Forceware 196.75

drivers almost killed my cards =C

 

Same here. For me it happened when leaving sleep mode I believe.

 

This set of drivers has a bug with DX10 AA in Bioshock. Some light reflections over water causes pixel errors.

 

I'll keep this set on as the 196.34 beta ran my cards a bit too warm for my liking. I don't use hibernation or sleep on my rig and have my fans set to 70% via Precision (plus monitor temps via my G19).

 

I noticed that too and know what area you're probably in, but I'm wondering now in retrospect if it was pushing some part of the core too hot.

 

I am using these drivers since their release date, while monitoring temps and fan speed in the background, without having a single issue. Yesterday i had one, and i want your opinion. Keep in my mind that my card is at stock clocks, and i don't use a third party software for fan control.

I was playing BFBC2 and i heard my fan louder than ever. I continued to play and every five minutes the game freezed for 3 seconds, and resumed without a crash or a single artifact.
After 2 hours i finished my BFBC2 session, and i checked the gpu-z log file. As max temp was reported 81c, but as max fan speed 93%. I never saw in the past such a big fan use. I searched the log file, and i found out that after 1 hour of gaming, the fan raised and never returned in normal levels. ex. it was at 75% while the vga's temp was 57c. At 57c the fan should stay at 40%(default), and start raising after 60c.
What made me preoccupied, was that i had even realtemp monitoring my cpu speed in the backround. In realtemp there is gpu tab, which i checked and found out that the max reported gpu temp was 117c.

Do you think that is possible that my card reached such a temperature, even instantly ?

After all that, i had an NBA2K10 match (max temp 59c, max fan speed 40%), and ran vantage (max temp 77c, max fan speed 56%). Not a single artifact or a problem in both cases.

 

I guess NVIDIA knew something wasn't 100%

 

Never had any problems with drivers that cause my hardware to die.

I think that all these people that got their card melted just pushed their card too hard.
By overclocking and not using proper airflow.


*EDIT*

I did notice that the drivers is removed everywhere, so im going to remove it, just to be safe.

 

BFBC2 ( PC ) is basically flawed. The application at the current version is unable to spawn and gain the exclusive access to the video device.
Running such application without Vsync could indeed trigger, on non "golden" ( nvidia ) card your forth mentioned issue.
For the chronicle, it is mainly an directx issue, perhaps a broken inplementation of both's DX10 pipeline and software OpenGL in some sort of disguised butterfly effect, and I have no idea if ATI Driver is affected too, so; draw your conclusion.

 

As english is not my native language, can you be more specific ?
What i want to know is that if somebody thinks that a GTX285 can hit 117c without making the pc shut down and without hardware damages as a consequence.
I also want to add that till that happened, i check the gpu temp reported by realtemp, and is always the same with the one reported by gpu-z. Only yesterday gpu-z reported 81c and realtemp 117c.

 

If the analog/l²C thermal solution mounted to your card would hit 97° Celisus ( + 8° compensation), the device would throttle down.

 

I verified it with older driver version and monitoring temps. It's a driver bug with AA, is not a hardware malfunction.

 

As i said, the game freezed for 3 seconds and resumed normally after that. I talk about a complete freeze, during which i wasn't able to do anything. There was only a static image in my screen, something that never happened before in this installation of my system (1year+).

We cant say that the card was throttling during those freezes ?

 

If the fan stopped for a moment, for example.

I'd recommend un-installing this driver and going back to a previous one.

Also: so far nobody reported a problem with the 196.86 either.

 

Throttling nope, you will notice that for sure, the device shouldn't be able to resume. Furthermore on Vista/Windows7 you should get at least one “driver display reset” alert.
Nvidia is recommending to rollback, anyhow: these driver are based one the big bang first release. So, unless someone wirte some well-documented things in regards to this "incident"; do what u must.

 

Thx HH!

 

for my card 96C only does 88% fanspeed (this with 196.36) with default auto fan control, think ive done past 106C without shutdown

so yes most propably u been pretty high temps, like cooking

 

I use to manual control the fan speed and mine sandwich x2 GTX295 non-single PCB doesnt goes over 60° / 69° Celsius with a fixed fan speed of 89%.
Are u running furmark in a chamber without air flow and the ambient temperature is over 35°?

 

havent use 3rd party since 195.? as there was probs with compatibility, will wait till driver devs calm down a bit and rt catches up

and i rather create my own fan profile than use fixed speed of jumbo jet

i run furmark on settings i know to heat up my card, thats why i use furmark in the 1st place, to heat it up, usually to test fan profiles, or fluidmark to test physx drv

back to power nappin ->

 

Let's hope new drivers from NVIDIA include temperature checking for WHQL certification from now on because this auto-fan bug should never have happened in the first place IMO.