NVIDIA shows signs ... [2008 - 2017]

[Bouncing Zabaglione Bros.]

Didn't the 590s need to be overclocked to blow up? Overclocking isn't exactly guaranteed to go fine. Although there is a new sense of overclocking entitlement brewing about in some enthusiast circles, partly thanks to how much the marketing departments advertise it these days.

Apparently some have been blowing at normal speeds/power. Even if you overclock, you'd expect some graceful failure, not blowing the power circuit at the slightest speed increase.

That whole business of the power limits being in the driver and working via application name doesn't look good compared to AMD's Powertune either.

At the end of the day, a lot of people pointed to the power circuitry as a problem at launch due to these blown cards, even to the extent of Nvidia issuing emergency drivers to limit voltage and speed. Now here we are a few months down the line with Nvidia beefing up those exact parts of the cards. It seems to me that Nvidia are all but admitting there's a problem that needs to be fixed, or else why upgrade those parts that are known to cause the problem?

 

[Kaotik]

Didn't the 590s need to be overclocked to blow up? Overclocking isn't exactly guaranteed to go fine. Although there is a new sense of overclocking entitlement brewing about in some enthusiast circles, partly thanks to how much the marketing departments advertise it these days.

IIRC at least 3 cards died at default voltage/clocks around the launch reviews

 

[swaaye]

Hmmmm. It is interesting that NV would mess up what is the second implementation of that chip.

 

[Malo]

They didn't mess up the chip, they messed up the supporting voltage hardware?

 

[Bouncing Zabaglione Bros.]

They didn't mess up the chip, they messed up the supporting voltage hardware?

Looks like underspecced MOFSETs that will blow fuses on the board when too much power goes through them. Circumvent the fuses, and the MOFSETs blow.

It looks to me like Nvidia originally specced it for a lower speed/voltage, then at the last minute the upped the speed to try and compete with the 6990, and some of the power components were not up to the job.

That of course did not stop Nvidia selling it to the public and trying to hack in some half-assed voltage protection via drivers/app detection.

 

[swaaye]

They didn't mess up the chip, they messed up the supporting voltage hardware?

Of course. Unless they are completely inept over there, which is quite doubtful, I'd say they just got out of hand with the cost cutting.

Also, what BZB said.

 

[Alexko]

Does that mean Nvidia will need to replace everyone's 590's, including paying for re-designs for custom boards made by vendors?

Judging from the kind of laptop replacement they offered to "bumpgate" customers, I guess they'll just send people a GeForce GT 430 and call it a day.

 

[swaaye]

I was thinking about that bumpgate Compaq again....

Like I said earlier on, a friend of mine has a Dell which had a 7900 GS die. These are included in the bumpgate "controversy". He sent it in and they gave him a new 7900 GS. So why can't everyone else send in their notebooks for repair?

 

[Kaotik]

http://jonpeddie.com/publications/add-in-board-report/

Updated add-in board numbers.
nVidia 59.12%, AMD 40.46%
nVidia lost bit over 1.5% units, while AMD gained just a tad more, eating the extra from the minor players
Year to year nVidia has lost over 5% units while AMD has gained nearly 6% units, again the extra from minor players

However, these numbers include CUDA-workhorses and such, too, not just "normal graphics cards"

 

[Psycho]

Judging from the kind of laptop replacement they offered to "bumpgate" customers, I guess they'll just send people a GeForce GT 430 and call it a day.

More likely a HD6450, considering the graphics in the bumpgate notebook
(that way it's not their problem if it blows up too).

However, these numbers include CUDA-workhorses and such, too, not just "normal graphics cards"

I guess Tesla cards are less than 0.1 percent of the market (units), but wonder how many units the Quadros are..

 

[thatdude90210]

More likely a HD6450, considering the graphics in the bumpgate notebook
(that way it's not their problem if it blows up too)...

Nvidia - an AMD valued costumer since 2011.

 

[kyniskos]

employee of nVidia leaking quarter-numbers

http://noir.bloomberg.com/apps/news?pid=20601087&sid=aXtWEf4YxyGU&pos=6

This is an interesting story. Doesn't affect nVidia in itself, but still newsworthy.

 

[Grall]

What an idiot... He has (or had, anyway) a great job, undoubtedly great salary, why piss all that away and sit for half a decade in jail. Such greed, such stupidity.

 

[neliz]

http://news.yahoo.com/s/digitaltren...softpactwithnvidiacouldresultinfuturetakeover

In a Securities and Exchange Commission filing dated May 27 and reviewed by Information Week, Nvidia says, "Under the agreement, if an individual or corporation makes an offer to purchase shares equal to or greater than 30% of the outstanding shares of our common stock, Microsoft may have first and last rights of refusal to purchase the stock."

 

[AlphaWolf]

http://news.yahoo.com/s/digitaltren...softpactwithnvidiacouldresultinfuturetakeover

10 years old and being discussed here.

 

[trinibwoy]

And if I can, just a quick follow-up. 28 nanometer, how does that look relative to 40-nanometer? That was a tough industry transition. Can you give comment there?

28, we are far, far better prepared for 28 than we were for 40. Because we took it so much more seriously. We were successful on so many different nodes for so long that we all collectively, as an industry, forgot how hard it is. And so one of the things that we did this time around was to set up an entire organization that is dedicated to advanced nodes. And we've had many, many tests chips run on 28-nanometer. We have working silicon and, momentarily, about to go to production with 28-nanometer. And it's looking really good, it's looking much, much better than our experience with 40-nanometer. It's just a comprehensive, across-the-board engagement between TSMC and ourselves, and making sure that we're ready for production ramp when the time comes. So I feel really good about 28.

Lip service or do you guys think they really changed their approach on process adoption?

 

[ninelven]

Oh I don't doubt they changed their approach, but what it actually got them is another story...

 

[rpg.314]

Lip service or do you guys think they really changed their approach on process adoption?

If they blame TSMC for leakage all over again, assuming they are late by 4-6 months, then I'd guess lip service, but not otherwise.

Besides, if they haven't changed their process migration recipes, then they deserve to suffer.

 

[trinibwoy]

Q3 results are in. http://seekingalpha.com/article/307...sses-q3-2012-results-earnings-call-transcript

Revenue for the third quarter was $1.07 billion, up 4.9% sequentially. Non-GAAP gross margin of 52.5% was a record for the fifth consecutive quarter Gross margin exceeded our expectations for the quarter, primarily as a result of strong demand and the high-end segments of our desktop GPU products and record sales in our Quadro professional graphics products. Non-GAAP OpEx was $317.6 million, slightly below our estimate. Combined, these results generated a non-GAAP net income of $217 million, or $0.35 per diluted share, up 12.1% over the prior quarter..

 

[NathansFortune]

Good job by Nvidia. Where's Charlie now? He can go take a hike.