NVIDIA shows signs ... [2008 - 2017]

Not searching, it popped up at another forum and users began reporting them in by themselves. A natural Emergence I would say.
It's something that occurred in the last three months.

 

No I mean that people would have noticed a pattern of failures at the big sites like [H] or XS. You seem to be searching for one, accumulating anecdotal reports from all sorts of obcsure places.

 

Same place that got the RV770 scores first :/

 

So it's not just dutch, but chinese 8800's that are failing too?

 

Seems to go the way of the 9800 (radeon)

 

Try shipping 4870X2 cards in situ in a system via a courier - we've taken to removing ours after experiencing cards wrenching PCIe slots off board to boards inexplicably just not working with no obvious signs of damage.

 

Shipping equipment completely built is the last thing a professional should do. There's a reason components come in boxes.

 

Right, reiterating what Rys said above, dare I suggest we abandon the 1 liner exchange that has little to do with the thread topic (yes, even if G80s are too strained and thus fail in China or Holland, and R700s strain PCBs in horrible ways that make them non-functional). The alternative is to have this locked, and that would be a bit of a waste. Thank you!

 

Don't thank me, thank the efforts of our Dev Rel team.

 

Even though tweakers is not an english language site it can't honestly be called obscure. What it lacks in editorial content it makes up by a very large and vocal community.

That makes it a fair sample size for spotting trends. Which means that the handful links with reports of incidents with G80s hardly qualifies as an 'alarming rate'. There are literaly thousands of people on that site with such cards.

 

Well, that's what it looks like to me most of the time.

Oh sure they've got performance increase in HAWX. Were they faster than NVIDIA? Nope. So what they've done has nothing to do with improving gaming graphics. They've just used their fraction of API to improve their performance but were still slower than NV thus bringing no possibility for improving graphic quality.

That's the question you need to ask AMD because it was their decision to effectively ruin common DX10 API in favor of their DX10.1 update.

Condemn DX10.1 then. Condemn proprietary 6x0/7x0 tesselator (ended up being not used at all?).
What we're discussing here is NVs provided MSAA support in BAA. What's that has to do with APIs and PC platform fragmentation? NV can use whatever it creates for profit. NV has no reason to support anyone but itself. And AMD has exactly the same goals and rights. Or should we comdemn them for porting Havok to their hardware and not optimising it for NV GPUs? (Where is that OpenCL Havok port, btw? Ended up being AMD's marketing speeches again while PhysX got it's second AAA title released?)

"Noticable implementation" means quality improvements or at least visible performance advantage. HAWX has neither.

Ubisoft removed 10.1 support from AC because it was buggy. And from what i've heard it's inclusion in AC v1.0 was a mistake.
But even if it's true -- where was AMD when that happened? Doing marketing speeches again instead of pushing Ubisoft for proper 10.1 support in AC? See now what i'm talking about?

Mirror's Edge uses it's own renderer with it's own MSAA support implementation.

 

/Davros hands GEO a brick wall "here bang your head on this"

 

So, any news on what went down at the NVIDIA spoiler press event today?

 

Well there would have had to be an event in the first place before something could go down

 

But there was. Or will be. It was all over the place. Some people swore that it'd happen, and swore at people who said it wouldn't.

So there must be some news.

 

There was "one" editor invited too Nvidia this week. And it was not a press event. It was a one on one interview type thing.