The AMD Execution Thread [2007 - 2017]

[Geeforcer]

Selling Fab30 (aka Fab38) before it's even fully converted to 65nm would certinly be very interesting, and it'd even make a lot of sense - assuming they could find an interested buyer, of course.

Hey, I know a larger semiconductor company that knows how to run labs profitably. Maybe they should try them.

 

[overclocked_enthusiasm]

I did mean the entire Barcelona family not just servers. The mobile platform is mid 2008 and I think that will be too late if Barcelona, Agena, Agena FX etc. have been bested by Penryn. This asset light strategy they announced will involve either selling Fab 30 as you suggest, selling the remaining Spansion stake or breaking out the ATI consumer business for sale. At some point it will take an Opteron miracle to save AMD as how can they possibly compete with Intel's superior Fab muscle and R&D budget?

Perhaps they can figure out a way to use Charter effectively but that will reduce gross margins further and likely offset the CapEx saved that would have been used to operate the Fab capacity Charter replaces. AMD is just in a really bad spot right now. The ONLY way out is superior technology. Intel shows no sign of giving that back anytime soon and seems commited to using their fab process/capacity and R&D advantage to kill AMD. They are moving up their roadmaps as quickly as possible to further highlight this fab process advantage because AMD cannot match the ramp speeds. As a result, AMD has one bullet in their gun right now...slash pricing.

The recent Barcelona POV tests seem to indicate AMD is appealing to the customers that have Opterons already. "See how well these scale with no power increases? Just drop them in and you have upgraded". They have yet to demonstrate raw power, higher clocks and uber performance for these parts. It seems they are just trying to keep their established AMD base from jumping ship not really trying to capture additional share from Intel. I could be reading the tea leaves wrong but Barcelona seems to be pretty quiet this close to launch. Quiet because initial clocks and performance will be lower than hoped...or lower than what Intel will offer? Seems so to me. Or, it is delayed as the rumors suggest?

 

[BRiT]

But they everyone would have bought a G8x and noone would have waited for the R600. So from the strategic POV, it's better that they didn't say it

It might have been a decent tactical move, but a very bad strategic move, perhaps the worst one they could have made.

Not saying or giving guidance on the state of R600 was, they generated a lot of bad-will towards enthusiasts who were waiting for the card. This fostered the response of upgrading to a G8x AND being upset and mad with AMD/ATI. This is not how a company retains or wins over consumers. This is how you lose consumers, even the long time consumers.

Pair ATI's behavior on R600 with AMD's behavior on Socket 939 and you have a combined company that carries a lot of enthusiast bad-will. Consumers are not always a logical bunch, and at times remain a very emotional bunch -- just look at the various boycott Sony groups.

 

[cadaveca]

The last 3 companies I know that switched to "Asset-Light Manufacturing" (meaning we make 'em, on order only, and hold not stocked parts) are now peiced off and sold to bankruptcy trustees.

I hate to say it...but this really is a bleak outlook for AMD. Time and time again ATI has been bitten by the TSMC "leaky gpu" bug, time and time again they go right back and bend over and say "gently this time, ok?".


I've been watching investor meetings for a while now at both companies. Hector looks more and more stressed each time, looks less confident, and really seems to not have a decent handle on how the company is progressing.

The rest of the upper management looks the same...they are all hiding big secrets, and it's starting to wear very thin...we see a little bit more each time...


When upper management behaves this way, it's no wonder the rest of the company falls right in line...they think this is how it's done...heck...they get paid the big bucks, so we must do as they do...

A sad, sad, SAD state of affairs...I'm glad I made some cash when I sold my stock instead of losing a ton.

 

[3dilettante]

3dilettante: Nice post on AMD's recent execution, I guess that's an aspect I don't tend to focus on much, heh. I certainly agree with most of what you said, with perhaps one or two slight primary exceptions: I'd trust Ruiz much less than the CEO of another company which is losing, because not every loser is incompetent at predicting the future (some just know they're doomed), and not every loser feels it is their duty to hype themselves up to the point that people would think they plan to completely annihilate the competition.

I see your point, though I don't chalk it up to incompetence so much as desperation.
Many of these storm clouds over AMD's head have always been there, even in the sunniest times.

I think in part Ruiz has to hype the future because the future is all AMD has to offer for further financing.
Its present isn't all that enticing.

I don't think Ruiz is incompetent, so much as he's trying to bilk investors in the hope that AMD can use the ill-gotten money to survive.

I also don't completely agree with what you said wrt the various versions of R4xx; I actually think ATI managed that juggling quite nicely. It certainly didn't turn out as good as if they never had to change things around at all, but that's pretty much always the case; as least they managed to decdent backup plan. There certainly were some problems with ATI, I wouldn't think that was the most significant one though.

I didn't want to weigh too heavily on ATI's cancelled designs for the reason that I didn't have a similar knowledge of the cancelled cores for Nvidia.
I just wanted to point out that both AMD and ATI had a history with missing out repeatedly on the "next best thing".

Since both ATI and Nvidia were much closer to parity when it came to size and resources than AMD and Intel, design cancellations would not have the disproportionate impact that a misstep of similar size would have on AMD's CPU fortunes.

 

[Periander]

What guidance did they give for Q2? As in, should we expect a warning in a couple of weeks?

They guided flat to up slightly in the Q1 CC. I think they are unlikely to make that, but whether it will be bad enough for a warning is an open question. I've heard several reports that at least part of the Q1 CPU decline was the result of Q4 channel stuffing, and that unit shipments may actually be up this quarter.

That has to be balanced against what will undoubtedly be lower CPU ASPs and likely lower results from the former ATI.

 

[overclocked_enthusiasm]

http://www.legitreviews.com/article/507/3/

More proof all the talk about being so far along with DX10 drivers compared to Nvidia was total BS. Just wait till R600...just wait till GDDR4 2900XTX...just wait till DX10 games...just wait till Barcelona...see a pattern? It really bothers me that all the assurances of future performance made by ATI/AMD have turned out to be false. Why should we expect Barcelona to be any different?

It's that whole credibility issue again...

 

[Dave Baumann]

A lastest WHQL driver compared to a beta released directly for the title doesn't really mean much.

 

[BRiT]

A lastest WHQL driver compared to a beta released directly for the title doesn't really mean much.

Tell that to the consumers who are enjoying the CoH DX10 patch with improved performance now. There's no reason AMD has to hold back their drivers with similar performance improvements until they're WHQL'd months later.

 

[satein]

Tell that to the consumers who are enjoying the CoH DX10 patch with improved performance now. There's no reason AMD has to hold back their drivers with similar performance improvements until they're WHQL'd months later.

Somehow, something unexpected could happen without expectation I would respect that they did trying at their best but it might turn out not as good as whatever us, the consumer, expectation. Sorry again, for sounding rather off-topic.

 

[vertex_shader]

Tell that to the consumers who are enjoying the CoH DX10 patch with improved performance now. There's no reason AMD has to hold back their drivers with similar performance improvements until they're WHQL'd months later.

And they are enjoying problems with other games, every twimtbp game has a specific beta driver, what cause many problems in other games, and when finally nv release a whql driver, performance slower than with the game specific beta driver.

 

[pelly]

And they are enjoying problems with other games, every twimtbp game has a specific beta driver, what cause many problems in other games, and when finally nv release a whql driver, performance slower than with the game specific beta driver.

So long as there is a "Beta" designation to the driver in question, compatibility issues can be overlooked as the driver should merely be used as a preview of the level of performance being sought in a particular title. Certainly, every consumer who has already purchased a DX10 card is dying to get their hands on some DX10 content. Given the amount of time it takes to get a driver "WHQL" certified (think weeks here), each company can only respond with a "beta" driver.

Although it seems logical for each vendor to release the most polished beta they have in-house as soon as possible, there is definitely a level of strategy to these driver releases.

Looking at AMD, they have to be extremely careful here to avoid having R600 be met with further negativity. If the GPU is not performing up to expected levels of performance with current Beta drivers, they will not push for an aggressive beta driver release until the issues (hopefully) get sorted out. (The "expected" level of performance here it to at least match the 8800 GTS...although some will say R600 was supposed to excel at DX10 meaning it could be expected to perform closer to the 8800 GTX).

Looking at NVIDIA, the company also has to be somewhat careful. If the company has an ace up its sleeve in terms of performance...they will hold onto this card until they know exactly what R600's final DX10 situation will be. R300 showed that there was some exceptional talent at ATI and NVIDIA certainly doesn't want to be surprised here and not have any answer waiting in the wings. With that said, a strong new driver could very well act as another nail in R600's coffin. If the 8800 GTS can prove to be substantially faster in most DX10 (and DX9) titles (5fps or more I'd imagine) then AMD will have a major issue on their hands.

Per usual, time will tell.......Hopefully, each side has a "magic bullet" driver release in the making and we'll see some real fireworks!

 

[overclocked_enthusiasm]

A lastest WHQL driver compared to a beta released directly for the title doesn't really mean much.

The implications made by ATI/AMD to date concerning R600 have been the following:

1. Vista drivers are more mature and stable than the competition
2. R600 will really show its stuff with DX10 content

Now explain to me why Nvidia appears to be as far along as ATI with WHQL drivers for Vista (158.24) and now a Beta driver (158.45) to highlight the advantages of 8800 series with DX10 content? Where is the driver superiority that ATI promised for so many months? Why isn't it reflected in the benchmarks?

 

[3dcgi]

The implications made by ATI/AMD to date concerning R600 have been the following:

1. Vista drivers are more mature and stable than the competition
2. R600 will really show its stuff with DX10 content

Now explain to me why Nvidia appears to be as far along as ATI with WHQL drivers for Vista (158.24) and now a Beta driver (158.45) to highlight the advantages of 8800 series with DX10 content? Where is the driver superiority that ATI promised for so many months? Why isn't it reflected in the benchmarks?

You're not the only one to say this, but where was point number 2 stated by ATI/AMD? All I remember is some developer comments that geometry shading is faster with R600 and it seems that people took this comment and drew their own conclusions.

 

[Geo]

You're not the only one to say this, but where was point number 2 stated by ATI/AMD? All I remember is some developer comments that geometry shading is faster with R600 and it seems that people took this comment and drew their own conclusions.

There is at least one AMD-provided slide we've seen that claims up to 50x GS performance vs "competing implementations". The programmable tessellation engine is also claimed to accelerate Geometry amplification, even tho it is technically outside the DX10 spec. We've asked Eric Demers some questions about these and other R6xx technical issues, and hopefully will be able to provide those answers in the days to come.

 

[Dave Baumann]

The tesselator has nothing to do with GS geometry amplification; they are not related.

 

[Geo]

The tesselator has nothing to do with GS geometry amplification; they are not related.

Well, I said it was outside the spec. You added "GS" in that formulation, not me. It does do geometry amplification and AMD is promoting it as such, right? I can put the slide up if you want to see the context. . .it's pg 20 of the architecture deep dive. . .

 

[Dave Baumann]

The tessellator does tessellation, which is a form of geometry amplificaition, but we don't talk about the tessellator in relation to geometry amplification via the GS.

 

[Geo]

The tessellator does tessellation, which is a form of geometry amplificaition, but we don't talk about the tessellator in relation to geometry amplification via the GS.

Well, the context of the slide is Geometry Performance

So it starts to sound a little like HDR and SM3. . . (i.e. the popular conception vs the technical reality) And, yes, left vs right on that page makes it clear the tessellator is outside the geometry shader. That's what I meant by "outside the spec".

 

[overclocked_enthusiasm]

You're not the only one to say this, but where was point number 2 stated by ATI/AMD? All I remember is some developer comments that geometry shading is faster with R600 and it seems that people took this comment and drew their own conclusions.

Sorry I can't find the quote I was looking for to support #2...perhaps I was mistaken. I thought these tidbits from Richard and Orton were telling though.

http://www.hardwarezone.com/articles/view.php?id=2229&cid=30&pg=2
HWZ: With the merger, it is understandable that certain products have been realigned, but could we have an update on the R600 delays?

Henri Richard: The R600 will be out in the second quarter. The reason we decided to delay the launch was that we wanted to have a complete DX10-enabled solution top to bottom. A lot of people wrote that the reason it is delayed is because of a problem with the silicon, but there is no problem with the silicon. We are demonstrating it. We can ship it today. But if you think about it, looking at where the market is at, the volumes are going to be in the R610 and R630, so it makes sense for us to do a one time launch of the entire family of DX10 enabled products. That meant delaying the R600 for a few weeks, but frankly it doesn't make a difference in the life cycle of the product and talking with our customers and partners, they felt that it would make a bigger impact with one full launch. So we decided to do that.

Also increasingly in particular with Vista, as we've seen with the competition. It doesn't matter if you're shipping the silicon if the drivers are not stable. There is nothing more frustrating than having bought a new graphics card and having your system crash repeatedly because the drivers are not ready. Although we today, even by Microsoft standards, have the best and most stable drivers in the entire industry. The few weeks will give us even more time to continue improving the drivers. Again, the decision lies in the fact that we will have a top to bottom DX10 offering with drivers that will have a very very high level of stability and the only difference is a few weeks. So it seems to make a lot of sense to do it that way.

http://www.futurehardware.in/558419.htm
Speaking at yesterday's AMD Financial Analysts Day, Executive Vice
President of Visual and Media Businesses, Dave Orton, appeared to throw
down the performance gauntlet in favor of AMD's upcoming R600 GPU.
Having had over a month to study NVIDIA's G80, Orton did not seem the
least bit intimidated. In a slide entitled "R600: Why we lead in
graphics", Orton promised that even if the name of the company had
changed, that the commitment to GPU performance leadership had not. He
promised a "take no prisoners" approach to performance leadership for
AMD's new GPU.

More interestingly, in his verbal remarks Orton reported (at roughly
the 1:22:30 mark of the webcast) that one of R600's key advantages
would be "new levels of memory bandwidth to the graphics subsystem, and
bandwidth is critical to graphics performance."

http://www.extremetech.com/article2/0,1697,2099613,00.asp
"There's speculation about what's going on with CPU business," said Henri Richard, AMD's executive vice president and chief sales and marketing officer. "Where's Barcelona? I can tell you it's more of a killer product than anyone thought. I'm not bragging. I know what we've got."

Wednesday's meeting was also used to clear up what Richard characterized as a swath of misconceptions and rumors concerning the delayed R600 —or, what is to be AMD/ATI's first DirectX 10-capable graphics card.

"We pushed out the launch of the R600 and people thought is must be a silicon or software problem…it's got to be a bug," said Dave Orton, president and chief executiveof ATI. "In fact, our mainstream chips are in 65nm and are coming out extremely fast. Because of that configuration, we have an interesting opportunity to come to market with a broader range of products," he explained.

"Instead of having them separate, we thought, lets line that up, so we delayed for several weeks," Orton continued, referring to the R600 family as a whole, which AMD now says will come out at the same time (a matter of weeks as opposed to months, according to Richard) instead of just the high-end version.

The R600 will be called Radeon x2900 and will be available in XTX (top of the line), XT, and XL variants.

http://seekingalpha.com/article/32901
David Wong - AG Edwards

Right. Could you give us any quantification of sequential microprocessor units growth or declines in any of the segments?

Henri P. Richard

We typically do not give you those indications.

Robert J. Rivet

The only comment I would make to give you the best color would be ASPs were the small piece of the equation in the sequential movement quarter to quarter. It was mostly units. ASP were a piece of the equation but the driver was the unit drop quarter on quarter. That was in all segments, whether it is desktop, mobile, server.

Henri P. Richard

I would like to add one point. We need to wait a couple of weeks until all the analysts are publishing the sales out figure for the quarter, because we know that customers really dispose of inventory that they had built in the fourth quarter, and so you might be actually surprised by our sales out market share numbers versus what you are looking at today, which is sales in.

Gurinder Kalra - Bear Stearns

Just a question on your R600 launch. Can you provide us more details in terms of timing, whether it will be a hard launch or a soft launch, how much product do you expect to have and in what categories? Would there be a performance product or a performance mainstream product?

Henri P. Richard

We are going to announce those products in the second part of May. The location has been picked. It is a series of 10 products covering the entire stack with DX10 capabilities, leading edge performance and more importantly, a solution that works with [WECO] qualified drivers.

Gurinder Kalra - Bear Stearns

Would you expect to have substantive quantities of product available at the launch?

Henri P. Richard

Absolutely. We do not do soft launches.