Things seemed poised to turn around for AMD in 2011. But the abrupt departure of CEO Dirk Meyer on Monday afternoon – at the exact same time that rivals Intel and Nvidia ceased their hostilities and a week after Nvidia jumped into the processor racket – indicates that AMD's board of directors sees challenges that aren't obvious …
It's never easy being a CEO at AMD
With all of the challenges imposed by InHell's illegal activities to thwart AMD, it will never be an easy task for any AMD CEO to keep everyone happy be it customers or the board. I suspect there was more than just the emerging market issues that caused an ideological chasm between Meyer and the board. AMD has survived for 35+ years of InHell abuse and I fully expect them to continue delivering the products that consumers desire.
AMD turned negative...
At the point when Intel started coming back at AMD technically with it's Core 2 desktop CPUs and subsequent Xeons based on similar technology AMD made the big mistake of becoming defensive and trying to consolidate the small gains that they had made against Intel.
This mistake, that simply lost time against a newly fast-moving Intel, was compounded by the purchase of ATI. Way to piss-off the company making the best desktop chipsets for your processors and (though it's been quite a while coming) push them into the arms of your deadly rival.
Speaking as a long-time AMD fan Fusion has sounded great...for far too long. Sandy Bridge is here now, technically excellent and aggressively priced. Add to that the Intel-Nvidia kiss-and-make-up and it's hard to see AMD returning to it's position as a small but serious competitor in the chip biz. Shame really.
It's a long running story
And it's going to go on running.
There will be setbacks; there will be successes...
I just hope that...
AMD gets another Athlon moment. I would totally hate to see Intel getting even more of a monopoly than they do even now. Not that I would say the Core series are bad; not at all. But I want to see competition on the consumer level. That's what keeps prices down and quality up. On the server level, AMD, IMHO, are still well in the game.
ARM with Hypertransport
As I said in another post a few days ago: ARM with Hypertransport.
"higher clock speeds, more cores, and more memory and I/O capacity."
ARM doesn't need to match x86 (NOT x64) clock speeds because ARM's designed-in code density means that any given number of ARM clock cycles gets more work done than any given number of x86 clock cycles.
Multi-core ARM is already here. Outside the server market, how many cores is enough? Outside the server market, two is likely more than enough for the vast majority of the people the vast majority of the time. Intel are only promoting multicore because they've finally hit a brick wall with clock speeds, and Intel need to promote *something* to keep their chips competitive and expensive (and therefore profitable).
ARM doesn't greatly need more memory. It certainly doesn't need 64 bit architectural address space, AMD64 style, and nor does 99.7% of consumer electronics or desktop IT, now or in the next ten years. And on a more practical scale, any given application will be somewhat smaller on ARM (because of code density) than on x86.
More I/O capacity? Hypertransport.
Job done. When do I get a job in Cambridge?
I see "bake" is back...
You're not Bill Gates are you? Don't forget, PC's will never need more than 640K of memory!
Cores and x64
"It certainly doesn't need 64 bit architectural address space... now or in the next ten years"
Not in 10 years, no. But by 19 January 2038, yes.
"ARM doesn't greatly need more memory"
Akin to the "640k memory" quoted already.... IF ARM is to make desktop or (especially) server chips, it will NEED to handle plenty of memory. 4GB is a good start, but x64 is a less-complicated necessity to handle >3.25GB.
And as for your code-density comment, the executable code size isn't the problem. Running exe + dlls (windows) or libs (linux) amounts to a very small footprint indeed, compared to what actually eats up the RAM: multimedia. Images (bitmaps), audio, etc. These are things that can't be blitted in a compressed form, and would be too CPU-intensive to keep in a compressed form 100% in RAM and decompressed each time the screen area needs refreshing (think of your wallpaper for instance).
"When do I get a job in Cambridge?"
Never. Lack of understanding.
Call it quit?
"Without Meyer, what will AMD do next?"
Errr... carry on? It's not like the CEO takes all the decisions on his/her own, they've got countless advisers on all matters, including technical ones. Reminds me of people blaming Bill Gates for calling the start button "Start" (because of course Bill Gates PERSONALLY labeled the button himself, if you check his Outlook calendar he booked a whole week for that).
"Not an easy job", don't know about you but mine is not particularly easy, yet I don't get millions in wages and bonus. I'd gladly take the job for half his wages, and even if I screw up, I wouldn't be the first CEO to display incompetency.
I don't think so
According to any number of reviews, there is nothing "technically excellent" about Sandy Bridge per se. By all accounts it's a lateral performance move for Intel to try and compete with AMD's more user friendly and cost effective products.
The largest volume of chips are sold to people who have no technical clue or care regarding brand. Thus AMD needs to carve it's own nich in both the enthusiasts market as well as mainstream bland PCs. AMD has made some progress in the mainstream channel with HP and others. It's not easy for AMD to compete when Intel routinely makes illegal bribes or threats. The ATI acquisition is just starting to yield benefits but AMD has a big lead on Intel with Fusion.
The loss of Nvidia IMO is no loss at all as they are back-stabbing operators much like unscrupulous Intel. Nvidia is desperate and may not survive now that AMD has acquired ATI.
In your opinion...
Opinions are great: I have a lot of those too, but when the hard facts blow my opinions out of the water it's time to re-evaluate. Check your favourite hardware tech sites: Sandy Bridge delivers X58 performance at a similar price point to the mid/high-end Phenom II with significantly lower absolute power consumption (never mind power/performance). It's not a breakthrough innovation but to argue that that doesn't represent some kind of technical excellence is pure semantics.
Of course you are right that the vast majority of chip sales are made without reference to the technicalities or the fact that Intel behave in a monopolistic, underhand and possibly illegal manner. But those are hard facts too, that's the world that AMD has to survive in and it will only do so by bringing products to market that are significantly superior to Intel's offerings. Will Fusion be such? Who knows? And therefore, as of now, who cares? You and I may love AMD's product but the faceless market could not care less: they have delivered good (but not Intel-beating) chips late too often recently.
The same goes for Nvidia: it doesn't matter if they unscrupulous back-stabbers (though I'd like to know what your evidence is for that), it can only be a mistake for AMD to upset them to the advantage of Intel. Not nice but true.
So your recommendation is AMD to repeat Via
That is what Via did. It is still printing some C-7 and nano CPUs, but it is not putting out anything new there. The SMP multicore offerings were shown only as a prototype and never saw the light of day. However, it has entered the ARM market long ago so we can see an example of what fate may wait for AMD if it does. Nearly all of the cheaper also-run mee-too Android tablets run WM8505 or VT8500 Via SOCs.
That has not done it a lot of good though. The once mighty supplier of most chipsets in cheap computers is nowdays on the periphery picking up scraps. Overall - it did not do it a lot of good.
So while AMD may add Arm Micro Devices to its portfolio, it still needs to fight the battles with Nvidia and Intel on the main fronts to remain relevant.
Same problem as microsoft
So AMD was busy competing with it's traditional rival in an area that is rapidly becoming less important.
It's like MSFT congratulating itself on beating Linux on the desktop - while busily ignoring phones and tablets
Seems a bit of an oversight to have not mentioned AMD selling off ATIs mobile graphics division to Qualcomm at the beginning of 2009.
Perfect example of AMD balancing the books but missing where the market was going.
As reported in Ars Technica's write up here:
Seems like it was a lose/lose situation. Remember in 2009 there was no iPad then! The same board would probably have called him mad for carrying on with that project.
not in consoles?
What about the Wii?
To me it looks like Mr. Meyer was on the path to cloud computing. AMD has a good reputation with server chips, and is now developing a great APU. This looks to me like the perfect combination for cloud computing. Low cost desktops connecting to huge server farms. While the cloud hasn't taken off yet, I can see it coming in the near future.
The Wii uses an IBM powerPC and an ARM - although it does have a fairly low end ATI graphics chip
AMD can't do it all
I'm very glad AMD gave me a socket over a span of time, and gave me decent, fun, value based and good performance as well. A very fair deal between me and AMD.
But in the meantime, workstation, laptop, and server markets and their needs mean Intel has better offerings. I can't argue with the I5 number cruncher and an ability to drop to atom power levels on idle. Intel back when they convinced Apple to join the X86 grouping did so on a Power per watt basis that they are rigorous in working through. And Energy supply in future IS given its cost going to mean everyone's machinery HAS to use less juice.
Therefore, it is an inflection point for AMD. I can't thank them enough for what they did in recvent AM2+ and AM3 times, but that socket and the processors are also reaching an end point. I may yet, and AMD may yet get 6 more months out of them, but their time is ending.
I hope AMD can offer me interesting things in future again. I think over the past 2 years, they did well with what they had.
- Geek's Guide to Britain INSIDE GCHQ: Welcome to Cheltenham's cottage industry
- 'Catastrophic failure' of 3D-printed gun in Oz Police test
- Game Theory Is the next-gen console war already One?
- Analysis Spam and the Byzantine Empire: How Bitcoin tech REALLY works
- Apple cored: Samsung sells 10 million Galaxy S4 in a month