Charlie D Claim Watch

Page 7 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

BenSkywalker

Diamond Member
Oct 9, 1999
9,140
67
91
The accuracy we are referring to includes many things. Such as how the 512core part will likely not see the light of day. That the card will not blow the 5870 out of the water. That there are leakage problems, yield problems, etc. That the supply will be limited.

So built in redundancy, performance being close to current(ie- same not pre existing) generation and leakage and yield problems on a new process are news to some people? OK, I guess I just vastly overestimated the level of education people on this forum have regarding the entire process from start to finish. For the record, all of those things have been entirely commonplace for every new part on a new build process. If those are the things people are saying he is accurate on I guess I can understand.

You can't take him so seriously Ben..... You'll have a heart attack.. If him being wrong a few times means you can't trust a word he says then don't... but don't try to convince others to ignore him based on what amounts to a logical blasphemy.

If you claim hard numbers, expect to be called out on it. This goes for any author of anything. If he stated things like 'seeing clockrates spread wildly in the 500MHZ-700MHZ range depending on the part' then it is very easy to retain credibility while reporting a rumor. When you state absolute numbers and present them as fact, you are putting your credibility on the line. I state anything in a thread you don't think looks right and you ask for a link to back it up which is perfectly reasonable. Charlie never backs up anything as a journalist and to some, that is OK. There is a way that he could report rumors while not looking like a total fool- even keeping his bias. Don't nail yourself down to exacting numbers when they are in a state of constant flux, only fools do such a thing.
 

Daedalus685

Golden Member
Nov 12, 2009
1,386
1
0
So built in redundancy, performance being close to current(ie- same not pre existing) generation and leakage and yield problems on a new process are news to some people? OK, I guess I just vastly overestimated the level of education people on this forum have regarding the entire process from start to finish. For the record, all of those things have been entirely commonplace for every new part on a new build process. If those are the things people are saying he is accurate on I guess I can understand.



If you claim hard numbers, expect to be called out on it. This goes for any author of anything. If he stated things like 'seeing clockrates spread wildly in the 500MHZ-700MHZ range depending on the part' then it is very easy to retain credibility while reporting a rumor. When you state absolute numbers and present them as fact, you are putting your credibility on the line. I state anything in a thread you don't think looks right and you ask for a link to back it up which is perfectly reasonable. Charlie never backs up anything as a journalist and to some, that is OK. There is a way that he could report rumors while not looking like a total fool- even keeping his bias. Don't nail yourself down to exacting numbers when they are in a state of constant flux, only fools do such a thing.

All aboard the fool train then... Profoundly ironic given a certain claim of hard numbers in another thread...

Oh fuck it.. I'm gonna go have a conversation with my cat...
 

Lonyo

Lifer
Aug 10, 2002
21,939
6
81
So built in redundancy, performance being close to current(ie- same not pre existing) generation and leakage and yield problems on a new process are news to some people? OK, I guess I just vastly overestimated the level of education people on this forum have regarding the entire process from start to finish. For the record, all of those things have been entirely commonplace for every new part on a new build process. If those are the things people are saying he is accurate on I guess I can understand.

If they are so obvious and commonplace... why didn't NV manage to plan for them?
 

yh125d

Diamond Member
Dec 23, 2006
6,907
0
76
I'm honestly trying to find the accuracy that people are talking about does it really come down to-

Going to ship later then ATi
Going to be bigger then ATi
Going to be hotter then ATi?

Is that what everyone sees when they talk about the accuracy? I'm trying to grasp this, we have his biggest claim left on the table, and noone wants to say he's right so far- Fermi can't be mass produced. I'm ignoring the finer details, are the three points listed above why people consider him accurate?



What information? That is what I'm asking. People say look at the OP, I go through the list and none of it is right, so I look to the broader picture that he has left, and the only things I come up with are the three general points above. All of those things were known before Charlie said anything.

You could try reading the OP if you wanted.

Correct claims:

"According to sources in Santa Clara, GT300 has not taped out yet, as of last week."

"So, if all the stars align, and everything goes perfectly, Nvidia could hit Q4 of 2009. But that won't happen." - kind of an odd one, but the point being that things will be far from peachy and wont be out before '10

"If there is a second respin, then you might have a hard time hitting Q1 of 2010." - There were 2 respins, missed Q1 '10 (for card availability)

(in another article) "This means if everything goes perfectly, not so much as a hiccup, we are looking at mid-December....the chances of this happening are somewhere around zero though." - reiterating, no fermi in 2009

"No matter what happens, at this point, there isn't enough time to go from tapeout to anything more than a few hand-built publicity stunt parts before 2010." - wood screws anyone?

"If it tapes out tomorrow, and needs only one respin, don't looks for parts on sale before March. Given NV's execution of late, a second spin is almost a given, so we are now well into Q2/2010." - April 12? for actual availability

"In the end, there won't be any GT300s this year, and early next year is really iffy" - again

"NVIDIA IS KILLING the GTX260, GTX275, and GTX285 with the GTX295 almost assured to follow as it (Nvidia: NVDA) abandons the high and mid range graphics card market." - This is a point of contention, some contend that since there are still technically parts "for sale" he's wrong, but considering how many models have been EOL'd, how so many are continually out of stock and how prices have risen above prices from a year ago it's safe to say NV has pretty much abandoned GTX2xx cards. Theres more GF6 series cards for sale

"IT LOOKS LIKE Fermi A2 silicon has finally taped out, so the timetables are a little firmer once again. There is no chance of a real launch in 2009" - once again

"and that is with two of the 16 shader clusters clusters fused off. With them on, it would likely be well above 250W, far too hot for a single GPU card" A 448 + 1 cluster, aka a GTX480 does pull well above 250w with higher clocks than he thought at the time. one with both clusters but the lower clocks he thought at the time would be correct, as it'd probably consume as much as a GTX480 (or more)



And that's only some of his claims. You'd be an absolute ignoramus to pretend he got nothing right
 

BenSkywalker

Diamond Member
Oct 9, 1999
9,140
67
91
If they are so obvious and commonplace... why didn't NV manage to plan for them?

You are assuming that they didn't? Their margins improved last quarter versus year ago, markedly. They maintained enough supply of their previous gen parts to keep the pipes fed until the launch of their new part and have ramped production on the new parts inside of the quarter that they stated. I know enthusiasts have their own schedule, but based on nV's stated business goals they are right on schedule. That isn't to say having a part out six months earlier couldn't have possibly benefited them, but looking at the angle they are taking being late in the consumer enthusiast segment was worth the trade off to include the HPC functionality that they are counting on to help them expand into new markets. I'm not saying that they are headed in the right or wrong direction, just pointing out that it is the direction that they have explicitly spelled out to investors for quite a while now.

From a *business* perspective nVidia has thoroughly bested ATi in the last six months. I understand the huge rift between financials and what enthusiasts want, but in terms of them executing as a *business* they have bested ATi the last two quarters. I'm not saying that is what anyone on these boards should care about, simply pointing it out as you asked about nV planning for these events to happen as they did. All indicators are that they did plan for it, nearly perfectly(again, not saying it is the ideal sitation that enthusiasts want to see, just looking at it from a business perspective).

And that's only some of his claims. You'd be an absolute ignoramus to pretend he got nothing right

They were later then ATi is what all of that comes down to. I won't nitpick but in summation all of the particulars he claimed that I'm seeing are pretty much wrong in terms of what tapeout/respin happened when(initial tapeout is A0, A3 is the third respin- fourth time it was sent to TSMC). I'm sure if he had that right he would have pointed it out, nVidia did worse then what he claimed which he doesn't seem to ever want to miss.
 
Last edited:

yh125d

Diamond Member
Dec 23, 2006
6,907
0
76
Charlie never backs up anything as a journalist and to some, that is OK.

Do you know absolutely anything about journalism? Because that's one of the fundamental aspects of it. When you're dealing with stuff that's under NDA like this you -can't- "back it up" without getting sued/getting people fired. He's not choosing to not back stuff up because he's pulling it out of his ass, it's because he can't


The cardinal rule of journalism is you don't give up your sources. That's how the business works.
 

Daedalus685

Golden Member
Nov 12, 2009
1,386
1
0
You are assuming that they didn't? Their margins improved last quarter versus year ago, markedly. They maintained enough supply of their previous gen parts to keep the pipes fed until the launch of their new part and have ramped production on the new parts inside of the quarter that they stated. I know enthusiasts have their own schedule, but based on nV's stated business goals they are right on schedule. That isn't to say having a part out six months earlier couldn't have possibly benefited them, but looking at the angle they are taking being late in the consumer enthusiast segment was worth the trade off to include the HPC functionality that they are counting on to help them expand into new markets. I'm not saying that they are headed in the right or wrong direction, just pointing out that it is the direction that they have explicitly spelled out to investors for quite a while now.

From a *business* perspective nVidia has thoroughly bested ATi in the last six months. I understand the huge rift between financials and what enthusiasts want, but in terms of them executing as a *business* they have bested ATi the last two quarters. I'm not saying that is what anyone on these boards should care about, simply pointing it out as you asked about nV planning for these events to happen as they did. All indicators are that they did plan for it, nearly perfectly(again, not saying it is the ideal sitation that enthusiasts want to see, just looking at it from a business perspective).

By what metric have they bested ATi?

The problem is that Nvidia went on and on about how fast and great the thing would be... most of us knew it would not be, and some of us enjoy reading rumour mongering..

I'd guess that 90% of us don't give a shit about the company Nvidia, only about how the parts perform. Those folks were let down by this particular product for one reason or another (perhaps they really thought it would be 60% better than the 5870... perhaps they wanted a cool card, whatever).

Your stance was that since Charlie was wrong about a few things he is not worth reading... Maybe he isn't telling us anything new but he was about the only person that gave out anything about the process of the GPU being manufactured. It was genuinely interesting, and useful beyond him being right or wrong. If every rumour is based on its exactness after the fact then almost none would be 'right.'

What on earth is the problem with using a rumour site like Charlie's as an interesting and possible source of information?

I can assume Fermi would be big, expensive and power hungry... but I'm not sure that I would have gotten quite the idea of how bad the leakage and efficiency would actually be without his ravings...
 

Janooo

Golden Member
Aug 22, 2005
1,067
13
81
The initial launch shipment. He stated how many parts were going to ship before the product was EOL, that is going to be covered by the first shipment to hit retail.
...

Do you have a quote? Where does he say that it's only initial launch shipment and nothing else?
I don't believe 9,000 risk wafers is equal to one launch shipment.
So do you have the quote?
 

GaiaHunter

Diamond Member
Jul 13, 2008
3,628
158
106
They were later then ATi is what all of that comes down to. I won't nitpick but in summation all of the particulars he claimed that I'm seeing are pretty much wrong in terms of what tapeout/respin happened when(initial tapeout is A0, A3 is the third respin- fourth time it was sent to TSMC). I'm sure if he had that right he would have pointed it out, nVidia did worse then what he claimed which he doesn't seem to ever want to miss.

They were later than their own schedule.

What are NVIDIA yields atm?

Do they have a 512 GTX 480 for sale?

Are they actually selling any Geforce based Fermi at all?


You look at the fact the GTX 480 is 700 MHz and 480 shaders and will start selling in a couple of weeks to say Charlie is wrong.

I see the exact same things as a proof that NVIDIA did have the problems Charlie claimed.

I guess you just don't know what kind of information to extract from a rumour site.
 

BenSkywalker

Diamond Member
Oct 9, 1999
9,140
67
91
By what metric have they bested ATi?

Revenue and profits. This goes back to the SEC filings, maybe it's easier to serach the forums for the last time I broke them all down and linked to them, but last quarter as an example nV had higher gross profits then ATi had total revenue. This is why I clearly stated as a business, nothing at all to do with it from an enthusiast perspective.

What on earth is the problem with using a rumour site like Charlie's as an interesting and possible source of information?

Because of how much he gets horribly wrong. Idontcare, who is a forum member here, worked in microprocessor design for quite some time and would be a far better source to ask about the process if you want to get into details- he worked for TI IIRC so his breakdowns aren't biased towards one design philosophy or the other currently being used. Some of the things Charlie does get wrong are elementary and confuse people needlessly. Anyone following microprocessor design for any length of time knew all about what Fermi's issues were going to be, Charlie tried to make a soap opera over a normal operation for such a part.

I don't believe 9,000 risk wafers is equal to one launch shipment.

That is based on Charlie's claims versus what Anand stated they will ship at launch. It is well within possible outcomes that nV lied to Anand, but it seems that board partners are backing up what he was told.
 

Daedalus685

Golden Member
Nov 12, 2009
1,386
1
0
Revenue and profits. This goes back to the SEC filings, maybe it's easier to serach the forums for the last time I broke them all down and linked to them, but last quarter as an example nV had higher gross profits then ATi had total revenue. This is why I clearly stated as a business, nothing at all to do with it from an enthusiast perspective.

ATi is a part of a large company, they are the part that makes only GPUs. Nvidia has many more products than ATI. The two are not comparable. Perhaps we could compare as a ratio of Profit to R&D but I'm sure there are clever ways of AMD moving around some of the nicer numbers from its devisions to make the whole look better.

The point is it is far too complicated to just compare them like that. We can say that Nvidia makes more profits than AMD, thus are a better company than AMD and leave it at that.. but that is about it. There is no way we can really evaluate just ATI anymore.


Because of how much he gets horribly wrong. Idontcare, who is a forum member here, worked in microprocessor design for quite some time and would be a far better source to ask about the process if you want to get into details- he worked for TI IIRC so his breakdowns aren't biased towards one design philosophy or the other currently being used. Some of the things Charlie does get wrong are elementary and confuse people needlessly. Anyone following microprocessor design for any length of time knew all about what Fermi's issues were going to be, Charlie tried to make a soap opera over a normal operation for such a part.
That is a good reason for folks not to take it without grain of salt, or why folks that do not understand should avoid him... But frankly if one does not have at least a fundamental knowledge of the industry they should not be regarding rumours anyway. That goes for most things.. If a person is not capable of separating the good from the bad in everything they hear there are more serious problems than the guy who is wrong some of the time.
 

Janooo

Golden Member
Aug 22, 2005
1,067
13
81
...
That is based on Charlie's claims versus what Anand stated they will ship at launch. It is well within possible outcomes that nV lied to Anand, but it seems that board partners are backing up what he was told.
None of them stated that the whole 9,000 wafers were used. What if they were split to 2-3 batches (one now, one in about a month, ...)?
Let's assume that they used all of them. If NV is going to have 50K cards at the launch, 9000 wafers, 104 candidates per wafer then 50K/(9,000*104) is about 5% yield. Which means it's 5 good dies per wafer. Are you sure this is the case? That's what you imply by saying that NV used all wafers.

Charlie talks about 10 good dies per wafer. If there are 50K cards then there could be 2 batches (2*4,500), if there are 30K cards at the launch then maybe 3 batches (3*3,000) or one more 6K of 9,000 risk wafers. The 2nd and the 3rd batch will have better yields assuming TSMC improves manufacturing. I would say we can easily see 150K - 200K cards from the risk wafers and for sure it's not going to be one launch shipment.

Then the fun starts. What's next? Fixing Fermi1 or going for Fermi2 as soon as possible?
 

Daedalus685

Golden Member
Nov 12, 2009
1,386
1
0
None of them stated that the whole 9,000 wafers were used. What if they were split to 2-3 batches (one now, one in about a month, ...)?
Let's assume that they used all of them. If NV is going to have 50K cards at the launch, 9000 wafers, 104 candidates per wafer then 50K/(9,000*104) is about 5% yield. Which means it's 5 good dies per wafer. Are you sure this is the case? That's what you imply by saying that NV used all wafers.

Charlie talks about 10 good dies per wafer. If there are 50K cards then there could be 2 batches (2*4,500), if there are 30K cards at the launch then maybe 3 batches (3*3,000) or one more 6K of 9,000 risk wafers. The 2nd and the 3rd batch will have better yields assuming TSMC improves manufacturing. I would say we can easily see 150K - 200K cards from the risk wafers and for sure it's not going to be one launch shipment.

Then the fun starts. What's next? Fixing Fermi1 or going for Fermi2 as soon as possible?
If I were to guess I'd say a small fix in the fall with fermi 2 as soon as 28nm is around. They will not spend any extra resources fixing fermi 1 if they can divert them to ensuring two is sound. That is where the money will be. The architecture looks promising. Fermi two could be a godsend to HPC and gamers alike.. they won't want it any later than it has to be.