at all..

except now, boinc is saying my gtx480 only does 192 GPLOPS peak... wtf?!

 

[Ken g6]

boinc is saying my gtx480 only does 192 GPLOPS peak... wtf?!

Maybe that's how many ripples it can simulate with fluid dynamics?

 

[RussianSensation]

except now, boinc is saying my gtx480 only does 192 GPLOPS peak... wtf?!

Curious, where does BOINC show the GFlops peak of your graphics?

 

[dajeepster]

Curious, where does BOINC show the GFlops peak of your graphics?

you go into the messages tab... it's near the top... when you first start boinc, it list all the configurations.. well.. most of them... of your machine.

you need a recent version too... at least a 6.10.xx version

 

[RussianSensation]

Oh thanks found it: 1408 GFLOPS. I think the program is just misreading/miscalculating your GTX480.

 

[Rudy Toody]

well... seems the drive allocation table was corrupted... or something along the line.

I did a full format and then wrote 1's to the entire drive (or was it zeroes?)... after that was done, I was able to install windows with no problems at all..

except now, boinc is saying my gtx480 only does 192 GPLOPS peak... wtf?!

Should I ask my rigs to hold another séance?^_^

 

[dajeepster]

Oh thanks found it: 1408 GFLOPS. I think the program is just misreading/miscalculating your GTX480.

well... so far I've only gotten my GTX480 it work with collatz reliably.

GOOD :
Collatz Conjecture

ERRORS:
setibeta errors out
seti errors out
milkyway errors out
gpugrid errors out (gpugrid knows this though...they've just got a couple of cards to start working with)

Haven't Tried Yet:
F@H

 

[Assimilator1]

your three rigs need to find a new line of work... cause it wasn't the psu... new psu and still same problem.

...........

*edit* wd has updated... for thier newer drives at least

Didn't you check the PSU with your shiny new DMM?

Btw yea I wouldn't relie on memtest alone for RAM issues.
Have you tested the rig with OCCT?

 

[dajeepster]

Didn't you check the PSU with your shiny new DMM?

Btw yea I wouldn't relie on memtest alone for RAM issues.
Have you tested the rig with OCCT?

it turned out to be the harddrive... but on a new note... I now have a shiny (sorta, gun metal black) psu i can build a computer around