Slashdot Mirror
AMD Confirms Commitment To x86
MrSeb writes with an excerpt from an Extreme Tech story on the recent wild speculation about AMD abandoning x86: "Recent subpar CPU launches and product cancellations have left AMD in an ugly position, but reports that the company is preparing to jettison its x86 business are greatly exaggerated and wildly off base. Yesterday, Mercury News ran a report on AMD's struggles to reinvent itself and included this quote from company spokesperson Mike Silverman: 'We're at an inflection point. We will all need to let go of the old 'AMD versus Intel' mind-set, because it won't be about that anymore.' When we contacted Silverman, he confirmed that the original statement has been taken somewhat out of context and provided additional clarification. 'AMD is a leader in x86 microprocessor design, and we remain committed to the x86 market. Our strategy is to accelerate our growth by taking advantage of our design capabilities to deliver a breadth of products that best align with broader industry shifts toward low power, emerging markets and the cloud.' The larger truth behind Silverman's statement is that no matter what AMD does, it's not going to be 'AMD versus Intel' anymore — it's going to be AMD vs. Qualcomm, TI, Nvidia, and Intel."
The larger truth behind Silverman's statement is that no matter what AMD does, it's not going to be 'AMD versus Intel' anymore — it's going to be AMD vs. Qualcomm, TI, Nvidia, and Intel."
Considering the execution of Bulldozer, you could possibly add AMD to the vs. list.
A feeling of having made the same mistake before: Deja Foobar
An inflection point is when the second derivative of a curve is zero, and is generally where a curve changes from concave down to concave up. I'm assuming the guy was talking about a profit over time curve or something.
To offset political mods, replace Flamebait with Insightful.
Obviously it can also change from concave up to concave down, before the flamers and math nazis reply.
To offset political mods, replace Flamebait with Insightful.
AMD is a leader in x86 microprocessor design, and we remain committed to the x86 market. Our strategy is to accelerate our growth by taking advantage of our design capabilities to deliver a breadth of products that best align with broader industry shifts toward low power, emerging markets and the cloud.
This is a completely meaningless statement. You could say the same exact thing about any microprocessor company. For example:
"Freescale is a leader in PowerPC microprocessor design, and we remain committed to the PowerPC market. Our strategy is to accelerate our growth by taking advantage of our design capabilities to deliver a breadth of products that best align with broader industry shifts toward low power, emerging markets and the cloud."
This statement is true even though AMD and Freescale aren't competitors.
This is the kind of garbage that makes employees think that their managers are clueless and don't know how to fix the company.
And the men who hold high places must be the ones who start
To mold a new reality... closer to the heart
'We're at an inflection point. We will all need to let go of the old 'AMD versus Intel' mind-set, because it won't be about that anymore.'
I actually find that a little saddening. I miss the good old Intel vs. AMD days when there was a legitimate choice to be made by performance enthusiasts. Mind you, those days are long gone, but I will still hold a special place in my heart for my AMD Athlon64 3400 :)
In the past, I always advocated for, and employed AMD chipped systems. I was once burned by my advocacy when I lost several AMD mobos after they all got fried!
This was a contributory event to my getting fired, though a poorly written application was partly responsible. My employer could not listen because other AMD systems survived. They did because they were to be running the application next.
What is the experience of slashdotters using these systems? Do they still consume lots of power or overheat?
that had to google "inflection point"? From a marketing standpoint it might be good to have a CEO who isn't an engineer :P.
or a CEO who picks up a word or phrase from an engineer and thinks, 'Hey, that sounds good, I'll use it in my next meeting or press statement!'
A feeling of having made the same mistake before: Deja Foobar
This is why it's good to have a background in math, even if you're not employed in an STEM field. All sorts of processes can be described in mathematical terms, knowing what those terms mean helps you understand the world better. People often say "calculus? I'll never use that after high school!". But the truth is, I use my calculus education every single day without ever touching an integral or derivative.
Give me Classic Slashdot or give me death!
Let's say AMD is planning - or thinking about, at least - stopping the manufacture of x86 processors. What's a responsible company spokesperson going to say? "Yes, we're working on an exit strategy and are hoping to be out of the business by 2014" - does anyone believe that would be stated? If it was, their x86 business would tank immediately, and all employees working on x86 now would update their resumes and get while the getting is good.
Several years ago, we had an important faculty member accept a dean-ship at another university. The lead time was going to be a bit more than a year. In the meantime, this faculty member still had research projects going full bore. So what did he do? He told his staff that the research projects were going to continue, and would remain at our university for the foreseeable future. Guess what happened a year later? Yup - the "foreseeable future" he spoke of 12 months before turned out to be almost exactly 12 months long.
#DeleteChrome
"Our strategy is to accelerate our growth by taking advantage of our design capabilities to deliver a breadth of products that best align with broader industry shifts toward low power, emerging markets and the cloud."
We will continue to make chips for servers, and low end crap. We can't compete with Intel for the consumer market in the short to medium term, however we are still relevant in business circles.
Consumers prepared to be gouged by Intel as soon as they figure this out. Also other than to just "say it" this has been the truth for some time, years in fact. I don't know if it is AMD stumbling or Intel just continuing to hit home runs, but there hasn't exactly been a whole lot of competition since the days of the ye old Athlon 64 series of processors. Ever since Intel came out with the Core 2 Duo, AMD has been unable to come up with an answer. Perhaps it had something to do with diversifying by buying up ATI, diverting capitol or focus away from core business. Ironically the AMD/ATI brand of video cards has a better reputation than the AMD CPU division, if only my opinion...
My understanding is that Radeon cards are still competing neck-and-neck with Nvidia's offerings these days, especially per-dollar. I may be mistaken, though, as my video card is still an 18-month-old ATI Radeon 5850 (back before Nvidia even had a DirectX 11 card on the market, and before the AMD-ATI buyout), which can still play everything I've thrown at it on full settings at 1920x1080.
Even if their CPUs are lack-luster (even at the lower price point, it would seem, where they used to be quite competitive), they may be able to survive mostly on the GPU market without too many troubles. For a while, at least.
Indeed. Which is why words and phrases like "pushing the envelope" and "quantum leap" are so often used wrong, and marks the CEO (who reflects on the company) as a dummy.
I just upgraded my PC from a Intel E6600 to a AMD Phenom II X6 1100T. I chose AMD, for one reason. How the heat sink / fan attach to the motherboard.
I have dogs, and kids and my PC doesn't reside protected under a desk. It gets bumped all the time from them playing and those stupid plastic plug brackets that Intel uses to attach the heat sink and fan to the motherboard were absolute garbage. Someone would bump my PC and the heat sink would hang off and cause the CPU to overheat. Not to mention after re-attaching it a few times, the knobs break and you have to buy a new heat sink and fan.
At least AMD still uses the clamp style that clamps to the socket. There is no way that is going to come off unless you intend to take it off. I won't buy Intel again until they re-design how the heat sink attaches.
When people say "x86" they usually include the 64-bit extensions in this category. Every x86 CPU made today, whether from Intel, AMD, or even Via, supports the AMD64 extensions.
Every x86 CPU made today, whether from Intel, AMD, or even Via, supports the AMD64 extensions.
Some of the netbook Atoms didn't a year or two back; isn't that still the case today?
When we contacted Silverman, he confirmed that the original statement has been taken somewhat out of context and provided additional clarification. 'AMD is a leader in x86 microprocessor design, and we remain committed to the x86 market. Our strategy is to buzzwords buzzwords buzzwords buzzwords buzzwords buzzwords.'
x64 is an extension to x86. What we need a a whole new class of computers designed and built for 64bit architectures. But that calls for a complete redesign of the most popular OS and probably MOBO architecture as well.
The problem is, who would want to do that?
Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
I love you guys but recently only have been buying Intel i5 and i7 because your Math coprocessor still stinks badly compared to Intel. For video compression and really heavy maths, I really wanted to use your 6 core processors, but they were slower than the 4 core i7 I bought instead.
Give me a 6 core that runs like a raped ape and has a really good math coprocessor and I'll be back. give me an 8 core that can also do multi chip on the same motherboard so I can build a 16 core for a cheap price, and I'll be back with a whole lot of friends.
There are lots of us that actually do real computing that has really heavy math. I know you guys can do this.
Do not look at laser with remaining good eye.
>are so often used wrong
Oh the irony. :)
(grammar: "... are so often used incorrectly, ...")
Corporation, n. An ingenious device for obtaining individual profit without individual responsibility. - Ambrose Bierce
Intel is pushing forward because it's beneficial to them at the moment not to rest on their laurels.
AMD is underperforming, yes, but not so much that Intel is given any real leeway to slack off;
That is to say, if the i5/i7 lines were only a 5% increase over C2D performance for 1/3 higher price, AMD would have destroyed them, so while AMD hasn't been "real" competition for Intel for quite some time now, they've been good enough to keep the industry trudging along.
If AMD outright left the market, there would be absolutely no incentive or real immediate threat necessitating actual improvements until desktop ARM chips actually started getting established.
"x86" in this context means desktop x86 chips, x86_64 chips and AMD64 chips.
x86 is a subset of x86_64, or said another way, x86_64 is an extension of x86, just like SSE or MMX but with a lot more new instructions and hardware requirements.
It sounds like everything you do is x86. The alternative to that architecture isn't x86-64, it's Itanium-64 or ARM or any of the various big iron RISC chips.
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
He just read, "The Innovator's Dilemma," by Clayton Christensen. He sees new innovation in x86 chipmaking as having diminishing returns, making the entire architecture susceptible to other architectures and competitors where new innovation still provides increasing returns.
I've been a /. reader for the past 10 years and your ramblings are the most useless post I've ever encountered here. 99.9999% (all except you) of the readers here understand that AMD will not abandon the 64-bit arch. Btw, the corrrect term is x86-64..
Shame- I usually support the underdog- and always wanted AMD to be able to run Intel neck-and-neck.
Nowadays though AMD seems to stand for A Mediocre Design
I hope they can recapture their mojo and challenge intel again- if for no other reason than to provide a lower pricing incentive to intel.
"That's the way to do it" - Punch
When the Pentium 4 came out, it was frequently called the "7th generation", but it was never called the 786 or 80786, either formally or informally
But they are all x86 compatible, because they can all run code compiled for 8086, 80186, 80286, 80386 and 486 processors.
My new hobby will be referring to processors as having x87 architecture, as a distinction to indicate they support floating point instructions.
People do refer to x87 when talking about the FPU on x86 chips. It's commonly used when differentiating it from SSE - modern compilers will emit SSE instructions instead of x87 ones unless you specify a backwards compatible target architecture (PII or earlier).
I am TheRaven on Soylent News
Some of the netbook Atoms didn't a year or two back; isn't that still the case today?
As far as I can tell, the N270 (Diamondville series) was the last Atom that didn't support 64-bit. A quick Google search indicates that Intel hasn't officially discontinued it, but it seems to be almost impossible to find any new products that contain one. Newegg doesn't have any netbooks using this, though they do sell a 10-pack of Intel Atom N270 motherboards. Since they don't sell individual units, I assume these may be surplus stock (and probably intended for use in embedded systems). Even if the Diamondville series is not officially discontinued, it's definitely on its way out.
It will be deployed in 2014:
http://www.eetimes.com/electronics-news/4230160/ARM-unveils-64-bit-architecture
"Indeed, the first processors based on ARMv8 will only be announced sometime in 2012, with actual server prototypes running on the new architecture expected in 2014."
"Inflection point" is sometimes used more broadly and metaphorically to mean something like, "a time of drastic change (or potential drastic change)." He's not using it in a technical sense.
We could just breathe life back into the Alpha architecture.
As a matter of fact, I believe that MS supports it right up until Windows 2000 (multiple RCs, no release).
I'd love to have an EV12 processor in my next machine. ^_^
I am John Hurt.
Because the processor (theoretically) supports the x86 instruction set?
I am John Hurt.
All netbook Atoms do support it now. The early N270 and N280 were the ones that didn't. Then the miniaturized Atom lines (Z, E) apparently lack it.
Precisely. It meant "the point where AMD goes from a desktop chip maker that also makes mobile chips, to a mobile chip maker that also makes desktop chips".
No, engineers talk in acronyms. Marketing droids use terms like "inflection point."
"Wrong" is an adverb. Coincidentally, M-W uses the word "incorrectly" as a synonym.
The road to tyranny has always been paved with claims of necessity.
China needs a processor company, and even without AMD being leading-edge if their products are sold inexpensively there is a huge potential market worldwide.
"This post is an artistic work of fiction and falsehood. Only a fool would take anything posted here as fact."
This is why it's good to have a background in math, even if you're not employed in an STEM field. All sorts of processes can be described in mathematical terms, knowing what those terms mean helps you understand the world better. People often say "calculus? I'll never use that after high school!". But the truth is, I use my calculus education every single day without ever touching an integral or derivative.
Why baffle people with BS when you can use real language :)
The new line will factor into integral processes where derivatives will end-product quality!
Sounds like something I'd read in a Dilbert strip...
A feeling of having made the same mistake before: Deja Foobar
Indeed. Which is why words and phrases like "pushing the envelope" and "quantum leap" are so often used wrong, and marks the CEO (who reflects on the company) as a dummy.
wasn't quantum leap that tv show starring Jonathan Archer
This is why it's good to have a background in math, even if you're not employed in an STEM field. All sorts of processes can be described in mathematical terms, knowing what those terms mean helps you understand the world better. People often say "calculus? I'll never use that after high school!". But the truth is, I use my calculus education every single day without ever touching an integral or derivative.
I was thinking about setting up a Reggie Pole for my wife to dance on.
They don't beat the comparative Intel chips at said tasks anywhere near well enough to justify the heat and cost tradeoffs.
The Bulldozer release showed AMD's commitment to low-end computers. I opted for Phenom II over Bulldozer because of the architecture changes. My next setup, if not another Phenom II, will be Intel.
I object to power without constructive purpose. --Spock
wow
I object to power without constructive purpose. --Spock
I thought the last x86 processor produced was the Pentium Pro
Assuming I understand your pseudo-purist definition of "x86" correctly, one minor pedantry...
The Pentium Pro wasn't the *last* of Intel's "real" x86 processors, it was the *first* of the RISC-with-x86-wrapper (*) designs that make up all chips today. AFAIK the original Pentium line was the last.
(*) Some have claimed that the core isn't actually that RISC-like; the point here is that it's not native x86.
"Slashdot - News and Chat Sites Deviant". (Click "homepage" link above for details).
I know:
http://slashdot.org/comments.pl?sid=2543280&cid=38161632
x64 is an extension to x86. What we need a a whole new class of computers designed and built for 64bit architectures.
How would "[designing] and [building]" a computer "for 64bit architectures" differ from what's being done now?
But that calls for a complete redesign of the most popular OS
I know of no desktop, notebook, or server OSes that would need "a complete redesign" to work on 64-bit architectures - they already work on them; presumably whatever would make the "whole new class of computers" different from the 64-bit computers being sold now is what would require that "complete redesign".
The Pentium Pro wasn't the *last* of Intel's "real" x86 processors, it was the *first* of the RISC-with-x86-wrapper (*) designs that make up all chips today. AFAIK the original Pentium line was the last.
The poster to whom you're responding said
in which case the last "x86 processor" was the 80486 - the Pentium wasn't sold as the 80586, it was sold as the Pentium.
And, given that the internal microops in Pentium Pro and later are not exposed to programs running on the processor, it's irrelevant to a discussion of whether those chips are "real" x86's or not. About the only processors where the question could reasonably be raised as to whether they're "really" x86 or not were the Transmeta chips, where an x86 interpreter and x86-to-native-code translator were, presumably, written in or compiled to the native instruction set and ran from, as I remember, a reserved chunk of main memory.
Listen, 10 years have done nothing for you, your still an immature little pos who likes to troll others to make yourself feel better about your worthless life. Your posts are stupid and redundant. Where was your dumb ass when x64 rolled out? You seriously forgot the incompatibilities? I guess working at your tier 1 support job lets you escalate it to your many superiors, but still how f'in dumb can you be? Nobody is talking about abandoning x64 and btw and the correct term is not x86_64 it's x64 for short, nobody uses x86_64 except for argument or research purposes.
This faggot keeps trailing my posts and prolly has mod points on his real account applying against me, if I cared I'd report abuse, but I don't.
Seriously dude, gtfo off slashdot and go catch some sunlight or something, calling you a loser would be an insult to the word.
Marketroids talk about graphs and curves all the time. They will have cost vs. performance, cost vs. time.
Imagine a basic bell curve. You start at zero, which is no sales, as there has been a pre-release advertising campaign and the products have been reviewed before going on the shelves. Then they start to sell. The inflection point is where word of mouth has gone out about a product, and market demand is highest.
After competitors start bringing out equivalent products, the demand starts to slow down. The peak of the curve is when you have made the most sales ever in one week or month. Then demand starts to fall.
Eventually, demand will fall down to zero if they don't do something about the price. Then it's time to bring out a new advanced model, drop the price or offer a two-for-one deal.
You don't want to bring your next generation product too soon otherwise your developers get pissed off they that haven't had time to develop applications for the current generation, and your customers get pissed off that they are having to upgrade, or have just brought a current generation product. Like the Adam Osborne effect with "luggable computers".
Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
(Note; I posted the above comment, did not realise I was logged out at the time).
"Slashdot - News and Chat Sites Deviant". (Click "homepage" link above for details).
x86 is a subset of x86_64
No, it is not. The registers aren't even the same. You can run 32-bit code on it by flipping into the right mode, but that doesn't make it a subset of 64-bit mode.
Engineers try not to talk ever. Marketing won't shut the hell up.
You're talking Windows 7 x64, which is just x86 extended a bit. What I'm talking about is 64 bit Chip, not x86 with 64bit extensions, and all the baggage that comes with it.
Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
You're talking Windows 7 x64, which is just x86 extended a bit. What I'm talking about is 64 bit Chip, not x86 with 64bit extensions, and all the baggage that comes with it.
If "a 64-bit chip" is a chip with an instruction set that's not derived from an earlier 32-bit instruction set, why do we "need" such a chip? What problems does the "baggage" of 32-bit instruction set support that all x86-64 chips, all 64-bit PowerPC chips, all 64-bit SPARC chips, all 64-bit MIPS chips, all PA-RISC 2.0 chips, and all z/Architecture chips have, and that future ARMv8 chips will have (look for "Allows AArch32 applications under AArch64 OS Kernel" in the ARMv8 presentation), cause that are so severe that we "need a whole class of computers designed and built for 64bit architectures", with "64bit architectures" being those that lack that "baggage" (and thus meaning "IA-64" at this point, unless you know of a plain to resuscitate Alpha)?
And why would that require "a complete redesign" of Windows, rather than a rewrite of the low-level platform support code and a new compiler for the new architecture, given that, for example, Windows did support IA-64, and supported it as a 64-bit architecture (unlike its old support for Alpha with a 32-bit address space and 32-bit pointers)?
Limited Transistor counts and physical limitations of die size to start. Much of the current x64 chip is just to keep legacy x86 shit running.
Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
Limited Transistor counts and physical limitations of die size to start. Much of the current x64 chip is just to keep legacy x86 shit running.
[Citation needed]. At least going by area on the chip, a significant amount on a quad-core Sandy Bridge processor goes to graphics and memory and I/O access and to the shared L3 cache. How many transistors go to, for example, support for segmentation and references to AH and AL and so on? If you're talking about "legacy" in the sense of stuff x86-64 keeps around because it's in the x86 family, rather than its support for 32-bit code, that's another matter - but that's not an issue of being a chip with support for 32-bit code, it's an issue of being a chip with support for x86 CISC code. A reasonably clean 32-bit architecture could, I suspect, be extended to 64 bits without a lot of transistors lost to support for 32-bit code as well.
Given how ARM has been all conquering on portables & embeddeds and in a position to take on laptops, I'm not surprised that AMD is thinking along these lines @ all. If they were to drop 32-bit support from their CPUs, then the 64-bit only CPU would be a RISC CPU, from what I understand, and AMD would find it easier to match promised performances. As for Windoes XP, MS no longer sells it or supports it, and so XP only needs to support existing boxes. If it was just an issue b/w XP and 7, AMD could have dropped 32-bit compatibility already. I think Windows 8 is going to be the point where one sees support for CISC instructions being dropped.
Incidentally, does AMD have any IP rights to the Alpha? If it does, that's another processor it could resurrect, and offer in a range of bins - from low power to high performance. Slap the Android on it, and they'd have a viable platform for tablets & laptops, while w/ FreeBSD, they could even return to servers. I do think that they'd be incredibly stupid getting into the ARM market, overcrowded as that market already is w/ multiple suppliers. Or else, Intel would have stayed there w/ XScale & succeeded.
I disagree w/ the GP that Intel is forcing AMD to stay in the x86 business. First of all, anti-trust only means that Intel cannot deliberately cripple AMD in order to gain market dominance, but if AMD, for other business reasons, decided that the platform is a dead end and chose to abandon it, that wouldn't make it an anti-trust case. Besides, a credible argument can be made that there are still other CPU vendors one can choose - MIPS, SPARC, POWER, ARM. Also, if Intel is discovered manipulating AMD business decisions, does anyone think that that itself would come under an anti-trust scanner for being anti-competitive? Does anyone remember Adaptec being under any anti-trust scanner when they acquired Symbios, thereby becoming a monopoly?
Even at the worst of times, the x86 decoder was a few percent of the chips. Today it's more like 0.01% of the chip. The "cruft" instructions have generally been moved to microcode, essentially a software implementation inside the processor and take up no hardware at all. Keep repeating it but nobody who knows what you're talking about will take you seriously.
Live today, because you never know what tomorrow brings
Kudos for the troll. I barely watch the thing, and was about to correct you.
I know tobacco is bad for you, so I smoke weed with crack.
Dragon chip has x86 acceleration instructions, not a full implementation, and they take 5% of die space. Second, ever try to design a superscalar decoder for x86? The hardware implemented algorithms become large and complicated, but in the real estate budget it goes under the heading superscalar decoder, and not x86 support.
I know tobacco is bad for you, so I smoke weed with crack.