How AMD turned Barcelona into a right royal mess

[Reader warning, another slab is below. Any chance of carriage returns? Ed.]

One thing AMD could do in the short term to make some badly needed cash and patch things up with users (like me) is to release updated K7 chips for 939 motherboards. I'm not really up for upgrading my whole mobo/ram/cpu at the moment (and if I did it would be an intel based kit) but I would be interested in a nice speed bump in the form of a K7/939 CPU if the price were right. How many 939-based mobos are out there in circulation? Quite a few I'm sure. I bet at least 10% of their owners are in the same situation. Do the math AMD. Sell me the product I want, keep me happy, and make some money.

Great article from a paid Intel hack. Is this the same gentleman who worked at Intel? 

Why is that not disclosed?

This is getting hard to watch. AMD is like a crime scene

Don't lie...err on the side of caution. Over promising is the same as a lie. People will forgive honest mistakes but not deception. Be honest, take your lumps and you'll get customers back.

Charlie,

As an avid fan of yours, this reader, as you know, has been skeptical of AMD’s outrageous and unsubstantiated claims since the beginning of the year. I’ve said so, privately and publicly to you. Further, as you also, know my interest is far from altruistic. I am investor first and an enthusiast second. I have done quite well with the other guy. You know this, too.

Perhaps this letter can be misconstrued as an “I Told You So”. It is not. Upon reading your piece I came away very angry. Why? You have been more than fair and patient with these desperate lying bastards. AMD fans bought the spin and hype, Wall Street bought the spin and hype, and private investors (some of my friends) lost a lot of goddamned money buying the spin and hype. I warned them, too. 

More than this, I suspected the day would come when you would write something of this nature. And, that day has come. You are far more tolerant than I. Had I been in your position, or rather, if I were in your position, I would have raked these sons of bitches over the coals.

More over, and true to form, your professionalism shines through, as there is obvious disappointment in your tone, but none of it is personal, KUDOS. 

As for me, they will never get a free pass on this one. Since the ATI buyout, AMD and its respective minions, screwed (stuffed) the Channel, screwed their partners (939), screwed their loyal fan base, screwed investors, basically, everyone they came in contact with. 

With this piece, I now see they have screwed themselves, again. But then, I knew this all along.

Your Fan,

SPARKS 


Maybe if AMD didn't spend so much time spitting on intel, and crowing about their own greatness, then maybe those "People With An Agenda" wouldn't be so quick to call a hypocrit a hypocrit.

Nice article Charlie,
I think we get it, Its ok to rip into Intel when they have problems but if you have anything bad to say about underdog AMD you suddenly become a troll. Well I havn't been under any bridges lately so i think i can certainly discount that one. I think this launch is somewhat unique in how bad it has gone for AMD. I believe this is why so many people have something to say about it. The K7 launch went rather well in my opinion there was no real erata in the processor the only problem I remember at the time was the buggy irongate chipset of which i was an recipient. Other than that there was high demand for the K7 and I was still able to get one with no problem. This was the second AMD chip I had bought and the third non-Intel if you count the Cyrix-IBM 686 The first AMD chip being the 386 DX40. i've had all AMD chips after the K7 except for now that I have the ASUS G2 with the core 2 duo. So i think I can speak with a little experience when it comes to Amd chippery and not as a troll as you would call us. I believe the press should stop supporting companys and just give us the news as it happens. The fact is, this Barcelona and Ati's latest offerings are dogs with fleas and there is no escaping that. Especially after Amd touted all the goodness of the "native quad core" and how we needed it so badly rather than have Intels approach. The same goes for 64 bitness and how we all desperately needed that too. I don't know about you but 64 bit just isn't too important to me yet. Talk about blue crystals! And heres another revalation for you, I use "Prescott' computers at work all the time and they don't seem defective to me in any way.

Msr. Ballmer Stated Late Summer that NO true 4 Core Exist Nor Will it in '7. Nov was Whole Eight days Ago, yet when barcelona is done, finally crafting of "7 "new commands anticapated as doubling or square of l1 cache instructions, 7 multi core needs will start to be anticapated, So therefore ULTIMATE Will Just Flow into that quality.Its big sstep to change so much & add tremendous amount of instructions as well. Multi level problems that need multiple number of fixes.
thomas s von drashek

"From this point on, what needs to be done? For AMD it is quite simple: shut up and deliver."

Gaaaaa! I am blinded by a flash of
obviousness.

Good gob Charlie...

Straight forward and to the point... what comes from bad launches is simply bad press ... AMD simply needs to execute, proper execution avoids these type fiascos.

please just stop. dont point your finger at those fanning the flames after the fact, when you did more than your share of fueling the flames before.

biggest issue is you did not simply put your rumors out as rumors, you tried to spin them as special first hand knowledge. to make matters worse you replied to any doubters like a pure ass. 

how about instead of telling people to get over it, you eat crow.

Well my 4800+ (939) is over two years old. Time is running out AMD and it aint looking good

The Intel guy is right: their tock is great, meaning they've got their architecture right. If they ramp up the thing this is over.

For that reason I say: look at inverted head-and-shoulders and buy that AMD stock.

Well, 

I'm definitely an AMD supporter, since I'm using all their processors since AMD DX4 100 Mhz :) ... which was ... I think ... my first one :).

Usually the articles on this site are more "spicy" related to AMD and I'm glad (or sad) to say that this one is a very well written article.

And I definitely agree with your last words: shut up and deliver... because I had high hopes for Barcelona ... and even if the architecture is outdated, well ... it would have been nice to be able to deliver ... something :).

Let's hope they get back on their track ... soon enough.

"Certain companies are sending out baiting mails to second tier gullible journalists"

As opposed to a certain other company that pumped a first tier gullible journalist with bad info not too long ago, isn't that so Mr. 'Dancing in the Aisles' Demerjian?

"Credibility is easily chucked away and very hard to regain."

Indeed.

Hector needs to step down or be removed immediately. 

Intel needs competition and AMD is not just a joke but it's also seems like management would make a mess out of a "McDonald's".

Literally, they are running this company into the ground and the result is very ugly. 

I've bought AMD and supported them throught tough times. 

I still love my opterons, which run great but this embarrassment must end... Hire away some people from Intel and put these pathetic failures working at AMD now out of their misery. 

There is just no excuse for what has happened. NONE!

As someone that likes performance at low prices ($195 for a 2.66ghz C2D) think I have to hope.

Sure, they are sucking right now on the CPU side, but they hit a home run with their revamped R600 gpu. 

Although my next system does have a C2D in it, it also has an ATI GPU. hopefully whatever little money they get out of that will help them get through this tough time.

I'll raise my glass and say "here's hoping" that we see another good processor like the barton go toe to toe with intel's finest at a lower price.

Seemingly the world is against AMD right now because of some failed parts. As we keep getting closer to the size of a single silicon atom the going is going to be tougher.

Unfortunatly AMD can't do anything right for most people right now. For instance if they would have held off until say June '08 to put out a flawless spider platform people would be crying the blues that they're taking too long and they are going to die. 

Alternately they put out a somewhat troubled version right now and the same people are PO'd that it's buggy and not beating Intel. So how can they win in the mind of buyers? Some may say that they should have put out a flawless piece 6 months ago that was faster than anything Intel has to date. That's simply not an option. 

Intel is back, and with a grudge. They are going to overcome AMD in every way they can. 

Even if AMD would have been readily engineering Spider back when the P4 was out and released it one year ago, Intel would still over take them. Intel as 30 times the market share, ~100 times the R&D budget, and no debt. 

Complaining about AMD's woes is no worse than wondering why Sierra Leone's armies got destroyed by the USA (in a hypothetical war).

I just hope they keep up the fight and keep Intel honest.

AMD could take a lesson from politicians - the coverup is almost always worse than the initial crime. A simple mea culpa would have alleviated much of the press and AMD is now risking their reputation with the deception / creative interpretation, lies of omission and spin.

The K10 product can be fixed with future steppings and more importantly improvements in the Si process technology. The reputation will be a lot harder to fix if lost.

Various fans and advocates can disagree over the performance of the product and the projection of where it will be, but it is hard to argue anything other than AMD has really handled this whole project miserably from the start in terms of PR. 

From the is it called K8l or K10 minor minutia, to the "mid-year" launch (with no specific timeframe), to the 40% better claims, to the 3.0GHz demo mid year, to the pushout of the June analyst meeting, to the "focus on energy efficient products" at Barcy launch per customer demand(my personal favorite), to the announcement of Henri Richards leaving days before the "biggest launch of 2007", to the K10 production is "as expected", to the Tahoe launch fiasco and lack of review samples out to the key review sites, to the pushout of the Nov07 analyst meeting at the last minute (nothing to do with K10 launch, mind you), and to the latest TLB bug.

Only a few of these actually have anything to do with actual product performance, but the continued 'stretching' of the facts leads me to...to borrow a political phrase.... "AMD fatigue"...it would be nice if AMD could speak plainly and tell the public and their potential customers how things are.

If / when AMD claims that bulldozer wil be 40% better than current products will we laugh and smile or take them seriously? If Intel claims that Nehalem will be similar to the jump from P4 to Core2 will we laugh or take them seriously?

Products come and go, reputation does not.

Indeed, it is getting amusing, charlie apparently from these comments works for intel AND amd.

And he starts again with the 'words of wisdom' and already starts on the 'they have a kick-ass product out soon, in 2010'

It's clear that the inq pushed the promises of AMD, because they bought into it, and did so more than once, being burned again and again, so perhaps it's true for the inq too, shut up and report on the REAL stuff from AMD not promises for the freaking year 2300, same as AMD should stop feeding promises for a while so too the inq should stop eating them.
There's a diference between reporting 'AMD says they hope to have a speedy part' 
and 
'AMD IS about to release a part that will make intel very worried'

(This of course is not inq-exclusive, many sites do it)

[Rubbish. But have your say. Ed.]

I agree with Sparks, very professional article, well written. AMD is sound technically and ATI is, if I may say, hitting-on-all-cylinders. Personally, I'm no AMD fan-boy, and I don't care for Intel [it's an architectural thing].

So, I suppose, if I would've wrote the article I may have focused a bit more on AMD's management. How inept can they get? It would be surprising if AMD didn't wind up being a case study at every business school -- i.e. how to augur your company straight, perpen-damn-dicular into the ground, like a Chinese mine-shaft.

Hey, Charlie, any news on getting rid of that ruinous rat-face Hector Ruiz?
[Captain-of-the-ship you know, and all that.]


Absolutely fantastic article. I always read the news on the Inq. DAILY. This article reaffirms my supporting of AMD. To be honest at this point if AMD makes another bad slip I am not sure if they will be able to rebuild the trust they lost until some point far in the future. I bought into the 939 then the AM2 now I recently bought the entire SPIDER platform. If AMD falls I don't want to pay the outrageous prices that Intel WILL charge when they become a monopoly.

[Any chance you could you use a carriage return, ever Robert? It's like a slab. Ed.]

AMD urgently needs to get its act together, in everyone's interests. We cannot afford to have Intel dominate the market without competition.

Check how the Conroe quad prices scale some time, small increases in performance for massive increases in price. AMD cannot compete with Intel in performance and Intel knows it, it is reflecting in the prices.

Either something is done or the £1000 processor may be popping up some time soon.

How much of this is from AMD dorking around with SOI substrates? It wouldn't seem that AMD's chip designers are that much different from Intel's. 

Intel's shown a commitment to building on epi. AMD's gone to SOI. To be sure, SOI is much more glamorous and elegant. It makes for sexier press releases. But, is it 'not quite ready for prime time' ?

TPW

I have been a dedicated amd fanboi my entire life, my first pc was a intel pentium1 100mhz, that was the only intel part I ever owned, as soon as I had my own money I purchased amd stuff, and was very happy with it, my first part was a duron 800mhz, then a 3500 single core K8, then numerous dual cores, and then a 65watt 5200 1mb cache.

news had come after purchasing the 5200, that barcelona, was on its way, and I thought this 5200 will be fine until it gets here, well after many push backs, and speculation I was getting disheartend, every bone in my body wanted amd to come through, well on the day they were released I was like YES a finally faster amd part. that was until the dreaded errata and poor benchmarks. I was gutted, to say it mildly. I want amd to do well, its keeping the cpu industry fair and competitive, so I thank AMD for that. but after seeing the 2.3ghz quad do poorly, that was the last straw for me, I went to my Etailer and proceeded to order some new components, new seagate drives 32mb ones (fantastic drives BTW) some 1066mhz ram, and then I got to the CPU page.
my palms went sweaty, my heart raced, as if I was about to commit the biggest sin of all, I was about to order a Intel CPU!.
it was really funny, at the time I was about to click buy on that lovely 2.4ghz quad core GO SLACR, My entire body locked up, as if it was telling me DONT DO IT, I could here Hector pleading with me for another chance, 
after staring at the page for 10 minutes quietly rationalizing why I should not buy it, I clicked the buy button, and it was done!

the next day I recieved a email saying my parts were ready, and even at that point my mind kept saying there is still time to back out of it, and again the picture of Hector on his knees saying please Stew just one more chance, this went on for 3 days, I couldnt bring my self to go and pick up the new parts, I just couldnt do it.
on the 4th day after reading more reviews I snapped out of it, and went and got my parts.

It was such a struggle for me to get a intel cpu. I have loved amd since I knew whata processor was.
The point we all should remember is AMD is responsible for a lot of things 
1.competition
2.price of all CPU's
3.Pushing Intel to develop better cpus' vice versa

If amd didnt exist there would be a monopoly, and we would still have P4's with netbust architecture.
really everone should buy an amd cpu, even if you are a intel fanboi, becasue if amd goes broke, so will you with inflated cpu prices, and slow develoopment 

Thanks for reading

Charlie,
This was a good, unbiased analysis of the situation and the first comment is that you're an Intel hack? WTF? Ignore ron. He's having one of his "not so fresh" days.

AMD is like the guy in the vaudevillian performer that tries to spin plates on sticks. AMD were OK when they just had one or two plates to spin. Now with the ATI merger they have four or five plates that they need to keep spinning and they've shown they can't do it. Unfortunately for them, Intel can keep about 10 plates spinning and are more plates (digital health, UMPC, ...) It is damn hard to do what Intel has done for years and AMD is finding out just how hard.

Is anyone else sick of reading the same story over and over?

The arc of AMD's existence since Sanders left seems like an object lesson in humility. As a culture, the AMD response to finally not just catching up to Intel with the Athlon, but actually surpassing them, and then wanting to gloat about that fact, was entirely understandable. The achievement of technological parity had been a long time coming, and I don't think many people would begrudge AMD their glee, especially not us end users who benefitted from the price corrections brought about by true competition.

However, where AMD failed was by treating that achievement as an end unto itself, rather than as the beginning of REAL competition with Intel. Intel quite obviously adjusted to the new market condition, and their current corporate structure and product line reflect that. Also just as obviously, AMD had become so lost in a bubble of self-satisfaction that they were unable to quickly and adequately respond to Intel's adjustments.

Which puts us where we are now, with an AMD/Intel situation reminiscent of many a day gone by. Sanders' decisions while he was running AMD weren't infallible by any means, but we at least knew he was always trying to bring the fight, and he carried an air of doing the best he could to achieve that end, which made it easy to believe in the company. To me, Ruiz doesn't project those qualities, and that worries me. Maybe it's just a difference of styles, and Ruiz is doing everything that anyone could reasonably expect, I really don't know. All I have to go on is AMD's performance since he's been running things.

Which means that while I don't doubt Ruiz' intelligence, at this point, I have to question his competence to lead an industry giant. Whatever the reason that AMD squandered it's lead over Intel, the fact remains that the lead WAS squandered, and it happened during Ruiz' watch. Even if blaming Ruiz for this is unfair, even if Intel is still too much a monstrosity for AMD to keep up with, the fact remains that AMD is currently bleeding money, and that is not sustainable. If Ruiz, or someone else at AMD, cannot stem this bleeding, then the future looks pretty grim. I remember what Intel's pricing structure looked like before they had a legitimate competitor, and I don't want to see that again.

So if this is a major problem for AMD and system builders like myself would not pay £180 for a Phenom with this problem. Especially at that price then they might as well drop the price until the new 9650 revisions etc.

I am glad I didn't buy a Phenom and board, memory etc I decided to upgrade some parts on my current 939 rig.

Oh my god, it's Charlie the moron again.

Mike please fire that nitwit, he's mostly so of track and biassed that one would believe fairy tales are for real.

Idiot.

Look, Charlie is just another talking head making a living reporting on stuff for which he has not a clue. 

Like many a time he has done the same to nVidia, ATI, and Intel. He pi**es in the shoes of hundreds of hard working engineers who are putting their heart and soul into these projects. 

Many of these dedicated people may loose their jobs because Charlie makes a living dramatizing and belittling their work. If he had a clue of how complex and difficult it is to deliver one of these projects, he might just find something else to **** on.

It's all well and good to blame AMD's enemies for keeping this alive, but really, the fault lies with AMD itself.

If AMD had ever come out and been honest about the problems they were having, rather than ever-spinning that things were "on track" , there would not be this slow drip-drip-drip of the truth coming out, more being discovered with each passing day and week and month.

When has AMD ever publicly apologized for their claims of a 40% rout of Clovertown, the "3GHz" Phenom demo (now expected to be 2.6GHz @ 140W in March 08), the "launch" of Barcelona in September, fooling review sites into not applying a performance-killing-but-necessary fix before benchinmg Phenom, etc, etc?

In fact, why is AMD not RECALLING their Phenom parts? Wouldn't that be better for their reputation, and their sales in the long term? Is anyone going to buy a 9500 or 9600 now that the news is out, and replacement 9550 and 9650s are planned for Q1?


I'm not suprised (in any event) when the chips don't work. Infact, from a completely non-technical standpoint, I understand (proud owner of both P4 Northwood & various X2 chips up to this good-ol 6000+), sometimes they just don't work. I had serious problems on the first-gen AM2 chipsets (more-so considered hybrids of mostly 939 parts), and I've had my share of problems with the P4 (likened but worse, overall, than the speed-ceiling on my late X2).. As it would seem, - troubles, however comparable, are here to stay. But at the end of the day, so are the suprises! 

PS. You won't be finding any specific CPU badges on my systems as - I've always bought our parts for the price-performance ratio over anything else..

It's unfortunate that AMD botched this one. I can't say they rested on their laurels with Opteron and X2 because i don't work for them. Was a native quad core a good idea? Hum... Maybe, but not at 65nm. Why didn't they "slap" two brisbane cores on a MCM and call it a day. It was s 65 nm part with a small thermal output. Didn't anyone at AMD think that could have worked? I bet two of those X2's would have scaled nicely at 65nm on a MCM. The bottom line is that AMD needs to succeed. Intel needs competition. If the Barcelona was great and had parity with Intel's top gun we would be paying less than a grand for the best. These companies need each other. Well that may be a stretch. We the enthusiast need these companies to be successful.

It appears that the failure leads back to management. Hector and company need to be fired.

Well, it seems like the Quad-core Penryns also are suffering from some problems - delaying the availability until march.

theintelquirer.net

just keep hector where he is so we can see amd's stocks slides down like my tissue as I wipe my ass...hahahaha

From the beginning of K8 754/939 and even 940, AMD screwed up big time there, with killing 754 and 939 way too early, ....

Okies, that is old story, moving to DDR2 is a must, no choice , lets move on, right till, AM2, okies, it live about 1 year or so, AM2+ is here, woot, 4 cores, and so on, AM2 is a bit useless now, okies, there is AM3 coming, and even Socket F? Socket F+?

Meantime, Intel Socket T is too good, and live long time, 

Let move forward to 2008, AMD has two competitors, Intel and nVidia, both are beast, while AMD did "fix" ATI with recent 3870 and later January? 3870 X2, they did not "fix" their processors business, they no longer has the arch. advantage (Intel closing the gap) and price advantages.

More over, SB600 is live way too long with is limited lanes and SATA ports, hopefully the DX10 IGP 780G make it on time,

So right now, AMD need to hang on, while Intel and nVidia is laughing that they only need to compete with them self, looks at C2D and 8800 series, in about 1 year now, almost the same thing.

Anyone remember Charlies article at the end of 2006 concerning what to expect in 2007 ?
Back then Intel was screwing up big time and AMD could seemingly do no wrong.
It seems that Mr.D was correct that AMDs future was theirs to louse up and they did.
Acquiring ATI ,Contracting with DELL and trying to close one Fab and simultaneously open a retooled Fab was just to much managerial work for a bunch of engineers.
Several articles at fabtech.com support that conclusion.
Couple the opening of and ramping schedule of a new Fab with development of a true 4core CPU and the results do not look so terrible for the engineers ,
the marketing people cannot control themselves and they are making promises that are consistently false.
2008 will be a very interesting year.

Whatever happened to all those internal Intel emails that AMD lawyers wanted ?
Intel loose them ?
How and when will the EU drop the hammer on Intel for monopolistic charges ?
2007 was expensive for AMD.
2008 is shaping up to be Very expensive for Intel.

You haven't really explained how AMD got this wrong and why so many mistakes were made.

You haven't discussed the implications of the chip being released in such a buggy state - does this relfect a state of utter desperation,does AMD fear for its existence?

You haven't touched on the really rather conservative and disappointing underlying Barcelona architecure - even at sensible clock speeds barcelona would be no great shakes.

And most of all you haven't provided any sensible suggestions for how AMD moves forward other than the utterly obvious must-do-better.

I still don't really know how AMD managed to make such a mess of Barcelona.

Well, you can't be too far off-beam - you are
being flamed from both sides!

The big difference is that the K7 was competitive against Intel's offerings until the much delayed K8 launch. The K8 is getting it's butt kicked all over the place, so they need to rush K10 out the door, instead of waiting until it's good and ready.

If you think AMD is hurting now, wait a couple years until Intel is doing a die-shrink of Nehalem (assuming they don't sit on it), and AMD is still on K10. I was beginning to think that Intel might actually beat AMD to shipping native quad cores in volume.

These are scary Phenom errata, for which (as a low-level programmer) I see no feasible workaround in a consumer machine, which can run OS/applications not under the control of the BIOS programmer.

"244 A DIV Instruction Followed Closely By Other Divide Instructions May Yield Incorrect Results
Description: A DIV instruction with a divisor less than 64 that is followed in close proximity by a DIV, IDIV, or AAM instruction may produce incorrect results.
Potential Effect on System: Possible data corruption.
Suggested Workaround:" [none given]

"280 Time Stamp Counter May Yield An Incorrect Value
Description: Reads of the time stamp counter may yield an inconsistent result.
Potential Effect on System Undefined behavior for software that relies on a continuously increasing time stamp counter value.
Suggested Workaround:" [none given]

This is straight from
http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/41322.pdf
linked as "AMD Phenom Processor Technical Documentation / Revision Guide for AMD Family 10h Processors" at
http://www.amd.com/us-en/Processors/TechnicalResources/0,,30_182_739_15343,00.html