AMD CEO Dirk Meyer Resigns

angry tapir writes "Advanced Micro Devices has announced that Dirk Meyer has resigned from the post of CEO, and that the company is beginning to search for a new chief executive. Meyer resigned in a mutual agreement with the board of directors, and the company has appointed Thomas Seifert, the company's chief financial officer, as the interim CEO. Meyer was installed as CEO in 2008 as a replacement to Hector Ruiz, just as the company was making its way out of rough financial times. In October, AMD posted a third-quarter net loss of US$118 million."

Comments

AMD CPUs all over the place

[e4g4]

With AMD CPUs left and right, how is AMD posting a loss?

Re:AMD CPUs all over the place

[hedwards]

I haven't looked at the financial sheets, but I would venture to guess that a lot of it is R&D trying to catch up with Intel again.

Re:AMD CPUs all over the place

[RightSaidFred99]

They're in a price war with a competitor who is a generation ahead of them in manufacturing technology. Their margins are getting slimmer and slimmer.

Re:AMD CPUs all over the place

[Khyber]

Ahead in manufacturing tech, maybe. Architecture? Who made x64? Who has one of the lowest power draw/highest performance CPU/GPU combo for mobile systems that would shit all over tons of current in-service desktop systems, with an even better revision coming soon?

When it comes down to it, none of the hardware companies are really that impressive. The hardware right now is enough to do WAY BETTER performance-wise, the problem is that programs are becoming so bloated and unoptimized.

It's coming down to where even programs you pay for are wasting resources by beaming ads to you.

Welcome to the future.

Re:AMD CPUs all over the place

[Kjella]

What does volume matter if you don't have margins?

1. Intel has always been ahead on processing tech, often a generation in front or if not on a more mature process. That means AMD has bigger dies and lower yields, which directly inflate cost.
2. A lot of the expense is R&D, and with Intel having ~80% of the microprocessor market each AMD chip has to carry at least four times as much of the cost as each Intel chip.
3. Intel got a processor to match every one of AMDs, the reverse is not true. Intel makes high margins where they are alone and squeezes AMDs margin where they compete.

Seriously, take a look at something like 3D rendering performance, which is usually extremely multi-threading friendly. The 2500K which sells for less than the 1100T is beating it in everything but the POV-ray test. Never mind that it's much faster and better for everything that doesn't take advantage of six cores. The Opteron vs Xeon battle looks the same, AMD had the advantage a while but they're struggling badly now there too. On the low end Intel has the Atom which is raking in money meaning AMD is losing a lot of low-end sales. They're boxed in and in every market they deliver "value" processors. That means in other words low income processors. So with low income and high costs, you post a loss.

Re:AMD CPUs all over the place

[Xest]

What AMD CPUs left and right? My work laptop is Intel, my Netbook is Intel, both my Cell phones have Qualcomm ARM chips.

I don't even buy AMD chips at home now, since whilst Intel remained more expensive, AMD chips always ended up seeming to require excessive cooling, and AMD chips never seemed to give the performance I'd expect, yet the first Intel chip I bought in 15 years for my home PC did straight out the box.

Apart from extremely low margin budget PCs and laptops from your local electronics superstore I don't see much use of AMD chips at all now, they don't really seem to have done a good job of diversifying into an increasingly mobile world, and don't seem to have made any meaningful headway against Intel on the desktop. I understand they've had some success in the server market, but it alone wont cut it, and I'm not even sure the headway they made here is being sustained?

I'm not wishing ill of AMD or anything, I think it'd be a tragedy if they dissapeared because Intel would be largely competitorless on the desktop and have no real motivation to rapidly improve as they have with AMD in competition, but right now I don't see how AMD has been particularly well managed, they seemed to have largely failed to enter new markets well without improving their standing in existing markets also. In contrast, companies like nVidia and VIA seem to be doing a better job of diversifying into new markets, so I don't really see how AMDs woes can be anything other than bad management- presumably that's precisely why this guy is stepping down.

Re:AMD CPUs all over the place

[Anonymous Coward]

You do realise you're not the only person who buys processors? You're not even statistically significant, making the first two paragraphs of your posts completely meaningless.

Re:AMD CPUs all over the place

[Xest]

It's a shame you didn't apply this logic to the person I was responding to, else you may not have completely and utterly missed the point of the paragraphs in my post that you are apparently struggling with.

The OP stated that AMD processors are left and right implying that he believes they are common and perhaps even hold a majority marketshare. I gave a counter example of a personal anecdote with an illustration of my own circumstances where it's not the case that AMD processors are "left and right" demonstrating the point that his experience is not universal to everyone. I'm sorry if you needed this explicitly pointing out to you, most people don't.

My second paragraph merely illustrated the fact that people have their own good reasons to switch from AMD, and whether people have reasons to switch back is neither here nor there, it's merely a point made in the context of the fact AMD is struggling and a demonstration of one of many reasons why they may be struggling.

If you had not jumped straight onto my post with the incorrect assumption that I was trying to use a set of personal anecdotes to prove that AMD has a smaller market share than Intel then you might not have so wildly missed the point of my post. Most people here are already almost certainly aware of the established fact that AMD does not hold a majority marketshare in really any segment of the processor business, most are aware that even for PCs for example - one of AMD's strongest segments - Intel still holds ~80% of the market. Anyone can trivially find this out for themselves with a few seconds of Googling, it's not a point that needs to be proven in a post on Slashdot when it's already fairly common knowledge.

Re:AMD CPUs all over the place

[PsyciatricHelp]

Honestly I think where AMD went wrong is while INTEL has been spending money on companies like Dell to increase how many products contain INTEL, AMD went and bought another company to try and improve them selves. They may be a bit behind due to financial struggles however I am one that definitely does not support those who play the marketing game. Has anyone on here ever seen a TV ad for AMD?

Re:AMD CPUs all over the place

[robthebloke]

Ahead in manufacturing tech, maybe. Architecture? Who made x64?

That's a bit like saying ford is the worlds greatest motor company because of the Model T, and then neglecting to notice the failures of it's recent business practices. x64 was something AMD did right 8 years ago. In the last 8 years however, things haven't gone as well for the business....

Who has one of the lowest power draw/highest performance CPU/GPU combo for mobile systems that would shit all over tons of current in-service desktop systems, with an even better revision coming soon?

Intel ATOM and Nvidia ION?

Re:AMD CPUs all over the place

[shallot]

The OP stated that AMD processors are left and right implying that he believes they are common and perhaps even hold a majority marketshare.

Actually, I don't think that was the point. The point is that when you make anything by the million, it's not obvious why you'd nevertheless post a loss. (It's possibly to occupy a relatively small part of a market and still make a decent living. Indeed, it's also possible to do that and make a lot of money - see Apple.) I guess the answer to this naive question would be - RTFA :)

Re:AMD CPUs all over the place

[I8TheWorm]

Not always. The pentium pro line of processors had a heat problem that Intel "solved" by adding latency layers, and AMD began beating them in benchmarks. Right up until the Core2 processors, when Intel had solved the problem.

Re:AMD CPUs all over the place

[Kjella]

Yes, in benchmarks but not processing technology - litography size and yield. Also you mean the Pentium IV, not the Pentium Pro from the mid 90s.

Re:AMD CPUs all over the place

[Antisyzygy]

We'll see how that all turns out when the Fusion processors come out in laptops soon. I think they will probably take over the laptop market. The desktop market still is going to be Intel dominated unless AMD can push its bulldozer out quicker than they are expected to and sell it at a lower margin.

Re:AMD CPUs all over the place

[RightSaidFred99]

Your problem is twofold. First, the vast majority of users don't need anything beyond the Intel integrated graphics. AMD has had superior chips in the past (Pentium 4/X2 days) and they didn't "take over" anything, and not because of mean old Intel cheating.

This leads is to your second problem - AMD doesn't have the manufacturing might to make chips at the same price Intel does. AMD's only choice is to cut prices to the bone, and Intel can still match them and make good margins.

I expect some design wins for Fusion, but they'll continue to be a bit player like that are now.

Re:AMD CPUs all over the place

[tyrione]

With AMD CPUs left and right, how is AMD posting a loss?

They aren't. AMD's quarterly report is on the 20th of January for it's 4th Quarter. The 3rd Quarter reference is 3 months old.

Re:AMD CPUs all over the place

[Anonymous Coward]

Your personal bias is showing. Did you notice? Well cover it up already... for shame.
 
We could debate relative merits of AMD / Intel all day long. In fact, we will despite anything anyone might say, including me. So for spite.... To counter your anecdote... I'm working on an AMD powered desktop. There is an AMD powered laptop sitting here also. The HTPC? AMD. The media server in a 4u rack in the closet (yes I have a rack in my closet)... AMD. Why? Because price was more important than benchmarks.
 
All that being said... if I were rebuilding all these systems today, some of them would probably end up with intel inside. Not the least reason for which is I now get intel inside pricing. IE, employee discounts.

Re:AMD CPUs all over the place

[Hognoxious]

With AMD CPUs left and right, how is AMD posting a loss?

It's called "making it up on volume", did you sleep through the dot-com boom or something?

Re:AMD CPUs all over the place

[WuphonsReach]

AMD chips always ended up seeming to require excessive cooling

I'm really not sure what you've been smoking, but I suggest you stop.

The last "hot" AMD chip I had was back in the original 1GHz Athlon days, circa 2000 or 2001. It got that reputation mostly because it was about the point of the time that Intel added on-chip thermal regulation and AMD hadn't yet. So if you pulled a heatsink off a Intel chip, it would clock itself downward to compensate. The AMD chips, which didn't yet have that feature, would fail (spectacularly).

All of the Athlon64 and Opteron CPUs since then have included thermal regulation circuitry. It's been a non-issue for both manufs since '02 or so.

Back around '03 or so, AMD shifted their processor feature size down to 90nm (or was that the 65nm shift?). That allowed them to ramp up performance and turn down TDP (thermal design power). During the last 6 years, they've had CPUs that run on as little as 45W TDP, with 60W TDP chips being extremely common. Only the more powerful CPUs run at 90W or 125W now.

Guess what? Intel *also* has chips in the 90-150W TDP envelope. They both consume about the same amount of power, require the same amount of cooling (a lot) to dissipate that much heat, etc. Have you forgotten how loud and hot the old Pentium IV chips ran? Apparently so. Because the rest of us were getting equivalent performance out of AMD Opteron/Athlon64 chips that ran a dozen or two degrees cooler, requiring less fan noise to keep things running. The Xeons have been more hit/miss compared to the Opteron, and Intel was using a different calculation for TDP which also muddied the waters.

But for the last 6-7 years, you had a choice with both Intel/AMD. You could buy the 125W (or higher) part for maximum performance, or go with one of the 60-90W parts which required only very modest cooling fans and heat sinks. And there have been select 45W parts from AMD that let you get by with stock CPU coolers/fans that are near silent. Even the 60W parts from AMD come with stock coolers/fans that are nearly silent now and it is typically the GPU card that ends up making most of the noise in a modern system.

Re:AMD CPUs all over the place

[Hognoxious]

That's a bit like saying ford is the worlds greatest motor company because of the Model T, and then neglecting to notice the failures of it's[SIC] recent business practices.

Indeed, unlike GM they failed to go bankrupt.

Re:AMD CPUs all over the place

[Khyber]

"Intel ATOM and Nvidia ION?"

Not even close. Have you even seen the latest AMD Fusion? It shits all over Intel+ION.

"That's a bit like saying ford is the worlds greatest motor company because of the Model T, and then neglecting to notice the failures of it's recent business practices."

During the auto bailout and financial crisis, Ford was the only company actually able to afford to give loans to people. AMD isn't Bankrupt.

Re:AMD CPUs all over the place

[robthebloke]

Have you even seen the latest AMD Fusion?

Unfortunately not. I requested an engineering sample from AMD about 2 years ago, and am still waiting......

It shits all over Intel+ION.

Whether it shits over ION or not entirely misses the point. I'm sure Fusion *could* shit all over ION, but the problem can be summed up with a quick glance here. Where are these fabled Fusion netbooks? It's ATOMS, ATOMS, ATOMS and nothing but ATOMS in that market sector.....

So returning to your original question:

Who has one of the lowest power draw/highest performance CPU/GPU combo for mobile systems that would shit all over tons of current in-service desktop systems, with an even better revision coming soon?

The ATOM/ION has been available for over 18 months. It has one of the lowest power/highest performance CPU/GPU combo for mobile systems, with a new revision coming. It's adequate enough to replace tons of in-service desktops NOW. The ATOM/ION combo is a very valid answer to your question...... Fusion may become a valid answer in a few months time, but it isn't an answer *yet*.....

Re:AMD CPUs all over the place

[Khyber]

http://liliputing.com/2011/01/31062.html

I was testing this LAST YEAR. Oh look, it's available TODAY.

Re:AMD CPUs all over the place

[Khyber]

Sorry, wrong link:

http://blog.laptopmag.com/deal-of-the-day-hp-pavilion-dm1-with-amd-fusion-for-424

Sorry..

[dark grep]

I for one an sorry to see him go. I think he has brought the company well through some rough times.

Re:Sorry..

[Anonymous Coward]

I for one an sorry to see him go. I think he has brought the company well through some rough times.

Some CEO's that are great for riding through the rough times aren't the CEO's that you want when that stretch is over.

Re:Sorry..

[1984]

Winston Churchill, meet Clement Atlee.

Except I think that involved winning a war, not just surviving in a currently tenuous second position...

Re:Sorry..

[Cornelius the Great]

I'll agree with this. AMD's been seeing some triumphs lately- their graphics division has been very successful, even despite a minor delay with the Radeon HD6900 GPU. Nvidia might have the performance crown this generation, but their previous generation has been shaky and their 40nm chips haven't been as available as AMD's, allowing AMD to gain considerable marketshare.

I've noticed a few netbooks with AMD Bobcat cores appear at CES, and has enough performance and power efficiency to give both Atom and Ion some serious competition.

While Llano doesn't appeal to me personally, it's nice to see Fusion reaching the desktop shortly. I'm also anxious to see how the Bulldozer will perform once it's released in a few months.

With the delay of Intel's Ivy Bridge into 2012, AMD has a lot of potential to make this year a profitable one.

Re:Sorry..

[electrosoccertux]

I'll agree with this. AMD's been seeing some triumphs lately- their graphics division has been very successful, even despite a minor delay with the Radeon HD6900 GPU. Nvidia might have the performance crown this generation, but their previous generation has been shaky and their 40nm chips haven't been as available as AMD's, allowing AMD to gain considerable marketshare.

I've noticed a few netbooks with AMD Bobcat cores appear at CES, and has enough performance and power efficiency to give both Atom and Ion some serious competition.

While Llano doesn't appeal to me personally, it's nice to see Fusion reaching the desktop shortly. I'm also anxious to see how the Bulldozer will perform once it's released in a few months.

With the delay of Intel's Ivy Bridge into 2012, AMD has a lot of potential to make this year a profitable one.

so the guy that brought AMD to a position where they're successfully launching 3 products in one year (which they've never done before) is not someone you want to keep around? Are you kidding? It's about momentum and inertia, this is a silly time to do something like this to a visionary like him.

This is the same board that kept that dolt Hector Ruiz around for years while he ran the company into the ground. Color me unsurprised.

Re:Sorry..

[Cornelius the Great]

I hate to break it you, but Fusion, Bobcat, and Bulldozer have been in development for quite a long time- all of these projects started when Hector was at the helm. Dirk can hardly be credited with these product releases, other than keeping AMD afloat long enough to allow these products see the light of day.

Re:Sorry..

[Rockoon]

so the guy that brought AMD to a position where they're successfully launching 3 products in one year (which they've never done before) is not someone you want to keep around? Are you kidding?

Are you honestly asking this question? If you are going to pretend to know anything about the business world, then you should at least pretend to also know that some CEO's are specialists at bringing companies out of financial trouble and even bankruptcy.

For example (from my industry) there is Scott Butera, a CEO that has brought more than one casino out of financial trouble, who has just been picked up by the Foxwoods Resort Casino in Connecticut because of its very serious financial troubles (billions in debt, defaulting on loans..)

Often what these specialists bring to the table is their trusted contacts in the financial industries. The primary goal is often to maintain a credit line while the problems are resolved (because no large business can run without credit, regardless of how much cash they have.)

Re:Sorry..

[Raenex]

The primary goal is often to maintain a credit line while the problems are resolved (because no large business can run without credit, regardless of how much cash they have.)

That doesn't make a whole lot of sense. Large companies like Google, Microsoft, and Apple have billions in cash reserves, and can probably operate for a year or two by spending just cash.

Re:Sorry..

[Anonymous Coward]

I have to say, from personal experience, Mohegan Sun is a better gambling casino than Foxwoods is.

If you want food, or if you want to play Bingo, Foxwoods is the place to go.

Re:Sorry..

[Rockoon]

Well I dont know which is "better" or "worse."

What I do know is that Foxwoods problem was expanding their operations on debt one to many times. It was already the largest casino in the country but still in debt from its previous expansions when they chose to build the billion dollar MGM wing.

Re:Sorry..

[bhcompy]

And they keep those cash reserves for dividends, acquisitions, and keeping their credit rating high.

Re:Sorry..

[Raenex]

But the companies mentioned don't pay dividends, don't *need* to acquire companies (and can afford to do so anyways and still operate on cash), and don't *need* credit if they have cash to spend. Remember the person I responded to said "no large business can run without credit, regardless of how much cash they have". It just isn't true.

Re:Sorry..

[Anonymous Coward]

Except I think that involved winning a war, not just surviving in a currently tenuous second position...

Britain was hanging on (i.e. surviving in second place to the krauts) until the yanks got involved. Then it all changed...

Musical chairs

[syousef]

Hey, that's one clever way to get your mind off the recession: Play musical chairs with the company execs. Did you see the job open up at Microsoft? Time to apply Dirk! Where she stops nobody knows....weeeehheeeeee! You poor schleps can lose your jobs, we'll just keep going round and round!

Re:Musical chairs

[hedwards]

Actually, that's typically how it's handled when the CEO makes an abrupt departure, one of the other executive board members will step in while they find a replacement.

Personally, I don't blame him, I blame it on Intel and its successful attempts to bribe major equipment integrators to not use AMD chips.

Re:Musical chairs

[Anonymous Coward]

No amount of bribes in the world can account for the fact that Intel's latest processors have been significantly outperforming AMD's for the last few years now.

While I'd agree that Intel has been guilty of questionable business practices, only a fool would attempt to explain away all of AMD's current failures by blaming Intel.

Re:Musical chairs

[Rockoon]

No amount of bribes in the world can account for the fact that Intel's latest processors have been significantly outperforming AMD's for the last few years now.

Yet AMDs significantly outperformed Intels for quite a few years.. but only managed 50% market share at its height because of Intels illegal (no "questionable" about it) practices.

Re:Musical chairs

[serviscope_minor]

No amount of bribes in the world can account for the fact that Intel's latest processors have been significantly outperforming AMD's for the last few years now.

Really? Your sure about that? You're sure that if Intel lost out badly during the P4 versus Opteron era, they would not be behing on their R&D (due to less money) and AMD would not be ahead (due to more money)?

And Intel underperform on price/performance. They only win in the performance region that AMD don't enter, which is the very high end, and especially per-thread performance.

Re:Musical chairs

[DarkXale]

Thing is though that for most users, per-thread performance is the key; and despite AMDs processors being cheaper they still lose on the price/performance in most applications.

Re:Musical chairs

[red crab]

Months ago I was shopping for a new laptop; i was inclined to buy one with AMD in it. Was shocked to find however that no vendor apart from Acer had an AMD product. Dell, HP and Lenovo were blatantly advertising the new ix series processors from Intel. I had no option but to settle for an i5 520. Marketing had certainly a lot to do with it and AMD can't match Intel on that.

Re:Musical chairs

[JamesP]

Yet AMDs significantly outperformed Intels for quite a few years.. but only managed 50% market share at its height because of Intels illegal (no "questionable" about it) practices.

Exactly. And I was a big fan of AMD at that time.

But at the time of Prescott processors AMD began to decline in innovation. It stuck with Athlon XP (Barton) architecture
And the Athlon 64 was not much better

AMD needs a 'push' like the one it gave Intel at the P4 days

Re:Musical chairs

[Antisyzygy]

I don't know where you get your statistics, but AMD does win on price/performance in all but the highest end processors which most people either don't need or can't afford.

Re:Musical chairs

[Antisyzygy]

AMD needs a 'push' like the one it gave Intel at the P4 days

Such as Fusion? I predict a significant increase in AMD's share of laptop and ultraportables market in the near future.

Re:Musical chairs

[Antisyzygy]

Uhh. HP has quite a few laptops with AMD in it.

Re:Musical chairs

[JamesP]

Fusion is very interesting, maybe this is it

But I was thinking more about architectural improvements (for example hyperthreading http://www.anandtech.com/show/2594/8)

Re:Musical chairs

[Owy]

NReally? Your sure about that? You're sure that if Intel lost out badly during the P4 versus Opteron era, they would not be behing on their R&D (due to less money) and AMD would not be ahead (due to more money)?

Intel has a vast amount of business interests, not only in the processor market (at the time of Pentium iv I remember reading somewhere that the processor business was only 20% of all the Intel business), while most of AMD's was only in the processor market. So when Intel saw that they were losing a good market (specially because of the reputation provided) they started investing immense amounts of money from other markets. So, no they could will not remain behind for a long time.

Re:Musical chairs

[confused one]

So does Lenovo. *checks* and so does Toshiba, MSI and Asus... They're not the majority; but, AMD has never been strong in the laptop segment. Newegg has 111 models from 9 vendors in stock. Some are older models and there's some overlap (differences in installed memory, etc); but, parent didn't try too hard when he looked.

Re:Musical chairs

[Anonymous Coward]

AMD was the one bragging in those days that it sold everything it could produce. AMD did not have infinite production capacity.

BTW, the 50% figure is the high point of AMD's 4s and higher server market share. AMD never had anywhere near 50% of the x86 desktop/mobile or server market by units or revenue. AMD peaked a little north of 25% by units shipped, again which was capacity constrained.

Re:Musical chairs

[syousef]

Thing is though that for most users, per-thread performance is the key; and despite AMDs processors being cheaper they still lose on the price/performance in most applications.

It's not even about performance. I got bitten by instability issues after building an AMD 1800+. I still have the machine, but it never got used for anything important. Never again. Intel might be the devil, but they're the stable devil.

After intel's sandy bridge launch...

[aws4y]

I would resign too, AMD is always the bridesmaid never the bride.

Re:After intel's sandy bridge launch...

[exomondo]

I would resign too, AMD is always the bridesmaid never the bride.

Their best run was with the Athlon64 vs Netburst, but even though they had the superior product they didn't have the OEM deals.

Obviously it's to get the jump on their competitor

[Lord_of_the_nerf]

Does this mean in six months time Intel Ceo Paul Otellini will quit, but for higher severance package?

Any word yet

[publiclurker]

on how many million they will be rewarding him for losing that $118 million?

Re:Any word yet

[Dexter Herbivore]

No idea, but at least being AMD they can calculate it correctly.

Re:Any word yet

[Dexter Herbivore]

Are you seriously trying to compare a few poor design interface decisions like Bob and a Start button, with the greatest mass chip fuck up of all time? Maybe you should learn a sense of scale in regards to the magnitude of all time fuckups. I somehow don't think that Bob cost Microsoft $475 million dollars.

Re:Any word yet

[Your.Master]

No, he's comparing extremely out of date jabs to an extremely out of date jab. Maybe you want to complain about the Apple III not having sufficient cooling in an article about Dell desktops.

Re:Any word yet

[Anonymous Coward]

Hey dude. You know any good ENIAC jokes?

Re:Any word yet

[Dexter Herbivore]

You seem to be missing my point, there's minor PR mistakes and there's screw ups that potentially caused a loss of life. Let's not forget the more serious of the two.

Re:Any word yet

[Raenex]

I agree with the Anonymous Coward. Bashing Intel for a 1994 flaw? Come on. It's 16 years later.

By the way, it's not like AMD never makes mistakes.

Re:Any word yet

[Hognoxious]

I suspect his golden parachute will look like a knotted napkin compared to the ones failed bankers get.

Then again a banker would probably lose 100 times as much so it stands to reason he deserves more...

Hard call for GPU selection

[DoofusOfDeath]

I'm just getting going on GPU programming. I was thinking to go with OpenCL (pushed by AMD/ATI ) over CUDA (pushed by nVidia) because I thought AMD looked more likely to survive in the long term. But now it's getting harder to tell which company is safer to rely upon.

Re:Hard call for GPU selection

[Anonymous Coward]

Rely on a company? You must be new here.

Re:Hard call for GPU selection

[Anonymous Coward]

I'm just getting going on GPU programming. I was thinking to go with OpenCL (pushed by AMD/ATI ) over CUDA (pushed by nVidia) because I thought AMD looked more likely to survive in the long term. But now it's getting harder to tell which company is safer to rely upon.

OpenCL works on both AMD and nVidia GPUs , so you should be safe there.

Re:Hard call for GPU selection

[DeadBugs]

Actually NVIDIA was a major contributor to OpenCL http://www.realworldtech.com/page.cfm?ArticleID=RWT120710035639

Re:Hard call for GPU selection

[Cornelius the Great]

Intel has also promised OpenCL support on Sandy Bridge and later integrated GPUs. Not to mention S3 and VIA support.

I predict that Cuda will quickly become irrelevant and die a long, slow death (ie- just legacy support, no new features). Much like Cg did, after GLSL and HLSL matured. No one wants to be stuck on a single hardware platform, despite performance advantages.

Re:Hard call for GPU selection

[GigaplexNZ]

Intel has also promised OpenCL support on Sandy Bridge and later integrated GPUs.

As far as I'm aware, Intel promised that they'd get OpenCL running on the CPU (for Sandy Bridge) as opposed to the GPU after Apple pressured them for OpenCL support. OpenCL on GPU is scheduled for Ivy Bridge.

Re:Hard call for GPU selection

[Anonymous Coward]

I have long sworn off NVidia for personal use, because of, in my experience, inconsistent relability in product by 3rd party implementers, and poor drivers by NVidia itself.

However, I think the choice for GPU programming depends on purpose: do you want a wide/broad market distribution to work on anyones hardware, or will it be purpose built with a hardware scaling capability?

CUDA I see more as a dedicated, upscaling solution towards GPGPU, and there seems to be far more [interesting] GPGPU based solutions being put into production that are based on CUDA/Tesla/Fermi than on Stream/Fusion. I don't follow the field closely, but that's been my general impression from the news feeds and forums I semi-routinely check. I think if you are just toying with it to get a sense of what is possible, it would be a toss up, but if intending to field a serious, large-scale system based on a GPGPU, I would be very much more inclined [today] towards the NVidia CUDA solution on a Fermi card for development, with potential to deploy to a Tesla [array] without correspondingly significant changes in software.

Re:Hard call for GPU selection

[Twinbee]

From what I've heard, Nvidia are planning to allow CUDA to compile to other processors such as Intel's x86 (and presumably the new ARMs).

Remember, in the end, it's just a language, albeit a language that's biased slightly towards Nvidia's architecture. But as like all other languages, there's nothing stopping them from making that same code run on any platform.

It's also somewhat easier to use than OpenCL from what I've heard.

Re:Hard call for GPU selection

[Anonymous Coward]

OpenCL also compiles down to x86 (and, eventually presumably, ARM). The host machine just looks like another target (eg, you enumerate the OpenCL compute units on your machine to choose which to compile your kernel for; both the GPU and CPU appear in the enumeration).

Re:Hard call for GPU selection

[Anonymous Coward]

poor drivers by NVidia itself.

Anyone who says, "poor drivers by NVIDIA", is either an idiot and/or a troll. NVIDIA's drivers are literally the metric to which all other GPU companies aspire. They are the benchmark. So needless to say, I need not read any further to completely understand you're an idiot and/or a troll.

Re:Hard call for GPU selection

[thejynxed]

LOL, yet, time and time again, nVidia does the same stunts with their drivers as they did at the release of Vista, or did you forget all of that?

We won't even go into having to use our ovens at home to re-flow the shitty soldering on their video cards.

Re:Hard call for GPU selection

[Anonymous Coward]

thats simple if you want a massive and growing userbase to use your code/apps, go ARM and Freescale quad A9 at that This Year:D

good luck getting anything AMD video decode working as they cant even be bothered to release the Open Decode library for linux never mind open the UVD ASIC to linux developers.

Unification under DirectX

[vlueboy]

Someday MS might give us a standard wrapper for both.
Oh, who am I kidding? MS only cares about their own DirectX product as te be-all/end-all. But they might need it as bait for coders in your same dilemma, as XP and DirectX9 are still their own strongest rival for DX10 and upcomig DX11 adoption.

In other words, if corporate America finds a killer app for CUDA "soon," MS could start selling the idea that XP/DX9 upgrades to Windows 7 and 8 are their only upgrade path to gain built-in CUDA-like support.

Re:Unification under DirectX

[macshit]

Isn't the MS-only analogue of OpenCL/CUDA "DirectCompute"?

OpenCL is clearly the portable choice, but I presume Microsoft will try to gimp that on Windows to encourage people to use DirectCompute.

Re:Unification under DirectX

[GigaplexNZ]

MS already has DirectCompute which is an alternative to OpenCL/CUDA and already runs on existing NVIDIA and ATI GPUs.

Re:Unification under DirectX

[Rockoon]

Just to add to this..

..the predecessor to DirectCompute was a little .NET library that came out of Microsoft Research called Accelerator which was initially available to the public in 2006.

..thats several years before CUDA (2008) and OpenCL (also 2008)

Microsoft has actually been the innovator on this one.

Re:Unification under DirectX

[Bitmanhome]

CUDA was announced November 2006, so Microsoft wasn't that far ahead. But mass-market GPGPU really got started around 2000, culminating in the Brook project in 2004. Microsoft didn't start this trend, though they did jump on it quickly.

Re:Unification under DirectX

[zigfreed]

A quick wikipedia search indicates the CUDA SDK came out Feb 2007 and the Streams SDK v1 came out Dec 2007. Following links from Streams gets you to BrookGPU, an ATI project using the Close-to-Metal programming interface who's last major release was 2004.

Microsoft is only an innovator where it has a monopoly.

Mobile Failure

[DeadBugs]

http://www.brightsideofnews.com/news/2011/1/10/coup-at-amd-dirk-meyer-pushed-out.aspx It seems that that the selling off of their mobile business and the success of Tegra is behind this.

Re:Mobile Failure

[Antisyzygy]

As far as I knew they already DID sell off the mobile business and it was partially Dirk Meyer's fault so they removed him.

Dirk, we miss you already

[Anonymous Coward]

Thanks for steering us back from the brink, Boss. I really appreciated having a both a capable engineer and a courageous leader at the helm.

Re:Dirk, we miss you already

[Anonymous Coward]

Indeed...

I'm no silicon engineer...

[QuantumBeep]

I'm no silicon engineer, but I have to imagine that AMD has SURELY got something big in the works.

Think about it. They've essentially had three chips in ten years. Athlon, Athlon64, and Phenom. Everything else is minor variation and process evolution. That's not a lot, really.

As I see it, AMD is either biding their time, holding their market segment down with their really-stretched-to-the-limit Phenom architecture, while perfecting the next generation product...

OR

They've just been fucking off for the past ten years. We're going to find out soon which it was, because the Phenom II 1100T looks like a hard limit to me, and Llano doesn't look to be much of an evolution.

Re:I'm no silicon engineer...

[Trogre]

You forgot Opteron. Where else can I get twelve physical cores per CPU? Not Intel, that's for sure.

Re:I'm no silicon engineer...

[Anonymous Coward]

We don't run cryshit on servers.

Re:I'm no silicon engineer...

[Anonymous Coward]

Maybe in Crysis but not in serious multi threaded applications. You know, you need to make money once they You out of your mother's basement.

Re:I'm no silicon engineer...

[Rockoon]

You're forgetting that AMD has a very comfortable performance lead at the very high end

Yes, those are AMD 48-core system at #1... and #2... and #3

Then there is the old performance per dollar metric where AMD has the top 7 chips on the market right now.

Intel definitely has some good chips, but aside from a small group of them, they are terrible value (rip off) and also not something they are selling a whole lot of (if you are throwing down $1000 for the CPU, you are probably in the market for a server chip with significantly better memory bandwidth than that i7-980 offers)

Re:I'm no silicon engineer...

[Anonymous Coward]

I am a silicon engineer and you do an amazing job of referencing one graph that shows AMD leading.
Does AMD have a comfortable performance lead at the very high end? http://www.cpubenchmark.net/high_end_cpus.html
How about in OC systems? http://www.cpubenchmark.net/overclocked_cpus.html

Re:I'm no silicon engineer...

[drinkypoo]

In price per retired instruction, AMD is the clear winner, even when you have to use more processor packages to get more speed than an intel solution. Who cares what you say after that? The situation is even more clear when you bring in Intel's flagship, where they're not following AMD's lead vis-a-vis instruction sets, Itanium. Which is a fucking dog.

AMD is leading the value proposition today.

Re:I'm no silicon engineer...

[SpinningCone]

might depend on market - for regular desktops it's not really true anymore.Intel still does the "hey are you EXTREME?!?! pay $1000 for this" thing but the Phenom 970 is about $185 bucks and so is the similar performing intel i5 750. and the new sandy bridge processors are performing very well even against their 6 core gulftown cousins for bout half the price or less ($300 )

at this point in time the bang/buck isn't favoring AMD like it used to and when spending so much on a PC sliding $20-$50 for a much better CPU isn't too bad.

AMD will probably be fine though they had much better CPUs a few years back when the D series intels were sucking.

Re:I'm no silicon engineer...

[CAIMLAS]

I agree with you.

Now, if only Dell and HP,etc. would start making AMD based systems which were not 'entry grade'. You know, the shitty cheap systems with few upgrade options, no redundant PSUs, etc.

I'd suspect the AMD/ATI chipsets of not being as good, but I'm not really sure what the root cause is.

Re:I'm no silicon engineer...

[confused one]

Opteron is/was a derivative of Athlon, Athlon64 and Phenom. Just like Xeon is/was a derivative of Pentium2, Pentium3, Pentium4, Core2, and Core i7. They have additional cache, features to support multisocket and ECC memory for servers, and are tested to a different standard.

Still an AMD fanboy

[madwheel]

It's unfortunate, but regardless, I will be a die hard AMD supporter. They've helped keep the market competitive, have much better business practices, and always have the end-user in mind with regards to their CPU socket configurations. Or should I say configuration? One socket for a massive range of CPUs. I like being in control of my upgrades. I can't stand that Intel changes MB socket types with damn near every CPU and expect it to be alright to fork over a couple hundred bucks in addition to the CPU price. AMD has never let me down since I switched during the K7 era. I for one can not wait for the Bulldozer. I know right now the new Sandy Bridge chip is simply amazing but I can wait a few months.

Re:Still an AMD fanboy

[Rockoon]

Here is my predictions:

1) Bulldozer will beat Sandy Bridge on integer performance (clock for clock) but will lag badly on floating point performance (clock for clock.)

2) Nobody will complain about DRM in AMD processors because they dont have any.

Re:Still an AMD fanboy

[drinkypoo]

AMD has been the clear leader for consumer choice and value since the K6/3, which is a complete monster. Actually, the K6/2 is a beast as well, but it has a crap fpu still. Not that I expect anyone to do this today, but if you actually compile your full system for K6 (hello, Gentoo!) then you will beat the pants off a Pentium II of the same clock rate and cache... not to mention that a K6/3 system with external cache has an L3 cache because of the integrated L2. Unfortunately they were saving their pennies for the upcoming K7 launch so they didn't have the money to do a fully rebranded K6/3 launch as a new product which could actually compete with the P2 in the x86 market.

New intel processors are always astoundingly expensive until the next AMD processor comes out, so IMO it is safe to say that you should always wait for AMD to bring out a processor before a new purchase even if you don't plan to buy AMD. And if more people did that, Intel would drop their pricing and a new equilibrium would be found, but a lot of people who didn't understand (or won't forgive) the K6 FPU debacle will never forgive AMD for their one (very real) failure. If you look back at AMD's list of attempts to rival or even surpass intel they are all massive successes, including 40 MHz 386s, low-power 486s, the oddly competitive yet overlooked 586...

Re:Still an AMD fanboy

[Anonymous Coward]

1) Bulldozer will beat Sandy Bridge on integer performance (clock for clock) but will lag badly on floating point performance (clock for clock.)

Unlikely. AFAIK, AVX on AMD will be implemented via SSE instructions for the most part. Intel parts will contain true 256bit register ops. Expect integer and floating point performance on Intel to be double that of amd once the benchmarks are in....

Re:Still an AMD fanboy

[CAIMLAS]

Yes, the K6/2 was an incredible system.

I remember doing just that - beating the pants off friends and their Intel 233Mhz machines with my same-clocked K6-2 - even with a worse graphics card, I got higher FPS in games like Quake. Windows operation was visibly better and more responsive, as well. AMD had the edge in this department all the way up to the current, Nehalem based Intel systems, IMO. Unless I'm doing server wrok, I'll take an AMD Athlon64 CPU over a 5200 Xeon or similar.

cut that in half

[Anonymous Coward]

and try and go for manufacturing capcity and play to the NOW gen, rather then RICH next gen.

thats a strat that might pay off in long run and give you more umf to get back in race.
Over spending will kill.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Amd paid thier best engineers poorly

[assemblerex]

they all left, party over. Long story made short.

Re:Amd paid thier best engineers poorly

[Anonymous Coward]

As someone who was able to compare salaries when AMD bought ATI, I've got to take issue with that comment... AMD actually paid *very* handsomely compared to ATI (and that pay disparity still exists, and causes some resentment... but that's a separate issue). And of course, there are tons of highly competent, skilled, and creative engineers still in the company.

A lot of "top" ex-ATI talent has gone elsewhere (also starts with 'A'), but that phenomenon is almost exclusively limited to Silicon Valley. In general, the hop from job-to-job culture is far less practical when there are only a handful of ASIC jobs to be had in certain areas. Many "top" CPU guys are still around too, so far as I can tell (not my department).

The thing that I notice from all of these moves is that ex-ATI people are on the move upwards, largely displacing the CPU side. The poor execution of the latter group is a large part of that, no question. The trouble is, most of the moves upwards are being made by people in (you guessed it), Silicon Valley. The headquarters is still in Austin, but it's becoming little San Fransisco. The reason this is a problem? Well, it's building resentment in almost the entire remainder of the company, which is a rather large organization. The CPU guys are annoyed that everything is moving under formerly-graphics ownership (add that to the irritated sentiment that AMD overpaid for ATI...), and two-thirds of the GPU guys are annoyed that everything is moving under Silicon Valley ownership. Some changes are viewed as being unfair (such is life) and some are clearly undeserved (ATI had some big screw-ups too). The politics is actually pretty bad; it's much worse than any other place I've ever worked.

We've got a group meeting about this announcement this afternoon. I wonder what the spin will be...

captcha: tantrum

haha...

Dirk Meyer: Engineer who made marketing mistakes

[Futurepower(R)]

This article explains: Coup at AMD: Why was Dirk Meyer Pushed Out? Quote:

"Remember, Dirk Meyer's three deadly sins were:

1) Failure to Execute: K8/Hammer/AMD64 was 18 months late, Barcelona was deliberately delayed by 9 months, original Bulldozer was scrapped and is running 22 months late.

"2) Giving the netbook market to Intel [AMD created the first netbook as a part of OLPC project] and long delays of Barcelona and Bulldozer architectures.

"3) Completely missing the perspective on handheld space - selling Imageon to Qualcomm, Xilleon to BroadCom."

There is a comment at the bottom of this poor-quality article in the Inquirer that says Dirk Meyer "was the lead engineer who designed the Athlon, Opteron and the DEC Alpha. Let's not forget that from 1999-2006, AMD actually had better processors than Intel, and this was due to Dirk Meyer's technology."

Re:Dirk Meyer: Engineer who made marketing mistake

[Anonymous Coward]

AMD does not make netbooks, they make semiconductors. Why are you blaming them for somebody elses problem? Le me guess, you are a fucking dipshit.

Talk to the Oracle.

[Deathlizard]

Considering their market cap, and Oracle's interest in chip companies, It wouldn't surprise me if Larry Ellison isn't their next CEO.

Re:Dirk Meyer: Engineer who made marketing mistake

[drinkypoo]

OK, I should not feed you, but go forth and see how you like getting any Athlon L110/R690M netbook that is A) not heavily based on the same design B) all functionality working on some other operating system than Vista. Go!

Me too!

[Anonymous Coward]

I just resigned a coupe of weeks ago from my utterly shitty job, too. Coincidence ? Hard to believe!

Re:Dirk Meyer: Engineer who made marketing mistake

[Antisyzygy]

AMD doesn't make laptops or netbooks, they make processors. Go Intel if you want to pay twice as much for a moderate increase in performance. Go Nvidia if you want inferior hardware.

Wasn't the loss posted in 2009 not 2010...

[Anonymous Coward]

I may be getting bad data from my charts, but I think they posted a profit for third quarter (calendar) of 2010..

Re:Dirk Meyer: Engineer who made marketing mistake

[import]

1) Dirk wasn't CEO til 08, well after AMD64 and Barcelona. Bulldozer was scrapped *by* Dirk IIRC

2) AMD had nothing to compete with the atom in that market. Sure, they were first, but with a Geode IIRC slower than an Atom which is already pathetically slow.

3) This is arguable, it falls under the strategy where AMD needed to focus on core business. Whether or not you buy that or you believe they should've switched core business or whatever is where its arguable. I think from AMD's fiscal performance, Dirk probably did the right thing.

Re:Dirk Meyer: Engineer who made marketing mistake

[drinkypoo]

AMD doesn't make laptops or netbooks, they make processors. Go Intel if you want to pay twice as much for a moderate increase in performance.

AMD does make processors and chipsets intended for use in netbooks. Intel doesn't make netbooks either but you suggest I "go Intel if I ant to pay twice as much". I went with an AMD-based netbook and I got a machine that only works 100% with Vista. It's actually faster than the cheaper Intel-based netbook I bought around the same time, but it's also a pain in my ass. The part that works worst is the graphics hardware, of course. ATI is the Taco Bell of graphics. I love my AMD CPUs (I have an Phenom II X3 720-based system right now) but ATI video is a pathetic joke. And actually, I wouldn't mind having working power saving for the Athlon 64 L110 in my AMD-based netbook in Linux, either, but for some reason they disabled the mechanisms that are actually working even though they are present on the CPU.

I was fucked hard by AMD's deliberate failure to release Windows 7 drivers for this hardware which was still shipping after the release of Windows 7 (though I bought before) and I. Will. Not. Forget. They gave good support for Geode and I mistakenly thought that something they would put the Athlon 64 name on would be the same. I was wrong. If a netbook says AMD and/or ATI on it, I'm not going to give it a second look.

Re: The imbecilic Ford analogy

[wkcole]

Ahead in manufacturing tech, maybe. Architecture? Who made x64?

That's a bit like saying ford is the worlds greatest motor company because of the Model T, and then neglecting to notice the failures of it's recent business practices.

That's a spectacularly ignorant statement.

Ford is the only one of the "Big 3" US-based auto manufacturers that has not been put through a bankruptcy backed by government financing. They took steps well ahead of the financial meltdown to turn as much of their non-performing assets into liquidity (e.g. the blue oval trademark is collateral for a big fat loan, or put more harshly: they hocked the family jewels.) Bill Ford recognized that his name didn't mean he was the best strategic manager for the company and hired Alan Mulally to be CEO. They have jettisoned low-margin operations on their own terms and are making very good cars. They also offer the unchallenged best system offered by any manufacturer anywhere for car/{phone,media player,PDA} integration. I hate to admit that last bit because they are getting a lot of that from MS and I absolutely despise MS, but facts are facts. Ford's product pipeline and management are the envy of the auto industry: not just the resuscitated GM and Chrysler, but everyone. Do you think Ford's recent business practices don't mark them as the gem of US automakers? Do you think their record in the past 5 years isn't a greater feat than the Model T? If so, you're an idiot.

And sure, I'm biased. As of last month, my wife is a Ford employee. She was a contractor for them for longer than the dozen years I've known her and has made it through at least 3 rounds of "headcount reductions" since 2000 as Ford did what it needed to do to survive.

Re:Dirk Meyer: Engineer who made marketing mistake

[Antisyzygy]

Sure thing. ATI is a horrible company, they just had some of the top performing hardware released earlier and for cheaper than Nvidia and better drivers recently than Nvidia. Nvidia had to struggle just to release Fermi, and even then at a price hike compared to ATI. You are letting one failure color all your future consideration of AMD/ATI in the portable market and that is just plain stupid especially in light of the fact that its based on a Microsoft problem (i.e. domination of the market with inferior crap software that manufacturers have to cater to since people are stupid and keep buying windows). I have had other not entirely similar problems with Intel on desktops, so should I just swear off Intel because they failed me once even though their i3's i5's and i7's are a generation ahead? I can make purchase decisions based on my income and price/performance which in my opinion AMD wins right now, however if I were to have a need and disposable income for an i7 Id be damn sure to buy one in spite of their failure for me in the past.

Re:Dirk Meyer: Engineer who made marketing mistake

[drinkypoo]

Sure thing. ATI is a horrible company, they just had some of the top performing hardware released earlier and for cheaper than Nvidia and better drivers recently than Nvidia.

Better drivers? By what metric?

You are letting one failure color all your future consideration of AMD/ATI in the portable market and that is just plain stupid

One deliberate failure. AMD produced the architecture and then abandoned it. I am still paying for this failure. I still own the hardware. They're still not bringing out updated drivers that work properly.

especially in light of the fact that its based on a Microsoft problem (i.e. domination of the market with inferior crap software that manufacturers have to cater to since people are stupid and keep buying windows).

Oh, so now AMD's failure to bring out newer, properly working drivers for newer Microsoft operating systems than Vista is Microsoft's fault? It's Microsoft's fault that they fucked up power management under Linux?

I have had other not entirely similar problems with Intel on desktops, so should I just swear off Intel because they failed me once even though their i3's i5's and i7's are a generation ahead?

Haha roflsnort "a generation ahead" uh no.

I can make purchase decisions based on my income and price/performance which in my opinion AMD wins right now

Sure, for budget desktop processors, that's why I'm using one right now. But I wouldn't touch them for a netbook until they have a proven record of supporting the hardware that they sell that manufacturers are meant to put in them. So far they do not. They brought out an Athlon 64 for netbooks and abandoned it. They brought out another Athlon 64 for netbooks and there hasn't been enough time to tell if they will abandon it. Geode has been end-of-lifed instead of being updated so that's over too. So far they are 0/2 (1 pending.) That's a pretty shitty record by any standards.