Nvidia 196.75 drivers causing graphics cards to die
geek.com
geek.com
However I haven't played any of those games, or really done anything terribly intensive aside from Zbrush and Maya, so I'm uncertain if it's related or merely a coincidence. Although, to be fair, I've never had a fan of any description die.
Does anyone know of some way to check or ultimately claim for this?
It's because it's the exact same card. On most rebrands, at least they bother to change the relevant identifiers, but 9800GTX+ differs from a GTS250 only by the number on the box. As such, the drivers have no way of knowing which of them you have -- they cannot even make a solid guess from batch numbers because unsold 9800GTX+ were repackaged as GTS250's. Faced with this, the driver programmers decided that less people would probably complain if they just called all the cards by the latter name.
Fun fact -- differently harvested versions of the G92 chip have been sold as:
- GeForce 8800 GT $2
- GeForce 8800 GS $1
- GeForce 8800 GTS
- GeForce 9600 GSO $1
- GeForce 9600 GSO ASUS 512 $$
- GeForce 9800 GT $2
- GeForce 9800 GTX
- GeForce 9800 GTX+ $3
- GeForce 9800 GX2
- GeForce GTS 150
- GeForce GTS 240
- GeForce GTS 250 $3
- GeForce GT 330
Groups marked with $n are all exactly the same card, with no differences whatsoever. Many other sets have only very minor differences. $$ denotes the biggest ripoff award -- halved memory bus width and cheap memory make this card only half the speed of the normal 9600 GSO.So, if you wonder why NVidia is swimming in cash, it's because G80, and it's slightly modified shrink G92, are insanely good designs. G80 debuted in late 2006 and G92 in late 2007, and it's still selling.
Just to update -- if you are buying a card now, ATi is quite probably the better choice, at least until NVidia gets GF100 and it's derivatives out.
Unless you're planning to run Linux, in which case the Nvidia drivers are still far and away better than any of the multiple offerings for ATi chips. Even though the Nvidia drivers are closed, they're extremely reliable and feature just about everything but Kernel Mode Setting, which is honestly not that big of a deal.
The nv driver used by default in most distros is quite sluggish, and is plagued by all manner of quality issues. The radeonhd driver in comparison is quite stable, implements working VT switching, and full EXA acceleration with a start on 3D. (EDIT: nouveau will provide a very good replacement for nv, but so far its deployment is limited)
As for the closed drivers, is stability really a problem? While I haven't had an NVIDIA card for a year or so, I seem to recall a fair number of kernel panics in graphically-intensive programs, while I have yet to experience one using fglrx (although the comparison is a big flawed, since my needs have changed quite a bit). Speed is also quite acceptable and does what I need.
I'm willing to grant that nvidia is better than fglrx. Honestly I don't do all that much 3D work on Linux. But in my experience, calling nvidia "far and away better" does ATI a bit of a disservice.
All that said, I wish ati developers the best of luck on making the open source drivers competitive with nvidia's binaries. I'd love to see them put some pressure on nv so it's not just a one horse race anymore...