Charlie explains the whole Fermi debacle in detail

[Daedalus685]

Both the 4xxx parts and the GTX2xx parts had margins in the 30%-40% range, in most industries that would be dancing in the streets stellar. They were not producing loss leaders and they could have reasonably priced them quite a bit lower still without them falling into that category. That said, eveyone in this segment shoots for Intel's 60%+ margins as an ideal, so it certainly wasn't like it was an insanely profitable part either(it was directly comparable to nV's margins, if they sold as many units as nV had they would have been more profitable as ATi has lower operating margins then nV).

I would like to read up on the 2xx and 48xx parts having a 40% margin... Maybe Newegg marks them up 40%....

as always.... [citation needed]

 

[cbn]

They have no choice.
NV has no license for x86 and last year they lost their entire chipset business (accounted for one-third of their revenue!), handheld/mobile chip sales make nothing compared to desktop especially WS sales and the discrete market is rapidly going away (into the CPU a' la Llano and Intel HD) sans niche, really high-end gaming cards... they have to create/carve out their new market, be it as small as it is right now.

Didn't Nvidia buy VIA? Doesn't VIA have a license for x86.

In any event I would like to see a large die Fusion chip from the Green Team. (Then we as consumers put our tower coolers on it and solve any noise problems associated with bleeding edge chip designs)

 

[BenSkywalker]

as always.... [citation needed]

http://www.sec.gov/

Have fun.

 

[Daedalus685]

http://www.sec.gov/

Have fun.

/facepalm

Edit (to remove comments that may or may not have included the term ass and may or may not have been possible to take as a personal attack)

I asked for your source because what you said goes against the common thought, thus it would be interesting to see if it is correct... There is no need to be a tool, if you don't have a source then don't make claims.

Also.. not being American I have no idea what the SEC does.. but are they not solely for investments and so on?

Almost all rumour and news sites claimed that the parts were manufactured and sold at a very low margin, maybe even a loss... There are no laws to prevent this sort of thing, only to prevent leverage of massive funds to put another out of business.

You claimed the margin was 40%.. prove it, or shut your mouth.

 

[BenSkywalker]

Also.. not being American I have no idea what the SEC does.. but are they not solely for investments and so on?

You may not want to call me out about something you have absolutely no clue about. The SEC is the body that all publicly traded firms are required to report their full financials to by law. It gives you breakdowns on all the various elements you are looking for. Every quarter the results are discussed in one thread or another, I dig through the filings ever quarter which is a process then can take several hours and post the data contained within. Making an inaccurate claim to the SEC means jail time, a little bit better then the sources some others like to use in terms of accuracy to put it mildly.

You claimed the margin was 40%.. prove it, or shut your mouth.

I pointed out the fact that it was in the 30%-40% range, decidedly different then 40%(the exact figure fluctuated every quarter).

 

[Lonyo]

/facepalm

Edit (to remove comments that may or may not have included the term ass and may or may not have been possible to take as a personal attack)

I asked for your source because what you said goes against the common thought, thus it would be interesting to see if it is correct... There is no need to be a tool, if you don't have a source then don't make claims.

Also.. not being American I have no idea what the SEC does.. but are they not solely for investments and so on?

Almost all rumour and news sites claimed that the parts were manufactured and sold at a very low margin, maybe even a loss... There are no laws to prevent this sort of thing, only to prevent leverage of massive funds to put another out of business.

You claimed the margin was 40%.. prove it, or shut your mouth.

NVs margin is in their financial report, which was in the 40% region. The financial reports have to be filed with the SEC and if they are fraudulent they can get in trouble.
BenSkywalker just missed the whole part where the 40% number is information included in their financial reports and went directly to the SEC.

But that's a 40% company margin, IIRC, so across all products. That's not a 40% margin for the 2xx and 4xxx cards specifically, so it's a big of a false figure.

 

[Dark4ng3l]

You may not want to call me out about something you have absolutely no clue about. The SEC is the body that all publicly traded firms are required to report their full financials to by law. It gives you breakdowns on all the various elements you are looking for. Every quarter the results are discussed in one thread or another, I dig through the filings ever quarter which is a process then can take several hours and post the data contained within. Making an inaccurate claim to the SEC means jail time, a little bit better then the sources some others like to use in terms of accuracy to put it mildly.



I pointed out the fact that it was in the 30%-40% range, decidedly different then 40%(the exact figure fluctuated every quarter).

Look the SEC has nothing to do with this thread. If you want to argue with him you can always PM him or something but this is way off topic.

 

[BenSkywalker]

But that's a 40% company margin, IIRC, so across all products.

If you check out the SEC filings they have segment breakdowns.

 

[Leyawiin]

Love him or hate him, but he has been 100% correct about fermi. Perfromance, yield issues, heat, noise was all laid out quite well.

Nobody should have been surprised when the reviews did come out. I wasn't, I believed his articles 100%.

Even a broken clock is right twice a day.

 

[tweakboy]

At the end of the day all we want is a card that can play Crysis at 60fps without a drop,,,ty and I dont think were getting that with Fermi Furbi 480. ATI dual gpu pownz Fermi

Soo nvidia better come out with their GTX500 dual gpu soon, I believe just guessin name and time,, thx,

 

[SunnyD]

Love him or hate him, but he has been 100% correct about fermi. Perfromance, yield issues, heat, noise was all laid out quite well.

Nobody should have been surprised when the reviews did come out. I wasn't, I believed his articles 100%.

Thing is anyone with half a brain knows that a chip that large running at the targeted clock speeds at expected voltages on the process technology available at the time would obviously have yield issues, take a lot of power, give off a lot of heat, and require substantial cooling leading to excessive noise.

That being said, performance would suffer as compromises would have to be made (less shaders anyone?) in order to push the yield up in order to make margins.

Charlie was just regurgitating what anyone in or around semiconductor technology knew from the get-go. It's also why ATI is laughing it's ass off, and has been for 2 generations now.

 

[Daedalus685]

You may not want to call me out about something you have absolutely no clue about. The SEC is the body that all publicly traded firms are required to report their full financials to by law. It gives you breakdowns on all the various elements you are looking for. Every quarter the results are discussed in one thread or another, I dig through the filings ever quarter which is a process then can take several hours and post the data contained within. Making an inaccurate claim to the SEC means jail time, a little bit better then the sources some others like to use in terms of accuracy to put it mildly.



I pointed out the fact that it was in the 30%-40% range, decidedly different then 40%(the exact figure fluctuated every quarter).

I did not call you out on anything... I politely (ok.. maybe it wasn't that polite.. but I was not trying to be disrespectful, I did actually want to read the source.. which I have not yet done) asked you for a source so I can read and actually learn something, and you linked me what may as well have been google.com .... Sigh.. never mind.

 

[BenSkywalker]

Charlie was just regurgitating what anyone in or around semiconductor technology knew from the get-go.

That's what I've been trying to say, good summation of it

 

[Daedalus685]

That's what I've been trying to say, good summation of it

And what on earth is wrong with that?

Not everyone is in or around the semiconductor business... Someone has to actually present these thoughts.. sure I wish someone who wasn't such a jack ass out for blood would do it.. but stating things that many people often know is what 99% of folks in news, schools, work, etc. spend there time doing. There is nothing fundamentally wrong with it. The problem lies with his technique.

 

[BenSkywalker]

And what on earth is wrong with that?

That ties in with misplaced expectations on my part. Those of us that have been around here for a long time talking about these things is like saying water is wet, it's so obvious when someone starts talking about why it's wet and gets it all wrong and people say they are right it is maddening. We have had many a lengthy discussion about these trends dating back to the time when GPUs were a fraction of the size of CPUs and used a lot less power and were far simpler devices.

The registered members on these forums have included several engineers from 3dfx, at least one from ATi, one from PowerVR and one from Intel that I know of off the top of my head. The level of basic education about the design process in this forum must have dropped a lot more then what I was thinking. We used to be able to get into the finer nuances of the process trade offs and relative merits of using SoI and having an appreciable amount of eDRAM on die and how that compared to bulk manufacture without eDRAM and the cost/build advantage of each- now people don't properly grasp very basic design rules about leakage/yields on new build processes. The level of education revolving around this industry has declined more on average then I had thought on this forum. I've been here a while, it happens sometimes

 

[Daedalus685]

That ties in with misplaced expectations on my part. Those of us that have been around here for a long time talking about these things is like saying water is wet, it's so obvious when someone starts talking about why it's wet and gets it all wrong and people say they are right it is maddening. We have had many a lengthy discussion about these trends dating back to the time when GPUs were a fraction of the size of CPUs and used a lot less power and were far simpler devices.

The registered members on these forums have included several engineers from 3dfx, at least one from ATi, one from PowerVR and one from Intel that I know of off the top of my head. The level of basic education about the design process in this forum must have dropped a lot more then what I was thinking. We used to be able to get into the finer nuances of the process trade offs and relative merits of using SoI and having an appreciable amount of eDRAM on die and how that compared to bulk manufacture without eDRAM and the cost/build advantage of each- now people don't properly grasp very basic design rules about leakage/yields on new build processes. The level of education revolving around this industry has declined more on average then I had thought on this forum. I've been here a while, it happens sometimes

Oh don't be so high and mighty..

Sure Charlie says some profoundly obvious things... but there is at the least some fun in reading his writings. I'd much rather hear him go on than only ever get to hear about how awesome Nvidia is.. At least we get both nut cases this way.

I guarantee that the average person knows no less about the industry than they used to... But what on earth is wrong with not knowing something? Isn't that why people come to a forum, to learn? Isn't it obvious that those that come to learn will outnumber those that teach?

Besides that.. how is a fundamental understanding of how a GPU is made going to allow one to expect Fermi to not be that great? It in no way guarantees that it will be hot and power hungry.. certainly it will be more so based on size but it seemed equally as likely they would have gotten it right. Someone has to let folks know what is going on.. we all wish it was someone better I'm sure.. but you know.

There are two things worse than ignorance... intentional ignorance, and those knowledgeable who refuse to teach. I'm sure I could make a case for those that pretend to be knowledgeable too, but that is kind of like the intentional thing..

Edit: please tell me you are not seriously trying to blame those that believe Charlie with the decline of the forums... It is not because of that that we cant have a technical talk (never thought I'd get more of that in at tom's than here.. lol) it is because many folks listen to the lord of the brand before the rules of logic. That is to say there are too many fucking troll fan bois.

 

[T2k]

Didn't Nvidia buy VIA?

Doesn't VIA have a license for x86.

They do...

In any event I would like to see a large die Fusion chip from the Green Team. (Then we as consumers put our tower coolers on it and solve any noise problems associated with bleeding edge chip designs)

 

[ronnn]

I'd love to know your idea of "sucks".

A chip that has low yields, is unfixable and can't be made to sell with a profit because the competition is doing better. Especially if investment was high and returns will be very low.

Not saying this is true, just regurgitating charlie.

 

[EarthwormJim]

My only complaint with SA articles is that they are written as if everything nVidia does is a personal attack against Charlie.

I don't know about that. I think Charlie really really really just hates Nvidia's CEO whom he calls their "Dear Leader." The non-ending spin after spin after spin from Nvidia's marketing and CEO is annoying, and it is different from most other tech companies.

Would AMD have put together a fake graphics board (held together with Home Depot screws) to show off to the world?

What I don't like about Charlie is that he seems to write as if everything he overhears or is emailed is factual.

I do enjoy his articles. They give a nice view inside the industry, however real it may actually be.


Nvidia just gambled wrong with this generation. They went with a huge architecture on a really immature processes. They're not 100% to blame, TSMC certainly deserves some.

Look at it this way though. The GTX480 is faster than Ati's part and it's partially crippled. Once yields get better, we could easily see a fully enabled GTX part (GTX485?) that might run at even higher clock speeds.

 

[SolMiester]

Charlie is such a sanctimonious jackhole that it's stomach-turning to try and slog thru his bile and gloating to learn anything. A legit site like Ars Technica could've told this story without the juvenile crap, but that's why they're legit and Charlie's just an ATI fanboy troll site.

This.....and I wouldnt want to contribute to his hit count!

 

[BenSkywalker]

I guarantee that the average person knows no less about the industry than they used to

The average poster on this forum certainly knows less about the industry then they did several years ago. I recall the nV DXTC compression issue years ago when we had threads where we isolated the DX call with a hex editor and figured out how to modify the games executable to enable significantly improved IQ when using it while the forum at large not only understood what the hell was being discussed, but actually helped track it all down. That was the general AT populace, or the time when we had a group that tracked down some of 3dfx's engineers to explain to them a problem their AA was going to have prior to the V5 launching- they didn't listen, we were correct and they were forced to address it in a driver release not too long after launch(had to do with their sampling being too sparse in terms of TMUs in relation to non adjusted LOD bias when using their multi buffer RG SSAA approach which didn't scale properly). Going from that level, to people not understanding leakage on a new build process...... it's a pretty drastic fall off.

Edit: please tell me you are not seriously trying to blame those that believe Charlie with the decline of the forums

I'll say that noone that understands much of what happens in the industry is going to believe what he posts.

it is because many folks listen to the lord of the brand before the rules of logic.

Charlie certainly isn't going to help on that end.

Nvidia just gambled wrong with this generation.

People thought that on these forums last generation too and nV ended up with ~66% of the market. Obviously we don't know how their gamble paid off yet, will be interesting to watch if they were wrong or not in the coming months/years though(if the HPC/GPGPU market ends up taking off, this may look like a stroke of genius by nV even if it doesn't dominate the gaming segment).

 

[T2k]

The average poster on this forum certainly knows less about the industry then they did several years ago. I recall the nV DXTC compression issue years ago when we had threads where we isolated the DX call with a hex editor and figured out how to modify the games executable to enable significantly improved IQ when using it while the forum at large not only understood what the hell was being discussed, but actually helped track it all down. That was the general AT populace, or the time when we had a group that tracked down some of 3dfx's engineers to explain to them a problem their AA was going to have prior to the V5 launching- they didn't listen, we were correct and they were forced to address it in a driver release not too long after launch(had to do with their sampling being too sparse in terms of TMUs in relation to non adjusted LOD bias when using their multi buffer RG SSAA approach which didn't scale properly). Going from that level, to people not understanding leakage on a new build process...... it's a pretty drastic fall off.

VS

I'll say that noone that understands much of what happens in the industry is going to believe what he posts.

VS

People thought that on these forums last generation too and nV ended up with ~66% of the market. Obviously we don't know how their gamble paid off yet, will be interesting to watch if they were wrong or not in the coming months/years though(if the HPC/GPGPU market ends up taking off, this may look like a stroke of genius by nV even if it doesn't dominate the gaming segment).

It's a golden topic at 1AM...

 

[ZimZum]

Didn't Nvidia buy VIA? Doesn't VIA have a license for x86.

The merger was rumored but never actually happened.

VIA's x86 license cannot be transfered and evaporates if the company gets bought out.

 

[bryanW1995]

I would like to read up on the 2xx and 48xx parts having a 40% margin... Maybe Newegg marks them up 40%....

as always.... [citation needed]

crap, I can't find them now, but I've read a bunch of charts over the past couple of years that display estimated total cost for different cards. iirc gtx 260 ended up ~ $128 and 4890 ~ $121, though I have not seen the charts in a long time. Original costs were higher obviously on both cards but prices were also quite a bit higher at launch. I think that the 30-40% number is, if anything, conservative.

 

[AnandThenMan]

If you check out the SEC filings they have segment breakdowns.

You seem to be somewhat of an expert on that site, so why not post your findings here for all of us to read?

Look at it this way though. The GTX480 is faster than Ati's part and it's partially crippled. Once yields get better, we could easily see a fully enabled GTX part (GTX485?) that might run at even higher clock speeds.

As the process@TSMC improves, AMD also benefits with higher clocked parts. (never mind any refresh parts AMD has in the works).