The AMD Execution Thread [2007 - 2017]

[overclocked_enthusiasm]

No it does not make sense, unless you can prove AMD put a lot of store by not doing awful executiuons. Any examples ?

K7?
K8?
K10?

All poor on release and needing massive intervention by IBM later.

AMD asking for someones head due to awful execution. Funny.

I think perhaps Orton oversold R600 in regards to launch timing and performance. While this may not be any different than the examples you listed above it does differ in that the intellectual property of ATI was the driving force behind the $5.4 billion merger. Given that this intellectual property may not be as valuable as advertised AMD may have felt they were sold a bogus bill of goods and the heat on Orton was too much.

 

[swaaye]

K7?
K8?
K10?

All poor on release and needing massive intervention by IBM later.

K10 hasn't launched yet. K7's only problem was that AMD had a tough time getting chipset and mobo manufs behind them. They beat Intel to 1 GHz with K7. And K8...what was wrong with K8? I remember no disappointment there. I remember Intel scrambling to put a souped-up Xeon onto the desktop to keep up with it.

 

[Fox5]

K10 hasn't launched yet. K7's only problem was that AMD had a tough time getting chipset and mobo manufs behind them. They beat Intel to 1 GHz with K7. And K8...what was wrong with K8? I remember no disappointment there. I remember Intel scrambling to put a souped-up Xeon onto the desktop to keep up with it.

K8 was late, and that misstep is largely what allowed the p4s to gain a clear advantage over the athlon xps. Had k8 launched on time, it would have launched when the top athlons were at least tied with the p4s, if not faster. Instead, amd tried to make faster athlon xps to make up for the late launch (I think it was late by at least 6 months, if not a whole year) and couldn't really succeed.
I'd imagine they had yield issues as well, since the 1.6ghz opterons did launch at least close to on time, and would have compared well with amd's athlon xp desktop offerings of any speed. Actually, that's likely what will happen with k10, slow speed ones that are competitive with the top of the current amd chips will launch server side first, and then eventually make their way to the desktop once amd gets all the issues worked out, then ramp clock speed.

 

[Geeforcer]

Let me guess, he resigned to "spend more time with his family"â„¢.

 

[LeStoffer]

In the end, I'm sad to see Dave Orton go, because he did a lot for the industry as a whole. Did he do the best possible job for his company? No. But he made a profound impact on the industry as a whole over his years at SGI, ArtX and ATI.

His greatest mistake of all times, I believe, will have been to sell off the company to AMD. As a CEO, he had a duty not only to strive to grow, but also to strive to minimize risk. While he did a good job overall in the former, he failed beyond words at the latter, both in this instance and in others.

So very true, espicially about the terrible, terrible mistake to sell ATI to AMD.

 

[dizietsma]

K7's only problem was that AMD had a tough time getting chipset and mobo manufs behind them. .

I seem to recall the VIA KT266a - 333 - 400 chipsets quite popular actually. The only problem was actually some versions of the chip itself.

http://www.geek.com/procspec/amd/thoroughbred.htm

"There's two flavors of the Thoroughbred core, the first hot-running and not-very-scalable Thoroughbred A and the cooler and higher-clocking Thoroughbred B. The "B" core is slightly larger due to a new layout designed to improve scalability"


The chipset problems only occurred with K8 release when VIA chipsets did not have a lockable AGP PCI bus I seem to recall.

 

[_xxx_]

So very true, espicially about the terrible, terrible mistake to sell ATI to AMD.

That was not a mistake. If they hadn't, there would be no ATI in any form today at all. R600 would have made them bankrupt, with certainty.

 

[Geo]

Can I have next week's lotto numbers too? I mean, if you're going to do that kind of fortune telling with certainty, can't it at least be for something useful?

/me waits for _xxx_ to reply, "I can only tell you what next week's lotto numbers would have been if AMD hadn't bought ATI. . . "

 

[_xxx_]

Hehe, no Lotto numbers My industry-related predictions just come from living in the very heart of one of the cruelest (internally) corporations ever. I witness rolling heads almost daily here, for way smaller f**kups than what ATI management did since R420.

 

[Arun]

Hehe, no Lotto numbers My industry-related predictions just come from living in the very heart of one of the cruelest (internally) corporations ever. I witness rolling heads almost daily here, for way smaller f**kups than what ATI management did since R420.

NV30 didn't make NVIDIA bankrupt. I mean, if they had absolutely NO CASH RESERVES whatsoever, and that their MCP business died at the same time, then who knows. It would still have been unlikely given how fast NV40 came, but who knows.

However, cash reserves made NVIDIA survive at absolutely no short-term or mid-term risk whatsoever. The same could have been true for ATI. The same is not true for AMD. This is the key problem that I'm not sure you realize here. There's more to a company than its income, the balance sheet is really just as important in the short- and mid-terms.

 

[_xxx_]

But at that time nV was much stronger financially than ATI was before the takeover, also literally without any serious competition until R300 made its rounds. Not comparable at all.

ATI on its own was at risk though. How much cash did ATI have before the takeover? How much was lost since? Simple math, the current numbers would be VERY red.

EDIT: and for the exact reasons you named I expect the GFX division at AMD to shrink quite a bit over the next year or two.

 

[Arun]

How much cash did ATI have before the takeover? How much was lost since? Simple math, the current numbers would be VERY red.

Correction: the current numbers would be VERY green. You are presuming things here with massively insufficient basis, as far as I can tell.

Look at Page 4: http://www.ccnmatthews.com/docs/ATI.financials.pdf
ATI's cash position in May 2006 was about $520M. This is a reduction of about $120M year-over-year, with a net income of about $100M in that same period. Let's summarize that by not considering the $25M of stock repurchase and say that with net income of $100M, they had a cash flow of -$100M.

So, the naive conclusion we can draw is that with a net income of $0M, they'd have had a cash flow of -$200M. That's not very accurate and it's kind of naive, but whatever - I think it's a fair worst-case scenario. About one year after the acquisition, I'd expect an aggregate net loss of about $50M at worst; remember that they'd have kept chipsets, making that profitable, and CE was profitable until Q1. That leaves them with about $270M of cash.

If they reduced capital expenditures and/or that CE/Chipsets were profitable, they could do just fine for quite a while. Add in the fact that getting $200-300M of debt for that kind of company is probably very easy, and I don't really see how they could possibly have gone bankrupt pre-R800, and once again that's a worst-case scenario. Most likely, things would have been a fair bit better than that IMO...

At this point, however, the problem is fairly simple: AMD is in a pretty much catastrophic position, and they NEED a good strategy *and* outside help to come out of this victorious. If you think R6xx is a fuckup (and I personally think people are exagerating that a bit; RV610 is fairly splendid if you look at it from the right perspective, and especially so from a financial perspective), then just wait until Barcelona. Then, and only then, will you really understand how much of a mistake this acquisition was. And I'm willing to stand by my words...

 

[_xxx_]

Correction: the current numbers would be VERY green. You are presuming things here with massively insufficient basis, as far as I can tell.

Look at Page 4: http://www.ccnmatthews.com/docs/ATI.financials.pdf
ATI's cash position in May 2006 was about $520M. This is a reduction of about $120M year-over-year, with a net income of about $100M in that same period. Let's summarize that by not considering the $25M of stock repurchase and say that with net income of $100M, they had a cash flow of -$100M.

Ok, since I do not have sufficient data: how much money was spent during the time between July last year and R600 release day? Just considering the R600 related costs. Several respins, dev resources, etc., all inclusive - how much do you estimate? Also think about marketing etc. costs (pulling numbers out of my arse I'd dumbly estimate at least several hundred mio. total, for next to nill earnings on R600 family).

If the business ain't profitable, you kill it or reorganize it (mostly by severely reducing the manpower and killing some projects).

Stupid/unbased guesses or not, my gut never betrayed me: IF there is R700, it might well be the last high-end card we'll ever see from the AMD camp. In a year or two, we'll see what happened.

EDIT: what does Barcelona have to do with the acquisition? I can't follow here. Or do you mean purely the financials?

 

[Anon Lamer]

The chipset problems only occurred with K8 release when VIA chipsets did not have a lockable AGP PCI bus I seem to recall.

Have you forgotten the infamous VIA pci bug in the 686b southbridge?

 

[Arun]

Ok, since I do not have sufficient data: how much money was spent during the time between July last year and R600 release day? Just considering the R600 related costs. Several respins, dev resources, etc., all inclusive - how much do you estimate? Also think about marketing etc. costs (pulling numbers out of my arse I'd dumbly estimate at least several hundred mio. total, for next to nill earnings on R600 family).

I don't know the exact numbers since it's sadly not broken up like this, never was and never will I guess. However, from May 2005 to May 2006, you had tons of tape-outs/respins too, and marketing/dev costs too! I'm not sure what's magical about the R600 period that makes it more expensive than any other in ATI's history by a significant margin. Those costs were always there, they're not new.

Stupid/unbased guesses or not, my gut never betrayed me: IF there is R700, it might well be the last high-end card we'll ever see from the AMD camp. In a year or two, we'll see what happened.

Well, you're speaking about what might happen now that ATI was acquired by AMD. Obviously, that might be different from what would have happened if ATI hadn't been acquired.

However, regardless of this, I agree that it would have happened anyway for financial reasons, as I have said in the past. The $399+ market isn't large enough to justify the design & tape-out costs going forward IMO, so dual-chip solutions make a lot of sense. The question then becomes whether ATI's 2/3-chips solutions are competitive with NVIDIA's 2/3-chips solutions. And that's obviously very hard to predict at this point.

So we'll move from up to 4 chips in a family to 3 chips. In fact, NVIDIA never moved to 4; they could have in the G7x era, but did a GX2 instead. In the G8x era, they didn't do a GX2 but created a monolithic high-end, which resulted in a large gap between the mid-end and the high-end. Obviously, the lessons of that should be quite clear.

Going forward, I believe we will move from 3 chips to 2 chips in a product family. But not by losing a chip at the top this time; instead, it will happen by losing a chip at the bottom. There are a wide variety of reasons (that I won't go into here) why it makes more sense to try to move IGPs up and kill the 64-bit discrete chips (G72/RV610/etc.) - including the fact that IGPs can actually be more profitable, partially because they don't need an extra PCB and RAM chips.

EDIT: what does Barcelona have to do with the acquisition? I can't follow here. Or do you mean purely the financials?

Yes, from a financial perspective. What I meant to do is justify why the acquisition will, in the end, have been very negative for ATI (in ways that perhaps couldn't have been predicted back then). The fact that the CPU division's bleeding has just begun will hurt ATI, and they'll be worse off than without the acquisition. Their position, relatively speaking, wouldn't have been so bad if they hadn't been acquired.

 

[Razor1]

Arun ATi wasn't in a great position before the merger, actaully during the beginning of the x1800 line they had around 250 mill in cash and 250 is liquidable assets, I remember this cause it was either Geo or Trini that corrected me, forgot about the short term assests lol.

So probably by the end of the 19x0 line they probably didn't have much cash reserves at all. Yeah they weren't as bad as AMD is now, but still the r600 could have pushed them over the line.

 

[Dave Baumann]

IIRC ATI had $600M is cash at the time of the merger.

 

[Geo]

They had very healthy cash reserves (in fact AMD used them to help finance the acquistion!), and quite frankly it's a rube's game to assume nothing would have been different if the acquistion hadn't happened. The relationship with Intel (and Intel partners) immediately cratering after announcement (and the financial impact that caused) is a very obvious example. I don't see any way possible that an independant ATI wouldn't have had a shot at an R7 generation.

 

[Arun]

IIRC ATI had $600M is cash at the time of the merger.

Yeah, that must about right I think. As I said above, ATI's last quarterly report was $520M of "cash + short-term investments". Also note that NVIDIA counts short-term investments under the same category as cash, presumably because that kind of thing is different between the USA and Canada). And all other forms of assets are useful to attract creditors, and ATI had plenty of those too. So I'm really not sure where people are getting the idea ATI had or would have had any liquidity problems.

 

[pelly]

Yeah, that must about right I think. As I said above, ATI's last quarterly report was $520M of "cash + short-term investments". Also note that NVIDIA counts short-term investments under the same category as cash, presumably because that kind of thing is different between the USA and Canada). And all other forms of assets are useful to attract creditors, and ATI had plenty of those too. So I'm really not sure where people are getting the idea ATI had or would have had any liquidity problems.

Honestly, I think people are just remembering the imapct NV30 had on NVIDIA (and certainly on Wall Street)...and applying the same logic towards R600. Since I'm confident that horse stopped moving a few months ago, I'll just point out that the two situations aren't apples to apples as the two companies are so drastically different.

Regardless, it certainly would have been interesting to see how things would have panned out without the acquisition. AMD aside, we still would have had ATI failing to regain the performance crown from NVIDIA. Add to that NVIDIA's gains in market share on both the desktop and mobile fronts....and it isn't a stretch to begin questioning how stable ATI would have been. We'll obviously never truly know....but I'm certainly miles away from believing ATI would have been significantly better off without the acquisition.