NVIDIA shows signs ... [2008 - 2017]

Status
Not open for further replies.
I think he deserves credit for his exposure of the Bump problem(Even if it was possibly blown out of proportion... I would not be surprised if many of the higher end cards start mysteriously failing around 2-3 years from now in droves).

Charlie comes across sometimes, most of the time with NVIDIA like he is gloating and blowing things out of proportion but I don't know if bumpgate was(blown out of proportion).

I wanted a 3g wan card for my xps m1530 a month or so ago and looked on the bay, a guy on there was selling all the bits off from a defective (yes graphics chip) xps m1530 of the same. What bothered me was he was just out of his 12 months warranty and Dell were supposed to be extending everyones warranty by 12 months, he had called them and been provided a ridiculous quote that no one would pay to fix any laptop.

People are being screwed to save money for NVIDIA or for that matter any large corporate that puts out a defective part, just look at SONY only just now fessing up to defective parts on their range. I did contact him but several of the parts had reached a point where they could not be taken off auction, just one example of either NVIDIA, DELL or other OEM shafting someone to save NVIDIA some money, I'm just waiting for mine to go bang now, thanks NVIDIA,

take care,
J.
 
Well see if nV manages to get some GT21X shipped in volume when back to school starts. Btw, when's that in US? 1st Oct?

Although even GT218 will technically invalidate Charlie's claims, I wouldn't count that pathetic part (of course, you may say it has a good performance/die-size, i don't know). I'm waiting for GT215/216 instead.
 
post 1252 said:
The demo shows 4096 bot planes handled solely on a single GPU parallel to game rendering.
Late further reply to this since I was just re-running the Froblins demo.
Froblins has 16,384 AI Froblins running their AI on the shaders as well as running tessellation.
Froblins AI is 2D only though so if 3D AI is x^2 computationally difficult it works out as a similar computational effort.
 
Last edited by a moderator:
Late further reply to this since I was just re-running the Froblins demo.
Froblins has 16,384 AI Froblins running their AI on the shaders as well as running tessellation.
Froblins AI is 2D only though so if 3D AI is x^2 computationally difficult it works out as a similar computational effort.

Don't you also need to know how much work each AI routine is doing? I don't think you can simply assume that both are equivalent.
 
You probably haven't been running for G80 for 2,5 years now?

This is the closest thing I could get to "evidence"

http://hardforum.com/showpost.php?p=1034145333&postcount=1
Put your card in your sisters easy bake.

again, there's the same amount of stories for "it had artifacts, I baked it, now it's dead"

came across another one of the success stories on ocn-
http://www.overclock.net/7268505-post33.html

and here too-
http://www.ocforums.com/showthread.php?p=6171248

the g80s seem to work if done right other cards don't.
 
I have an X1400 that's born-again in my wife's laptop that would argue with that "other cards don't" bit. :p

/me freaking loves the oven trick!
:D
 
I have an X1400 that's born-again in my wife's laptop that would argue with that "other cards don't" bit. :p

/me freaking loves the oven trick!
:D

exception to prove the rule:p
maybe i was too hasty in reading up what worked and what didn't but g80 do seem to be more receptive to the oven treatment.
 
Posted here due to premature closing of the other thread, sigh.
So RV740 avialability 6 months after it was announced means that AMD didn't suffer as much, eh? And it's price parity with 4850 surely mean the same thing?
6 months after its announcement is past its end of life date. They stopped making them ages ago in preference for D3D11. I won't deny that if TSMC had done a better job RV740 would have been around earlier and later (well, maybe not much later if the rest of Evergreen came earlier).

Obviously I have no idea if RV740 turned out to be a loss in terms of its development costs versus the profit it earned (if any).

Apparently there's a laptop model or 2 out there with RV740 inside. No idea what the sales numbers are. Dunno if they are a significant factor in AMD's major growth in the laptop market.

Meanwhile, NVidia's losing OEM/ODM contracts and has GT21x cards that are currently only available in OEM systems. Clearly those cards aren't ready for retail.

It's not "safe", it's easier, less risky and it's done so that later you won't make the same mistakes with a bigger chip.
That's why I put quote marks around the word safe.

I don't have enough info for any conclusions right now.
And it puzzles me when I see someone who apparently does.
It puzzles me even more to read about G300 delays while the initally planned launch frame haven't even passed yet.
All we have right now is a delay of GT21x series for which TSMC is the one to blame. That's all. How it'll end up with GT21x power/price/performance and GF100 we'll eventually see.
TSMC is to blame, but saying NVidia is blameless is remarkable. NVidia might have re-strategised GT21x introduction based on the 40nm fuck-up - in the same way some would argue AMD should have waited 2 months to introduce RV740. I'd agree there's possible justifcation in Nvidia doing this if the costs were stratospheric. But bearing in mind how much NVidia "bought" marketshare during the prior quarters, I'm not sure if avoiding those costs (and as a result losing contracts that GT21x would have fulfilled?) was the right thing to do.

GT21x failure, in general, could be a big part of the reason why NVidia's been losing OEM/ODM contracts.

GT218 has worse power and performance than a 55nm chip that's a year old, while AMD's 40nm chips (that have been up and running at AMD since the first quarter) are notably better in both power and performance than their 55nm predecessors. Its lack of competitiveness might be just another reason why NVidia's losing contracts (though that battle might have been RV710 v G96/98 and GT218 was simply too late).

Jawed
 
GT21x failure, in general, could be a big part of the reason why NVidia's been losing OEM/ODM contracts.
NVIDIA lost OEM/ODM contracts because of R700 line being more interesting to OEM/ODM than G9x line. GT21x chips are getting these contracts back if anything. They'll probably loose them again in the next cycle against EG series but from that pov GT21x is not a failure.
 
Yes AMD was able to push down costs to the level nV wasn't able to compete with thier older chips, for OEM's its all about the money since for low end computers its about volume.
Which the GT21x seem to be lacking, it is like this infinite loop that goes no where. GT21x winning contracts back, volumes and such.
 
Status
Not open for further replies.
Back
Top