Why Fermi Flopped From the CEO himself

I really think Nvidia is feeling the pressure of the world. They abandoned the chipset business which was stupid because the AMD community would welcome a SLi chipset anyday. Then they banked on 32nm being ready but TSMC is a private business and doesn't answer to Nvidia alone. They just had had a few bad streaks and like a spoiled child they are lashing out. Nvidia makes good cards but their business/market side is shady as f*ck and that's what I don't like about them. Starting with their retarded ass President.
 
OFaceSIG said:
I really think Nvidia is feeling the pressure of the world. They abandoned the chipset business which was stupid because the AMD community would welcome a SLi chipset anyday.
Click to expand...

Nvidia didn't abandon the chipset market they were pushed out. There is also no reason that AMD boards cannot run SLI other than Nvidia not allowing it.

Nvidia has been too arrogant and vindictive for their own good. Now AMD & Intel are both working to push them out of markets and put a hurt on them in general.
 
Terrible thread title. The article is about why Fermi was so late, not why it "flopped".

Fermi didn't flop.. gf104 is doing pretty damn well.
 
You guys do know that all that holds back sli from working on any xfire and similar board is just the driver, right?...

So you can find a softmod that will deal with that pesky issue IF your board is already compliant with multi PCIe vidcards...

http://forums.extremeoverclocking.com/showthread.php?t=330976



Back on topic: Interesting reading, i gotta wonder what would they be able to pull off if they learned basic communication skills between the teams...
 
BlueStorm said:
Terrible thread title. The article is about why Fermi was so late, not why it "flopped".

Fermi didn't flop.. gf104 is doing pretty damn well.
Click to expand...

I'm so sick of this, we're not talking about benchmark success were talking about MARKET success. And if you, for a minute, try and argue that nvidia has anywhere near equal DX11 market share, you are very mistaken. Even a steam survey showed that 86% or something like that, had ATi DX11 cards.

If we used your logic, then the USA won the vietnam war because we killed a hell of a lot more of them than they did us. (2million vs. 56,000)

I'm so sick of these nvidia fanboys saying that fermi didn't flop. It was a massive blunder, and the sales show it. It was a huge failure in the market.

How does it feel to be completely controlled by a marketing team, that you fill fight tooth and nail for a graphics card company?
 
Last edited:
lockheed2266 said:
How does it feel to be completely controlled by a marketing team, that you fill fight tooth and nail for a graphics card company?
Click to expand...


I doubt it's marketing team that fuels this. It's more like I brought it so it is better in almost every way mindset, e.g. console wars.
 
mutantmagnet said:
Original link in German.

Translation.
Click to expand...

First thing OP is that the CEO never said Fermi flopped, he explained why it arrived 6 months late to the market.

Fermi did not flop in fact in the 6 month it's been out it has made excellent ground catching up to amd who has been out for a year with it's dx 11 gpu's.

No nvidia doesn't have as much market as AMD in DX 11 AMD not only had a 6 month head start (high end)they had a 11 month head start on the mainstream market (5670//5750/5770) these are the cards that most people buy.

Nvidia has just recently catered to the mainstream with the GTX 460 on the high end of the mainstream and (Fail) with GTs450 due to it's performance and price point at launch, although it can be found for $109.99 after MIR which makes it a better buy than when it launched less than 1 month ago.

Anyway point is, Fermi did not flop. AMD just played a better hand at poker this time around. Nvidia wins some AMD wins some, it changes from time to time.
 
Last edited:
lockheed2266 said:
I'm so sick of this, we're not talking about benchmark success were talking about MARKET success. And if you, for a minute, try and argue that nvidia has anywhere near equal DX11 market share, you are very mistaken. Even a steam survey showed that 86% or something like that, had ATi DX11 cards.

If we used your logic, then the USA won the vietnam war because we killed a hell of a lot more of them than they did us. (2million vs. 56,000)

I'm so sick of these nvidia fanboys saying that fermi didn't flop. It was a massive blunder, and the sales show it. It was a huge failure in the market.

How does it feel to be completely controlled by a marketing team, that you fill fight tooth and nail for a graphics card company?
Click to expand...

Failed assumption on your part. I *am* talking about sales and share, but qualified by the amount of time the part has been out... monthly sales. You'd have to be a retard to believe you can compare share for a part out for a month vs a part out for 10 months to a year. But according to you, all of Fermi flopped because 2 high end gpus from the entire line up drew a bunch of power and the line up was late...okay then.

Fermi is an architecture, and it is ground up new and pretty bad ass.. but the product SKUs which came from it for the very high end didn't resonate very well. Oh well... the rest of the lineup is doing fine monthly.

Separate the arch from the products in your mind, they're two different things. I can just as easily accuse you of being an NVIDIA hater.. lot of good flaming does here.

I never said NVIDIA had more share. But good luck with your whole straw man fallacy tactics in the future.
 
Yeah i wouldn't say either that Fermi flopped... not even close when looking from a broad view.

The thing is that the gaming market, like it or not, is actually a small and low profit market vs the entire world of professional markets.
And in those it is undeniable that Fermi was a resounding success.

When looking at the whole market, i am sorry but there is no fanboyism, Fermi was a economical success, just not the one some of you guys wanted.


edit:Sorry for the horrible engrish, will try to fix it.
 
Last edited:
Revdarian said:
Yeah i wouldn't say either that Fermi flopped... not even close when looking from a taking a broad view.

You see the thing is that gaming market, like it or not, is actually a small and low profit market vs the entire world of professional markets. And in those it is undeniable that Fermi was a resounding success.

When looking at the whole market, i am sorry but there is no fanboyism, Fermi was a success, just not the one some of you guys wanted.
Click to expand...

I love it when the nvidia fanboys like you calling other fanboy and keep bragging about the so called professional market, it is like saying your gas hog car isn't shitty because the same auto manufacturer profits in building tank! that car must be awesome, that is beyond stupid.

Also when you talk about the "pro" market, Fermi-based Quadro was nowhere to be seen until August, whether it is a success or failure is remained to be seen, however it would be fun to watch if shitload of "professional usecomputer" on fire. Oh the gaming fermi is definitely a failure.
 
http://finance.yahoo.com/q/ao?s=AMD+Analyst+Opinion
http://finance.yahoo.com/q/ao?s=NVDA+Analyst+Opinion

I bet that market analysts are fanboys too, that is why noone recommends to sell nvidia stock vs selling amd stock, and instead those same ~3 recommendations are added to "strong buy" their stock.

Also
http://finance.yahoo.com/q/ks?s=AMD+Key+Statistics
http://finance.yahoo.com/q/ks?s=NVDA+Key+Statistics

Look at the balance.


Professional includes HPC. There is way more in fermi so i am sorry, but if i am a fanboy, all Wall Street is a fanboy too.

edit to add:
/shrugs, go figures, people here really do believes that the gaming market is what keeps the technology flowing atm, and that fermi exists in a single variant with different number of cores activated.
 
Last edited:
Won't read the article nor listen to Huang but I bet he laid the blame on TSMC and middle-management. Dear leader has been taking notes from Kim Jong-il.

Revdarian said:
Fermi was a economical success, just not the one some of you guys wanted.
Click to expand...
Bwahaha thanks for the laughs. Nvidia just recently took a $12 million write off because they are not confident they can sell product inventories they have in stock.
 
Lorien said:
Bwahaha thanks for the laughs. Nvidia just recently took a $12 million write off because they are not confident they can sell product inventories they have in stock.
Click to expand...

The loss has been larger then that:

Santa Clara (CA) - The replacement of faulty graphics chips as well as a much stronger than expected AMD/ATI pushed Nvidia deep into the red during the second quarter. The company reported a revenue decline of 23% sequentially and a triple-digit loss.
Click to expand...
http://www.tomshardware.com/news/Nvidia-Q22008,6124.html

So what? Nvidia has much money and can take it. No need for doom and gloom. Whats more interesting, is from the same link:
Huang also admitted that Nvidia underestimated the "price/performance" of AMD’s/ATI’s most recent graphics chip, resulting in a wrong position of its products and forcing the company to adjust its pricing during the quarter.
Click to expand...

Meaning cheaper cards. Last part belongs more in the Video cards forum, while the first part is more a "wall street forum" or "marketeers one-on-one" kind of thing for those that is concerned about PR of the companies that makes the cards.

Come on, how many of you guys are really interested in stocks and a products market penetration? Can't see where op is going with this thread in a video card forum.
 
Lorien said:
. Dear leader has been taking notes from Kim Jong-il.
.
Click to expand...
They're actually the same guy! Look at pics of them. All that news about Kim Jong-il being sick and out of public view when all the while he is actually in Santa Clara plotting to take over AMD.
 
There really isn't a point in arguing about semantics when simply rephrasing would probably end the argument.

As a product line to make the company money Fermi is an abysmal failure esp compared to the 5800 series (or the 4800 series for that matter). and the GT200 didn't do all that well either. This includes every bodies favorite 200 dollar card, the 460GTX. its an awesome product for us, as a money maker it sucks. it is pretty successful in regaining some market share.

I don't think anyone will disagree with the above, of course how much and why are always debatable but the term flop was pretty loose here

as far as the article is concerned I was impressed that he did allude (was hard for me to tell, I will have to reread that latter.) to some management issues. that is certainly a set in the right direction.
 
Valset said:
There really isn't a point in arguing about semantics when simply rephrasing would probably end the argument.

As a product line to make the company money Fermi is an abysmal failure esp compared to the 5800 series (or the 4800 series for that matter). and the GT200 didn't do all that well either. This includes every bodies favorite 200 dollar card, the 460GTX. its an awesome product for us, as a money maker it sucks. it is pretty successful in regaining some market share.

I don't think anyone will disagree with the above, of course how much and why are always debatable but the term flop was pretty loose here

as far as the article is concerned I was impressed that he did allude (was hard for me to tell, I will have to reread that latter.) to some management issues. that is certainly a set in the right direction.
Click to expand...



queue the response ...but but but 5 series in comparison had a 6 month lead...

Like it will ever catch up now since this comparison will be mute mid October.


Few people here don't seem to like the title but what is a CEO going to say about Fermi? He's going to give it the best possible spin/damage control as to not cause damage with shareholders. Learn to read between the lines.

Looks like a simple communication between the 2 engineer teams could have prevented the failed execution of Fermi.
 
Charlie takes his TSMC defence too far.

Anyone with any note of history remembers the great 4770 drought the 58x sold outs and price increases. Yield tore into AMD just as harshly but because they got a pipecleaner in the oven so quickly they found the issue while they still had time to make changes to the 5x line. It wasnt enough to save them from horrific yields as evidenced. But it kept them from putting out an AMD version of the GF100.

Is Nvidia FRAKKING stupid for not putting Pipe cleaners in sooner? YES but saying TSMC 40nm is issueless is absurd.
 
Dark Shroud said:
Nvidia didn't abandon the chipset market they were pushed out. There is also no reason that AMD boards cannot run SLI other than Nvidia not allowing it.

Nvidia has been too arrogant and vindictive for their own good. Now AMD & Intel are both working to push them out of markets and put a hurt on them in general.
Click to expand...

Nvidia was not "pushed out" They completely abandoned VIA when a VIA/Nvidia ION chipset could have ruled the HTPC root.

Also they had plenty of time to get a final Kickass Chipset out for AMD CPUs They have chosen not to and thus are at fault themselves.
 
LOL, the interview is actually pretty cool. Don't be afraid of the german, the video is him speaking in english.

He basically partially blames it on TSMC as they "parasitically (mis)characterized" the reality of what they were able to produce in terms of the interconnecting pipelines between the SMs and Memory and it was completely broken and they had to scrap it and re-engineer it to work...but there was also a problem between two departments that miscommunicated what was neccessary to get this part to work.
 
Thing is that if you see their moves, they prefered to broaden their market instead of spending R&D on a terminally ill one, with a public death sentence for 2011 with sandy bridge and fusion. (Actually we can argue that the intel one was as good as dead the second they showed the I5´s+ with gpu integration, sure nvidia's gpu's would be leaps and bounds better but at that low end market they weren't gonna be needed and would be better off getting in bed with ARM full time for Tegra, and if Tegra2 found competition they just need to go all out on Tegra3)
 
Tegra 2 failed Tegra 3 on Gloflo wont come out until end 2011 at the earliest.

Dont see too much profit in that compared to nforce.
 
Yup, DOA due to the competition, and yet they got someone like LG to sign with them, and the graphics are pretty much the best you can get on integrated phone soc's atm.
Tegra3 on Glofo won't have any problem tho, the method that is gonna be used is very matured totally unlike 40nm on TSMC (and i wonder what did TSMC told Nvidia, i mean if what Jen Hsung says is false he could be getting sued due to slander... but TSMC is just keeping quiet...)

nforce again was on a dying market, sure they pulled the plug rather abruptly, but fusion and company were already a "sure" thing so they prefered to invest for the future, would i have pulled the plug so abruptly?... hmm maybe not, but i don't have the projections of what would have been the cost of 1 extra license for a short known time + R&D... i don't count VIA because that would add the variable of what were the terms of the proposed deal, if there was any on the table.
 
Not best you can get considering the extra voltage needed to drive the chips. And I bet Marvel's latest triple core can lay it to waste in graphics performance.

TSMC isn't going to sue anyway and ruffle feathers. They screwed up and they know it. AND they have to compete with Globalfoundries now (Guess I will have to say it fully now because people cant take a fun bit of netspeak)

Tegra 3 has the potential to be somthing really serious in the mobile space. But MAN does it have competition. The 28nm processes the ARM cores will use are attracting major players AND will be the first to fab the Eagle core. Tegra 3 will have to offer an astounding experience if it wants to compete.

#1 Complete 3D BluRay decode and playback at 1080P
#2 FAR better 3D ability
#3 Getting serious about CUDA and the mobile space.
#4 In my opinion it needs to have the option of X86 emulation assist. Some cores out there can get to within 70 percent of X86. Surely something can be done to aid emulation that approaches that without going full X86.
 
Zachstar said:
Tegra 2 failed Tegra 3 on Gloflo wont come out until end 2011 at the earliest.

Dont see too much profit in that compared to nforce.
Click to expand...

If Tegra 2 is a failure why would LG and HTC be working on making smartphones and tablets with that?
 
Last edited:
yimingwuzere said:
If Tegra 2 is a failure why would LG and HTC be working on making smartphones and tablets with that?
Click to expand...

Look at why Boxee box went from it to Intel. Not even decent 1080P performance.

Others have noted how it does not meet specs.
 
Zachstar said:
Look at why Boxee box went from it to Intel. Not even decent 1080P performance.

Others have noted how it does not meet specs.
Click to expand...
I'll wait and see HTC's tablet first. If it still fails at 1080p H.264 then I'd most likely agree with you.
 
Zachstar said:
Look at why Boxee box went from it to Intel. Not even decent 1080P performance.

Others have noted how it does not meet specs.
Click to expand...

I know it sucks to have facts interfere with your whole agenda, but get your facts straight.

The chip is meant for embedded applications, and for LOW, LOW power draw. As such, putting it into a device with a power plug where more power draw is actually okay as long as you get more performance (see Atom) was probably a bad choice to begin with. Tegra is TEGRA.. inTEGRAted. It's meant as a mobile chip.. i.e. low power draw, but good functionality for a mobile chip in phones or tablets (i.e. for it's class of device).

The chip does 1080p where other mobile chips do not. The *PROBLEM* was not that it didn't do 1080p but that it didnt do the high profile spec for 1080p, which D-Link wanted Boxee to support. Well if you look at the product they're making and say "hey, this chip is made for low power draw, but we don't really care about low power draw... but we have processing capability limitation with this chip and *DO* want more processing ability", well naturally you'd move away from mobile class chips altogether.

The chip is still pretty good for tablets and phones, so we'll see how that plays out.
 
Ya go ahead and see how it plays out. Considering how long they have been delayed oh and 1080P AVC decode was touted by Nvidia a long time ago. And I doubt Boxee would just "Imagine" them having support for it. They were told it could handle what they needed and that was wrong.

As for low power draw. I laugh at that considering the reports of having to raise voltage significantly to get these to work.
 
Zachstar said:
And I bet Marvel's latest triple core can lay it to waste in graphics performance.
Click to expand...

I wouldn't be so sure, had to look them up since the announcement was only made the 23, 200mill triangles/sec sounds impressive but without other numbers to back it up, and they don't have past experience with doing 3d graphics (3d blue ray they do, but that is just streaming two blueray outputs at once ;-) )

I find interesting their concept for triple core, two high speed cores (1.5GHz) and one powersaver core (624mhz or so) makes sense with decoding dedicated hardware, which actually allows you to run the bare minimum while streaming video.
 
Snapdragon was supposed to be somewhat near Tegra 2 in 3D performance.

A certain phone has 2x speed of Snapdragon today and handles 90 million Triangles.

EDIT: Was wrong Tegra 2 is indeed 90 million triangles tho the phone obviously does it with less power...

The Marvell system is 200 million.

Edit: Wow I am failing tonight it was not 300 but 200

Yet another edit: http://en.wikipedia.org/wiki/Xenos_(graphics_chip) X360 = 500 Million... A bit less than half X360 PS3s is supposedly quite a bit less. I want videos!

Now I have no idea if they are really that close. But I remember hearing outlandish performance claims. I was repeating them when I sung the praises saying it would beat a Nintendo Wii. Then the games we see are nowhere near is detailed and some seemed to run just the same on a darn Ipad.

I would call the Marvel design begging for a portable game system
 
Last edited:
From their very own press release, it is 200mill.
http://www.marvell.com/company/news/press_detail.html?releaseID=1440

And no, Snapdragon is not near Tegra2 in 3D ;-) (you would have seen tests like the Unreal 3 ones otherwise)... in any case time to wait for LG's mobiles and Marvell's actual physical offerings to see them side by side.

Just to reiterate tho, i do trust Marvell's blue ray claims hehe.


Damn, we are living in awesome times all in all, aren't we?
 
I corrected myself on both counts before you posted.

Awesome perhaps? But 2010 has been marred by delays left and right. TSMC antics causing widespread high prices and hardware like the Ion 2 to be crap.

2011 please be better!
 
IMO the launch of the Fermi range flopped big time, and the biggest reason for that was it's delay, if it had been side by side with the ATI counterparts or even before then we'd have seen a much warmer welcome.

They continue to improve and pull it back a bit, the 460 seems a bit more popular, but the day 1 release stuff was pretty terrible, too little too late.
 
The 460 looks so good because Nvidia is kissing its profits goodbye with it in order to get something out. That makes the idiots at TSMC richer but does little for them.
 
You must log in or register to reply here.
Back
Top