3D, Business, Graphics, Hardware, Software Programs

Nvidia Big Bang drivers caused a spike in RMA departments?

Long awaited Rel180 drivers, also known as Big Bang II drivers brought forward many good things, such as multi-monitor capability when SLI is switched on (finally! it only took three years to implement the feature)… and these drivers were main argument in recent spike of “GeForce GTX260-216 vs. Radeon 4870 1GB” reviews that spawned on websites like mushrooms after the rain.

But, in the past 48 hours I spoke with three different Nvidia AIBs… their names will remain unknown, of course – I was told that since Rel180 came out, they all have a spike in tech support. Issues range from temp and fan speeds not being reported correctly, to game crashes. After doing internal testing, all three vendors started to advise users to roll back to a previous driver version – and miracle happened. Systems became stable again.

So, what is going on? I sent out couple of e-mails out, and judging by first response I got (off-the-record), it seems that Big Bang II drivers introduced rewrite at the way how temperatures are addressed. This caused all sorts of malfunctions in thermal monitoring software, and at the moment, there is no easy way to correctly find out what is the temperature of the graphics card – unless you can touch it, that is.

I checked the only two GTX280 cards I have, PALIT GeForce GTX280 1GB and EVGA’s GeForce GTX280 Superclocked and saw small, but noticable difference in temperatures between the two driver revisions (178.24 and 180.48)… 2-3 degrees, but still noticable. 

The “crash” we experienced is the fact that the fan will not spin up on Superclocked card, and this caused some errors with Crysis Warhead. On 178.24, there are no issues.  With Big Bang II drivers, the card is producing a lot of errors in Crysis Warhead and does not want to launch CoD5 at all. 

Al in all, weird situation.

  • Theo, the difference in temperature you noticed with your cards is with the cards under stress, I suppose. How do you stress test the cards, only launching the games? Pr do you use a specific software? Can you do me a favour? Could you launch GPU Caps Viewer – with the Open GL Stability Test – Fur Rendering? Please set the window size to cover all your monitor (1920 * 1280) for example, and please do not cover the monitor with any other window. Let it run for 30 minutes, and check the temps and the fan behaviour.

    Please let me know the highest stable temp you achieve (you can watch it directly on the GPU Caps Monitor’s main window).