Realtemp and nvidia cards crashing?

[Ketrel]

I have a question. A while ago, I posted a question about a potential defective card ( In case I did the related thread wrong, http://www.tomshardware.com/forum/252304-33-defective-card ).

I traced the beginning of that happening to when I started using Realtemp to get the temperature of my CPU.
I since stopped using realtemp and have not had it happen since.

So I must ask, has anyone else ever had something similar happen with their nvidia card (mine is geforce 8800 GT) when using Realtemp?

 

[unclewebb]

What version of RealTemp were you using? A memory leak was fixed after version 2.60 which might have been causing a problem for some users.

You can download the latest beta here:
http://www.fileden.com/files/2008/3/3/1794507/RealTempBeta.zip

If you have a previous version of RealTemp installed then download the above file, unzip it and copy the new RealTemp.exe into your RealTemp directory. Long term stability of RealTemp has been greatly improved.

Version 2.70 should be finalized and ready for download from TechPowerUp later this week.

 

[Ketrel]

I was using 2.60

Is my problem in the other topic a possible result of a memory leak or is it unlikely. All I know is it hasn't happend since I stopped using realtemp. :\ (Which sucks, cuz I liked it)

 

[unclewebb]

Personally, I think your problems mentioned in that other thread are separate from RealTemp. After 50,000 downloads there have only been 3 users that have reported problems with RealTemp v.2.60. Considering the number of users that overclock everything, I think that isn't too bad a percentage.

Two factory overclocked cards in SLI is probably your biggest problem. If you bring the clock speed down a little and a bunch of errors go away then I think it's pretty obvious where the main problem lies. How is your case cooling as well and have you run any memory testing programs. Simple things like memory failing can cause a lot of random issues for people.

Another thing to check is your power supply voltages with a digital multimeter. Lots of users don't have an adequate power supply no matter how many watts they're rated at. When running a 3D bench as well as Prime95 it's a good idea to measure your 12V and 5V rails to make sure there isn't a voltage sag.

When your system is stable, give RealTemp another chance and see how it goes. I'll post here when version 2.70 is officially released. There are lots of new features and so far it's rock solid stable even during extended tests.

Post some more info about your present settings like a few CPU-Z screen shots and see if your computer is stable enough to run Prime95 small FFTs on all cores for a few hours. With some more testing I might be able to trouble shoot this better.

 

[Ketrel]

Well I'm hesitent to give Realtemp another try. What I can say is this

1. I had already underclocked the cards back to stock due to heat.
2. I disabled SLI for the time being.
3. It happened using only one card.
4. It started happening within a day of using realtemp, and it hasn't happens since I stopped using it.

As far as the other things.

1. I've used memtest. (no errors)
2. I'm not sure what you mean by any of the other things.

 

[unclewebb]

RealTemp 2.70 is now available.
http://www.techpowerup.com/realtemp/

Well I'm hesitant to give RealTemp another try.

I understand that and won't try to twist your arm.

One user sent me a screen shot of RT2.60 that seemed to show a significant memory leak but I ran that version over night with the logging feature set to 1 second and with Gamer Mode turned on and it was still running fine the next morning and the memory usage was right where it was supposed to be. If there was a problem, the memory usage would have kept growing and growing but it didn't.

If your computer is stable then try running Prime95 small FFTs on all cores. You should also run something like 3DMark in demo / looping mode after the Prime test. Use CoreTemp or Everest to measure temps to eliminate RealTemp as a source of your problems. It's always a good idea to measure your power supply voltage with a digital multimeter while running a stress program like Prime95. Don't rely on software to report voltages on a computer because the sensors on most motherboards are not accurate.

RealTemp 2.70 changes how the MHz is calculated and also has a few other tweaks for better efficiency. When minimized to the SystemTray over night for 9 hours, the amount of CPU usage compared to the competition is obvious.

 

[anartik]

Kinda stumbled across this thread but I have been pulling out my hair trying to figure why my system all the sudden became unstable and this hit a chord. I have a Rampage formula X48 running at FSB 1800, E8500 at 4.275, OCZ DDR2 1066 at 1088, 2 WD6400's in RAID zero and an Asus 8800GTX overclocked near ultra speeds. The system is powered by a Corsair HX1000 in a Lian Li PC-V1010 with lots of cooling. I spent a considerable amount of time getting it stable so I could run anything from Prime95, 3DMark06 to hours of COD4. I had been using CPUID Hardware monitor and was not believing the temps I was seeing. The first time I ever used RealTemp was 2.7. Perhaps it is all coincidence but I started having wierd occurances like my monitor (Samsung 2693HM) turning off randomly, the monitor menu's coming up, screen locking up or the system hanging when loading games like COD4. As I recall the problems all occurred either during or after running RT 2.7?

 

[unclewebb]

"Microsoft's data strongly indicates that the problems were real. Damon Poeter at CRN dug through the documentation to find that on page 47 of the PDF, NVIDIA drivers were identified as the cause of over 479,000 crashes, or just under 29 percent of all the crashes Microsoft logged."

http://arstechnica.com/news.ars/post/20080325-vista-capable-lawsuit-paints-picture-of-buggy-nvidia-drivers.html

I've personally run RealTemp 2.70 for 65 hours without any issues and other users have reported to me that they have gone 100+ hours.

There is absolutely nothing in RealTemp that has anything to do with your monitor. I switched to ATI long ago because I got fed up with Nvidia and their drivers.