The AMD Execution Thread [2007 - 2017]

[cadaveca]

LoL. I was wondering when someone was going to single out Mr. Richard...


Seems to have lost his edge...lots of false hopes and promises, more the plays of a damaged ego, or someone scared. Seems he has more in common with Mr Ruiz that I intentionally thought.

I watched that press conference live...Henri also says back then that R600 was 64pipelines.

LoL. Before Mr. Richard ever said full product launch, I did. Dude's gotta stop reading forums and start working on fixing the issues at hand instead of burying them under the rug.


I wonder how much it would cost to force a takeover right now...

 

[3dilettante]

I'm trying to find some direct links, but stories about Barcelona being delayed into Autumn have been popping up.

This seems to have been precipitated by an earnings warning from Cray, concerning a shortage of parts for some products that are likely lacking Barcelona chips.

Allegedly, there is confirmation from sources amongst partners.


That leaves pretty much no time before Intel's 45nm ramp, unless Intel feels like taking its time.

Given the slow ramp up for previous AMD chips, that leaves essentially no time for Barcelona to exist on the market in significant quantity prior to 45nm competitors.
That means little time with ASPs being bouyed by the new product and a difficult road to Fusion (which may be marginalized by Intel as well).

That puts any future last-ditch financing in peril, as well as a lot of AMD's various platform initiatives (Torrenza, Puma, Fusion) that count on partners believing AMD can live up to its promises.

If this delay is true, a lot of partners that have taken a lot of crap recently are going to be pissed off.

AMD's credibility, while already shot, will be shot again--in the groin.

Cripes.

 

[nicolasb]

I'm trying to find some direct links, but stories about Barcelona being delayed into Autumn have been popping up.

A link:

http://www.theinquirer.net/default.aspx?article=40093

 

[Geo]

LoL. Before Mr. Richard ever said full product launch, I did. Dude's gotta stop reading forums and start working on fixing the issues at hand instead of burying them under the rug.

It would be either Orton or Bergman who would have fed Richard on that point in the first place, most likely. I think any time the former ATI execs make Hector, Dirk, or Richard look bad in public in such a fashion, they they burn some goodwill. I'm not one of those who think that things are at a real crisis point yet (and may never get there), but if events keep going downhill these kind of minor irritants get remembered when the really hard choices get made. Even in high-level business relations, personal relations and confidence or lack thereof in colleagues can make major differences.

I'd surely love to see the 2Q numbers show some significant improvement on the non-one-time front to validate their "perfect storm" story and get people breathing a little freer again.

 

[3dilettante]

http://www.siliconinvestor.com/readmsg.aspx?msgid=23600457

AMD says it is still on schedule, and the chip Cray is waiting for is not Barcelona, but Budapest.
That core is Barcelona with the updated HT links.

That means AMD's position has edged slightly from catastrophically bad to whatever it was prior to the news story.

 

[3dcgi]

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.

Speaking of GS performance why didn't we see coverage of the global illumination demo outside of hardspell? Was it not distributed to the press?

 

[overclocked_enthusiasm]

More smoke concerning Barcelona. Is there fire here or just rumors?

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3003

"Those hoping for nail biting, teeth clenching battles should apply elsewhere - the CPU war these days is a one horse race. If reports out of Taiwan are to be believed, initial performance results from AMD's Barcelona fail to impress and we've got at least a quarter before the race can even potentially get competitive. "

I think the clocks are lower than they would like...by alot...and the ease of the upgrade path will be hyped over the performance. This will leave AMD fighting to retain current AMD server placements and limit their ability to gain new share. Net neutral is best case for server market share if Barcelona is slower than expected. This will keep gross margins in the toilet and keep the price cuts high on the AMD agenda. This strat will bleed the company dry in 4-6 quarters IF Barcelona is a dog as the smoke would suggest. I think 65nm has been VERY unkind to AMD.

 

[overclocked_enthusiasm]

A link:

http://www.theinquirer.net/default.aspx?article=40093

http://biz.yahoo.com/ap/070604/cray_mover.html?.v=1
"Cray recently became aware that there has been a delay in volume parts availability for a key component of the quad-core XT4 until later in (the fourth quarter)," wrote Northland Securities analyst Chad Bennett in a note to investors. "We believe the most logical key component is the quad core processor from Advanced Micro Devices Inc."

Also, Henri "we don't do soft launches" Richard sold 19,597 shares yesterday....25% of his entire holdings.

http://biz.yahoo.com/ap/070604/advanced_micro_devices_evp_insider_transactions.html?.v=1

 

[overclocked_enthusiasm]

http://blogs.barrons.com/techtrader...ortedly-delayed-citi-trims-ests-stock-slides/
AMD: Barcelona Reportedly Delayed; Citi Trims Ests; Stock Slides
Posted by Eric Savitz
Citigroup’s Glen Yeung wrote in a research note this morning that “several sources†at the Computex 2007 show in Taiwan this week, including vendors, distributors and board makers, “have confirmed†that Advanced Micro Devices’ (AMD) Barcelona server processor, originally expected to ship June or July, now isn’t expected until September or October. Yeung writes that “yield problems†were most frequently cited as the reason for the delay.

Yeung says the Barcelona delay has negative gross margin implications for AMD. He cuts his EPS estimate for this year to a loss of $2.59 a share, from a loss of $2.48. Meanwhile, he also notes that a delay in Barcelona “opens the door wider for Intel (INTC) to consolidate its share.â€

AMD today is down 30 cents at $13.87; Intel is down 23 cents at $21.93.

 

[silent_guy]

Also, Henri "we don't do soft launches" Richard sold 19,597 shares yesterday....25% of his entire holdings.

http://biz.yahoo.com/ap/070604/advanced_micro_devices_evp_insider_transactions.html?.v=1

Don't pay too much attention to this 25%. It's basically meaningless.

The wealth of non-owner employees is usually not in the amount of stock holdings, but in the amount of stock options: You typically exercise options and sell them right away, because it's the safest way to avoid nasty tax surprises later on.
This can also be seen by looking at his number of trades over the last 2 years: he's used to be on a monthly selling schedule, then for the last 6 months, stopped selling but kept on exercising, building up a stock holding in the process.

 

[Geeforcer]

To perephrase... when "A" firm buys "A" firm, you get "A" firm, but when "B" firm buys a "B" firm, you get "C" firm.

While nothing seems to have gone smoothly in the last year or so for AMD, ATI, or their offspring, I am going to hold off on my Chicken Little impersonation for another month or so.

 

[dizietsma]

With Intel pushing back it's 45nm cpu's to the end of the year and 2008 for desktop this gave AMD a chance to make hay while the Sun shines, however this delay for them is an opportunity lost. It does look like yield issues if the top speed part at Computex is only 1.6GHz. They still need to gain 50% or so inspeed to get up to the 2.3GHz top end part on release.

This, coupled with the graphics card divisions delay and luke warm reception, are a couple of hammer blows it will take AMD a long time to get up from.

 

[Geo]

While nothing seems to have gone smoothly in the last year or so for AMD, ATI, or their offspring, I am going to hold off on my Chicken Little impersonation for another month or so.

Yeah, let's see 2Q results. Excluding one-times they were roughly in the neighborhood of -$425M for 1Q, and the exec team was coloring the quarter as "a perfect storm". So, reasonably, we should expect some significant improvement. Let's see if we get it. If we do, maybe "the doom is nigh" crowd will go back to sleep for a few months. If we don't, or heaven forbid it goes in the other direction even further, then it might be time to worry for real.

 

[overclocked_enthusiasm]

Th top end speed of Barcelona is rumored to be 1.6 GHz with a few posting at 2.0 GHz as the MAX. Sun has relegated AMD to low end like everyone else as well.

http://www.tgdaily.com/content/view/32349/135/

http://www.dailytech.com/article.aspx?newsid=7554

I know I am sounding like a broken record and that I hate AMD/ATI...I don't. I just am disgusted when management teams aren't honest with investors. Luckily I saw the wrtiting on the wall in December and sold all my shares despite management's reassurances at the December Analyst day. All those reassurances so far have turned out to be less than accurate to be kind.

 

[Arun]

Those benchmarks are slightly illogical, because as someone pointed out to me, that's actually worse in terms of IPC than K8! Even if there is a hardware bug right now that they'll fix or whatever, these are still hardly what I'd call impressive. And 2GHz, even if it's far from the final stepping perhaps (here's hoping for 2.4GHz+), is also very subpar (especially so for the dual-core variants if it affects that too...)

I've said for a long time now that the Barcelona Family will have no financial impact in 2007, however. And now that should have gotten even more obvious (although I was personally never expecting it before August or so anyway, heh). At this point, I'm even skeptical it'll have much of an impact on Q1 2008...

The only positive thing I've seen for AMD recently is this, because reducing 90nm capacity sooner than anyone had expected them to will improve their gross margins slightly.

 

[Pete]

Does this fear of low current/initial clockrates square with DT's Kubicki's blog quoting Kanter as saying, "Given the evidence I saw at ISSCC 2007, I'm pretty confident the chip [Barcelona] can make it to 2.8 GHz." Is his quote contradictory or complimentary?

Last two paras of that blog post, for context:

However, there is still a lot unsaid even this late in the game. While AMD's multi-core SPEC guidance claims the simulation runs on a 2.6 GHz Barcelona processor, guidance from the company as late as last month states the fastest debut K10 cores will top out at 2.3 GHz.

Kanter closes, "Given the evidence I saw at ISSCC 2007, I'm pretty confident the chip can make it to 2.8 GHz."

Edit: Dunno how this compares to the AMD benchmark in o_e's link, but apparently KK managed to run Cinebench on a quadcore Barcelona (813 @ 1.6GHz) and Xeon 3220 (1274 @ 2.4GHz). KK seems to be reporting "CB marks," which may be comparable to #s AT got for quad-core Core 2s (1313 @ 2.4GHz) and Penryns (1935 @ 3.33GHz). I'm not familiar with Cinebench scores (the various subscores themselves and what effect platform/bandwidth has on them), so I may be comparing the wrong #s. But that hasn't stopped me from calculating CB Marks per GHz: Barcelona 508, Xeon 531, C2Q 547, Penryn 581.

Thanks, 3dilettante, that's what I figured, but I wanted to be sure.

 

[3dilettante]

Kanter clarified elsewhere he believed Barcelona would reach 2.8 at some point in its lifetime, not necessarily at launch.

 

[cadaveca]

It would be either Orton or Bergman who would have fed Richard on that point in the first place, most likely. I think any time the former ATI execs make Hector, Dirk, or Richard look bad in public in such a fashion, they they burn some goodwill. I'm not one of those who think that things are at a real crisis point yet (and may never get there), but if events keep going downhill these kind of minor irritants get remembered when the really hard choices get made. Even in high-level business relations, personal relations and confidence or lack thereof in colleagues can make major differences.

I'd surely love to see the 2Q numbers show some significant improvement on the non-one-time front to validate their "perfect storm" story and get people breathing a little freer again.

I totally agree. The problem at hand, from my perspective, is that if the top guys leave themselves to hang out and dry so easily(that whole press conference made Mr Richard look a little wet behind the ears..unable to contorl emotions, etc...), then the deserve to take the flak for the comments they personally made to the press, without passing the "poop-covered buck".

The fact that such seems to have actually happened(Mr Richard looking to take some revenge on co-workers, I have ears very close to the ground), it's only a matter of time before such behaviour starts falling thourgh the ranks, and this is NOT what a newly-merged needs right now...they are all on the same side, after all, aren't they?

I see two very seperate entities with the AMD conglomerate now, and we are not talknig of AMD/ATI...we're talking those that are happy with how things are, and those with a very big axe to grind. Unfortunately the only sharpening stone is the conglomerate, and after a time, every stone must be replaced, as it wears away.


Don't get me wrong though...I do not want to see changes within the staff @ AMD...I want to see changes FROM the staff.

Too bad some of that good staff, they let go.

 

[overclocked_enthusiasm]

http://www.legitreviews.com/article/521/1/

Penryn up and running at 3.0 GHZ and 3.33 GHz today. Barcelona at 1.6 GHz??? Umm...this is going to be a bloodbath.

"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."

He might be right afterall about the "killer" reference.

 

[dizietsma]

http://www.legitreviews.com/article/521/1/

Penryn up and running at 3.0 GHZ and 3.33 GHz today. Barcelona at 1.6 GHz??? Umm...this is going to be a bloodbath.

"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."

He might be right afterall about the "killer" reference.

Thanks for the link. After a few minutes with my Windows calculator both the single and dual Penryns show about 8% improvement over the current QX6800 chip when equaliszed for similar clockspeeds. This sort of tally's with past 5-15% improvement with the new chips over the old.

I guess it all comes down to how fast AMD can get these chips up to. In the past they have had problems with K7 etc when first produced getting the clocks so this might be the same case and they can overcome that. I do think though that by the time Intel release Penryn for the desktop they could be able to put out a 11x333 part or 3.66GHz the AMD part will have to compete with.....

The worrying thing though for AMD is that even if they beat Penryn this year with Intel currently being on it's reduce process and tweak year then AMD only have one year before Intel's main architecture change comes in again with the Nehalem chip. IF this does as well as Conroe did then this must be a major worry because typically AMD have changed a processor design every 3-4 years or so.

My current rig for gaming is a single cored FX-55 and I really need to go up to at least 2 cores say by early 2008. With Intel I already know I can get an X38 motherboard and a Penryn or it's dual core brother at about 3GHz, with AMD I will just have to wait and see. They need to be more forethcoming or else my mindset will have already have written it off unless it is stunning. If you see what I mean ?