The AMD Execution Thread [2007 - 2017]

[liolio]

Sorry I've not read the whole thread and my points may be stupid or already adressed.

My point is AMD has or had a chance?

Intel has so much more money, they can work on different projects in parallel, while AMD is much more limited.
What we ask to AMD is commitment to excellence ie they can't afford one miss step.
But in the real world... every companies do mistakes.

More even without mistake, my opinion is AMD can no longer compete against Intel due to the process advantage of the later. Didn't Intel demo some chip (terrascale) @32nm, they delaying their complete shift to 45 nm because they can't afford to crush AMD (anti trust law could hurt them more than AMD will ever).

I feel like the situation is more like a "reality check" for AMD, I feel like Intel is in a situation were it can hurt a lot of hardware vendors AMD, SUN, IBM to name them.
Intel works on some many projects, "standard SMP cpu", larrabee, terrascale project, who knows what else, it's the same in regard of process Intel is in a situation where it can cram more transitors than every can and it doesn't look like thing will change in near future.

AMD/ATI has done some mistakes, it' a fact, but Intel has been impressive on its side and now everybody is fully aware of the kind giant the whole industry is facing.


What can do AMD? I don't know I think they should be more focused and admitted that for quiet some time facing Intel (alone)is no longer in their range.
Overall they should focus on delivering more focused product than what Intel can in near future.
AMD/ATI should push their GPU.
they should try to get in the console market.
If they can't compete with Intel, I think they can compete with Ageia, clearspeed (that they didn't manage to buy it) or IBM with the cell and toshiba with the spursengine.
AMD in on par with IBM and toshiba as far as proces are concerned and has a clearly advantage against the formers.

Anyway I don't know much, but I feel like AMD shouldn't try to go any further than four "traditional cores" on one chip and focused on market thatn need high perf per watt:
supercomputing, consoles, low end computers (tougher as Intel is their too).

But I don't feel like AMD is in a situation where they will recover quickly. They should focus on making on money on more tiny markets share or they will die.

 

[INKster]

More bad news inbound:

Big Losses Projected At Struggling AMD


SAN FRANCISCO (Dow Jones) -- Advanced Micro Devices Inc. will report fourth- quarter results next week with Wall Street projecting bigger losses at the struggling chip giant.


Analysts expect AMD (AMD) to report a loss of 36 cents on revenue of $1.79 billion, compared with a loss of 4 cents, on revenue of 1.77 billion in the year-ago period, according to Thomson Financial.

The Sunnyvale, Calif.-based company has been struggling to become profitable over the past year as it faced stiffer competition from archrival Intel Corp. ( INTC) and grappled with the cost of a major acquisition.

AMD reported a wider-than-expected third-quarter loss of $396 million in October, largely due to a $120 million charge for its merger with ATI Technologies in 2006. The company has also struggled with product roll-out delays and mounting operating costs that some analysts warned were not sustainable.

"We are lowering our estimates to reflect continued execution missteps, an increase in operating expenses due to engineering costs and continued pressure from Intel's roadmap," analyst Doug Freedman of American Technology Research said in a research note.

Analysts believe AMD could bounce back as a stronger competitor, especially with the graphics technologies it now has after the ATI acquisition. The company also got a boost recently from an Abu Dhabi firm, which invested $622 million in AMD.

AMD executives vowed last month that his company will be back in the black in 2008, and that recent production missteps will not be repeated.

"We blew it, and we're very humbled by it," CEO Hector Ruiz told analysts in December. "We're going to learn from it, and we're not going to do that again."

However, analysts say concerns about weakening demand in North America make it harder for AMD to stage a comeback.

"For AMD, the issue is how much did they get to participate in what is a comparatively weak North America, compared to the footprint they have all around the world," said analyst Crawford Del Prete of International Data Corp.

AMD tends to do quite well in emerging markets, he said, but it's not unclear if that was enough for the company in the fourth quarter.

"It's still pretty tough times for AMD right now," he added.

http://money.cnn.com/news/newsfeeds/articles/djhighlights/200801110425DOWJONESDJONLINE000421.htm

 

[liolio]

Anyway I don't know much, but I feel like AMD shouldn't try to go any further than four "traditional cores" on one chip and focused on market thatn need high perf per watt:
supercomputing, consoles, low end computers (tougher as Intel is their too).

I want to make this point clearer Ati would have better come with 1or2 cores+gpu than a 4 cores and try to face Intel on the enthousiast and servers markets.

They may have done good deal with people like dell and HP for mini PC and why not steal from Intel the deal for mini MAC.

 

[hoho]

Going back over Charlie's story, he's indicating B3 is the last stepping.

That also tells a few things about the future. AMD has said it will have 2.6GHz K10 with 140W TDP. As B3 will be the last stepping there probably won't be any higher clocked CPUs released before 45nm. As was said, 45nm is available some time in 2009. That would put 2.6GHz K10 against Nehalems coming later this year. Not a nice thing for AMD.

As for dualcores, I doubt AMD really wants to produce them. K8 is not that much slower than K10 but costs a lot less to make and likely can be clocked as high or even higher. I'd say 65nm K10 will be pretty much only quads and recycled triple core CPUs until 45nm is available.

 

[Albuquerque]

This just makes me even more depressed.

Even though I was originally cynical of the 38x0 series, I can now say without a doubt that I'm a true believer. If AMD's CPU business continues sinking, then the graphics parts are going to go right along with them.

 

[dizietsma]

I think results are out today. Last two quarters were $600, and then $400m loss so considering the Christmas period and better results from the graphics devision I am going to guess by keeping the series going and saying a $200m loss. It coul have been better and around -$100m I think if Phenom had shipped in bigger quanities.

 

[ShaidarHaran]

I'll throw my quarter in the hat for a $300-$400M loss.

 

[3dilettante]

This may break a few hearts...

http://www.semiconductor.net/article/CA6512566.html?industryid=47298&nid=3572

ZRAM is dead, long live ZRAM II.
The charge carrying capacity for the technology AMD wanted to use for its caches was apparently insufficient and scaled poorly.

Second-gen ZRAM is apparently a reprise of the bipolar transistor, or will be if it comes to fruition.
The description leaves me kind of skeptical, as it seems timing and a number of other factors need to be juggled. It sounds like a fragile solution.

It's also using a bipolar transistor, which has functionality that requires current flow at all times, unlike CMOS that only has flow during transitions.
On the other hand, at small geometries and high frequencies, everything leaks anyway. ZRAM would have to be refreshed, so power consumption might be a wash.

This does put off AMD's using ZRAM for a while, if ever.

 

[ShaidarHaran]

Interesting. I sort of assumed that Shanghai would be using ZRAM simply because of the massive increase in cache size over Barcelona. I wonder if Shanghai's supposed delay has anything to do with this. Perhaps that was the original plan and since it didn't work out, they now have to redesign Shanghai with run-of-the-mill SRAM.

 

[nAo]

This may break a few hearts...

Poor Arun..

 

[MTd2]

This does put off AMD's using ZRAM for a while, if ever.

The article says that AMD licensed both technologies, so I can't understand why you say this.

 

[3dilettante]

The second-generation ZRAM product has only been recently disclosed, and no working samples have been demonstrated.
No characteristics were given and no timelines are involved.

If we look at the lag time expected for when 1st gen ZRAM was to appear in AMD's products, we wouldn't expect a drop-in replacement with ZRAM II for Shanghai.

There's a feedback comment for the article that makes some points in engineering-speak.
I am not an engineer, but some of the points regarding the lack of details concerning what can be done with a bipolar device that lacks a current source seem vaguely plausible to me.

Unless scaling and voltage concerns are addressed, AMD is in danger of comitting to a technology that might be only marginally good for one process node (if it works at all), and I'm sure AMD's engineers are aware of this.

The fact that the presentation was challenged by engineers who also work with floating-body memory is interesting.

It's vaguely reminiscent of the entire pre-release cloud of confusion strategy AMD used for Barcelona's performance numbers, and we all know that ended.

 

[Geo]

I think results are out today. Last two quarters were $600, and then $400m loss so considering the Christmas period and better results from the graphics devision I am going to guess by keeping the series going and saying a $200m loss. It coul have been better and around -$100m I think if Phenom had shipped in bigger quanities.

I'll throw my quarter in the hat for a $300-$400M loss.

On top of the Phenom delay, wasn't there a filing that said they were going to accelerate/increase the ATI acquistion writeoff for "goodwill"? If that's true (i.e. get rid of what's left all at once), then I think I have to go with Shaidar's range and hope it isn't even worse than that. . .north of $500B to the bad (on paper, because of the writeoff) wouldn't shock me.

 

[2senile]

http://biz.yahoo.com/bw/080117/20080117006120.html?.v=1

I don't understand such things but it doesn't seem too bad.

 

[Arun]

It's actually damn good as the non-GAAP operating loss was only $9M... I'm waiting on the detailed PDF for some info including cash position etc. before I news this though (unless somebody else on staff beats me to it due to timezones!)

 

[Geo]

Operationally, that's much closer to break even than I was expecting. . . and yowza, that write-off is a whopper. Looks very much like the "take out the trash" strategy I suggested upstream.

This may very well position them to return to profitability by 2Q.

 

[MTd2]

http://biz.yahoo.com/bw/080117/20080117006120.html?.v=1

I don't understand such things but it doesn't seem too bad.

I can't understand very well, but it seems that things went really well for AMD, they finaly (almost) went out of the financial hell, if you discount their debts with ATI.

BUT, including the payments for ATI, they bled more than 1.6 Billion USD. Yes, that's 2x times worse than all the disasterous quarter combined, or 8 times worse than last quarter. The good news it is that they might have fully acquired ATI now, having payed 3.2 B in total.

EDIT.: Now I can sleep well. AMD out of the financial hell means cancer and spinal cord damage will be cured! WOOT!

 

[Bouncing Zabaglione Bros.]

So my question is: how did they do it? Or is it just financial smoke and mirrors?

 

[AlNom]

Just a thought, sales from HP (contributing)? They seem to be overwhelmingly AMD in their laptops and desktops (at least from what I can tell of Futureshop/Best Buy).

 

[Arun]

So my question is: how did they do it? Or is it just financial smoke and mirrors?

Seasonality, handheld recovering, and this:

“[...] We improved gross margin by three points sequentially, driven by increased shipments of new products, higher average selling prices and cost containment actions,” said Robert J. Rivet, AMD’s Chief Financial Officer. “We shipped a record number of microprocessor units in the quarter, including nearly four hundred thousand quad-core processors.”

So, Barcelona *is* actually shipping in very decent volume. Assuming their ASPs are $250 there, and it's probably higher than that, we're talking about revenue of $100M+! That's the difference between a bad quarter and a good quarter, really. Margins were also good; that was partially driven by seasonality (higher volume->higher fab utilization->higher margins) and apparently... Barcelona. Hmm!