AMD Licenses 64-bit Processor Design From ARM

angry tapir writes "AMD has announced it will sell ARM-based server processors in 2014, ending its exclusive commitment to the x86 architecture and adding a new dimension to its decades-old battle with Intel. AMD will license a 64-bit processor design from ARM and combine it with the Freedom Fabric interconnect technology it acquired when it bought SeaMicro earlier this year."

Comments

Oh snap.

[Anonymous Coward]

This is going to change everything.

Re:Oh snap.

[Anonymous Coward]

An over-priced slow server, ARM will grow to dominate the market. The same way Intel's slow and over priced servers have become commonplace.

Re:Oh snap.

[Anonymous Coward]

Don't forget the power savings!

Re:Oh snap.

[SomePgmr]

An over-priced slow server, ARM will grow to dominate the market. The same way Intel's slow and over priced servers have become commonplace.

Well we'd try something else, but it turns out monkeys with notepads and crayons are even slower (and more expensive).

Biodegradable, though.

Re:Oh snap.

[Anonymous Coward]

RISC architecture is gonna change everything.

The fat lady is singing

[Anonymous Coward]

The panel discussion that accompanied the AMD news conference was absolutely painful to watch. The only thing I learned is how completely clueless the CxOs of the 'cloud computing era' really are. Seeing company officers from Dell, RedHat and Facebook drool allover themselves like that was yet another painful lesson that the fratboys of the world have turned the tech industry into their drunken biatch.

Re:The fat lady is singing

[Anonymous Coward]

Do you (or any other kind commenter) have a link to the video? I'd like to see what you're talking about.

Re:The fat lady is singing

[hairyfeet]

What's sad is how bad the former CEO fucked AMD by doing a total slash and burn on their engineering and R&D and pushing for cheaper automated layouts that simply don't cut it. The Athlon64 guys? GONE. The Cryix guys? GONE. they pretty much have their backs against the wall because the former CEO burned the fucking company to get a short term bounce, which I'm sure he cashed out on.

And anybody who thinks ARM will save them might be interested in some magic beans I have for sale, as ARM frankly doesn't scale very well and from the early looks ARM64 isn't gonna be really any better for power than the CULV Intel chips while having a HELL of a lot worse IPC. Frankly, and this is coming from someone who has been building AMD systems exclusively for awhile now and is still hanging onto AM3+ for all its worth, the only real selling point they had was "bang for the buck" but by burning R&D and killing Thuban the former CEO left them holding the bag without shit besides Bulldozer, which we all know blows too much power, is too damned hot, and frankly their octocores get stomped by Intel quads on IPC while using a third of the power.

I have to agree with the engineer in that link, they should have done the same thing Intel did with Core, go back to their earlier K8 designs and start from there just as Intel did with P3 mobile but now they just don't have the money or the time. I truly hope the Athlon64/Apple A6 chip designer they hired back can come up with a design to save the company because right now? Right now they really got nothing. Hell the former CEO even pulled the plug on Krishna, which would have been a sub 20w quad core bobcat, which is why all we're seeing now is minor speedbumps on a 3+ year old design. I swear they got fucked raw by bad management and I only hope they pull through. Maybe if they would have done this 4 years ago they could have the niche Nvidia now holds, but now? Its just not enough.

Re:The fat lady is singing

[Anonymous Coward]

It usually normally takes a day or so for replays to be posted but it should show up on the AMD Investor Relations Website (same site that hosted the live webcast).

Re:The fat lady is singing

[Anonymous Coward]

I completely agree (although the Cyrix guys weren't a part of AMD if I recall correctly, they're now Via).

I don't get why Via and AMD don't do any collaboration. Via seems to have decent CPUs and some pretty bright sparks in their CPU design division but they use fucking awful graphics chipsets. Or Via and Nvidia for that matter.

Re:The fat lady is singing

[lightknight]

It's been a while, but wasn't VIA responsible for the really screwy AMD chipsets that used to make people curse under their breathe?

Re:The fat lady is singing

[lightknight]

Indeed. The one order the CEO can give to save the company is this: "Magical turn-arounds for companies who have been f*cked only happens in textbooks and fair-tales; as such, all resources for CPU design will go into creating a Phenom III with 12 cores and PCI-Express 3.0 and an Opteron design which employs liquid cooling (for the short term), as we are going to give it a major Mhz boost on top of the extra cores / cache we are going to staple on."

Getting involved in the already overgrown ARM market shows nothing but lack of vision. "We're going where everyone else is going, that'll be profitable!" You are going to be *that* guy who shows up late to the party, and wonders why all the booze is gone. Seriously, how do you mismanage stuff this badly? You're a CPU company, and you come up with the brilliant plan that despite being a major competitor in the x86 market, you're going to fix things by buying an oversubscribed design for a CPU in a market that...recursion error.

Think of it being like Ford, not using its own resources to think up a new car design, but paying Honda to license it the design for the Civic. Things are either absolutely atrocious, like AMD's stock should be worth a Haitian penny right now bad and we just haven't been told anything, or somebody doesn't know what he's doing. Go get the old guys your predecessor fired, and bring them back for more money. Find the DEC guys, and offer stock options if you have to to get them on board. Then follow their advice. After a year or two of punishment, AMD will be back on firm ground again.

Re:The fat lady is singing

[unixisc]

The more surprising thing is AMD going w/ a totally new architecture - the ARM64 - instead of one of the true tried & tested RISC CPUs out there. They could have licensed SPARC from Oracle, or POWER from IBM, or MIPS from Mips, taken an existing CPU design, made a chip quickly, and then in subsequent iterations, build on that.. There would also have been existing software platforms ready for them - not just Linux or BSD, but also things like Solaris, or AIX.

Re:The fat lady is singing

[scumdamn]

You mean like how Ford and Mazda worked together on engine design? Or how they collaborated with Audi? The Mazda 3, Focus, and S40 shared the same platform.

Really, what was your point again?

Re:The fat lady is singing

[Massacrifice]

Or how they collaborated with Audi? The Mazda 3, Focus, and S40 shared the same platform.

You mean Volvo. Germans don't share the good stuff. Only Swedes do that.

Re:The fat lady is singing

[Degats]

It's been a while, but wasn't VIA responsible for the really screwy AMD chipsets that used to make people curse under their breathe?

Yup. I had one where only one of the 3 memory slots ran at advertised speed and the onboard audio somehow conflicted with the AGP port, which caused the system to crash whenever I tried to do anything 3D.

And they're still responsible for the audio on some of the cheaper motherboards. It's the worst audio chip I've ever heard, the mic input didn't work properly and the driver was terrible and unstable.

All in all, it seems VIA seem to be incapable of writing stable/working drivers or firmware and the hardware's pretty shoddy too.

Re:The fat lady is singing

[Anonymous Coward]

they pretty much have their backs against the wall because the former CEO burned the fucking company to get a short term bounce, which I'm sure he cashed out on.

Welcome to American CEO 101. This is why, in very large part, American has been in decline for some time now. CEOs only care about a return tomorrow, without any regard for the fact they are killing the company day after tomorrow. The important thing is profit tomorrow and golden parachute day after tomorrow.

Re:The fat lady is singing

[hairyfeet]

Nice to see somebody else sees what is going on. They need to go back to the K8, get on bended knee and bring the Athlon64 guys back, but that will take time. They've got the Thuban design that was getting damned near 100% yields, they also have the Magny Cours socket, put them together, take a 95w X6 and unlock the multiplier and put 2 of them in the Magny Cours socket for the gamers, and for those that stuck with AM3 give them an 8 core Thuban. This would bring up the IPC enough to keep them afloat while they get the chip guys that know WTF they are doing to go back to K8, just as Intel did with the P3 and Core, and build them something great.

But if they stick with Bulldozer/Piledriver/Steamroller they are just fucked, royally fucked. Its not a good design and I don't think there is any way to fix it, it just has too many shared resources and the IPC is just too shitty. When you have 8 fricking cores and can't compete with a 4 core of the other guy? The chip is shit. Hell I'm still selling the socket AM3 chips because their IPC is still higher than the average user will need while being a good value and not blowing through the power, they need to get back to that short term along with making a multi-chip module for Brazos, bolt 2 Brazos modules together minus the GPU on the second one and you'd have a sub 25w quad that is cheap and has decent graphics for HTPCs and netbooks/notebooks. Hell plenty of OEMs are still using Brazos because its a good deal, a quad would sell like hotcakes.

I truly hope they pull out of the death spiral they are in but ARM is EXACTLY the wrong move, as you pointed out they are gonna be late to the party, by about 4 years, with Nvidia already having the job of good graphics paired to ARM that would have been theirs. They really have to rebuild after being gutted by their CEO and that takes time, so going back to what worked and just bolting two chips together would at least give them a product to sell that could at least keep up with and in some cases beat i5 and the first gen i7.

Re:The fat lady is singing

[hairyfeet]

I actually have one of the new Via HD audio chips in my board and...its quite nice actually. I figured I'd have to swap in a discrete but since I already had a SB X-Fi given to me by a customer and other than Via audio the board was just a steal, 4 RAM slots, Xfire ready, 8 SATA, ton of USBs, etc for $45 I could always just ignore the Via sound but I have to say at least for this board somebody got it right as the sound is quite nice, no crashy driver or hassles.

Re:The fat lady is singing

[Anonymous Coward]

Getting involved in the already overgrown ARM market shows nothing but lack of vision. "We're going where everyone else is going, that'll be profitable!" You are going to be *that* guy who shows up late to the party, and wonders why all the booze is gone. Seriously, how do you mismanage stuff this badly? You're a CPU company, and you come up with the brilliant plan that despite being a major competitor in the x86 market, you're going to fix things by buying an oversubscribed design for a CPU in a market that...recursion error.

Actually, it might be. Imagine them taking the K8 design, rip off the am64 decoder and attach an ARM decoder. Suddenly you've got a very fast ARM64 chip. No other company can compete for the high-power end of the market. With Intel attempting to move in on ARM's market, ARM might even pay AMD to ensure there is a high-power ARM chip readily available on the market. Realistically though, no this doesn't really bode well.

Re:The fat lady is singing

[Anonymous Coward]

You are confusing old Cyrix and new Cyrix, Via owns the Cyrix trademarks and patents but their design division is actually Winchip, the Cyrix guys went to AMD, the Athlon guys were the old DEC Alpha team.

Re:The fat lady is singing

[Anonymous Coward]

Imagine them taking the K8 design, rip off the am64 decoder and attach an ARM decoder. Suddenly you've got a very fast ARM64 chip.

Microprocessor design is not as simple and cleanly modular as you imagine it to be. Similar things have been attempted before and the results have generally been dismal. The first-gen Itanium supported both IA64 and x86 instructions, but x86 performance was horrible. There was also a PowerPC design which was going to have an additional decoder to support x86, but it never even made it to market because it was so bad.

The execution backend simply is not as decoupled and abstract from the native instruction set as you might think. The micro-ops it executes are designed to efficiently implement the native ISA, not any generic ISA. In fact, designers generally aim for the common case being one instruction translating to one micro-op, because that's the most efficient way of doing things. Many other design choices are driven by the native ISA too.

The other thing is that different ISAs have different rules about memory ordering, misaligned accesses, and a ton of other tiny details. Making an execution backend designed for one architecture's rules do the right thing for another's rules isn't always easy without sacrificing efficiency or performance.

Not going to help

[Anonymous Coward]

Next thing you know, AMD is going to give up on making a really good ARM processor too, and focus on the low-end, no margin market where Samsung and Apple will crush it with 3-generation-old designs.

And they'll call a dual-core part at 1.2 GHz the "AMD Fusion G8 Quad 3800+".

I can't wait!

Re:Not going to help

[Anonymous Coward]

I wonder if they'll migrate to producing 68000s next.

AMD might stand a chance

[MrDoh!]

They're losing the x86 battle, even with great chips, ARM might give them a huge boost, good for them to expand business.

Re:AMD might stand a chance

[CAIMLAS]

If AMD can push their engineering into ARM quickly, they might not only stand a chance but they might dominate fairly quickly, I'd think. They're not on par with Intel on die size, but IIRC they're pretty close - that knowledge is certainly applicable.

Remember, they've got good GPUs already. A lot of what they tried to do with the Mobility and later generations were very "ARM-like" already, it just didn't exactly work due to x86 limitations. I'd think they've got a pretty good chance overall. (If anything, it's a big market. Tegra# are really pushing NVidia along, after all...)

Re:AMD might stand a chance

[haruchai]

Fabrication has always been their Achilles heel. If their fab capabilities were only twice as efficient ( yes I know that's a lot but Intel is WAAAY above that), they would have little to fear from Chipzilla.

Re:AMD might stand a chance

[toejam13]

I'm sure this is just AMD hedging their bets against multiple processor ISAs. There are places where ARM is better than x86/x86-64, so it makes sense to try and dominate those niches. It falls in line perfectly with AMD being a less expensive alternative to Intel.

Given that Intel is trying to wind down its StrongARM line it inherited from DEC, AMD may see the ARM line as a place where it can finally be top dog. It has the expertise to give Broadcom, TI and Samsung a run for their money.

Taking a really big drink from the hypothetical Kool-Aid, I could see ARM64 processors being used as x86-64 replacements in palmtops and laptops. There are a couple of x86 to ARM translators on the market, which would solve the binary compatibility issue. I used FX!32 back during the NT4 and NT5beta days with my DEC workstation, and it made emulated binaries about 90% as fast as native. With advances in JITC translators and a cleanup of the x86-64 ISA to make it closer to meeting Popek and Goldberg virtualization requirements, I could see a good modern translator being 95+% as fast as native x86-64 code.

I've been expecting Apple to churn out a Power Book with an ARM processor and a binary translator. They did it with m68K -> PPC and PPC -> x86, so I wouldn't be surprised in the least to see x86 -> ARM. Now imagine it with an AMD ARM64 SoC at the heart of it.

Re:AMD might stand a chance

[lowlymarine]

An ARM processor doing binary translation for x86 would be like trying to tow an 18-wheeler with a Tata Nano. ARM may be low-power, but it's also...well, low-power. Even older Core 2 chips wipe the floor with ARM's latest and greatest from a performance standpoint.

Re:AMD might stand a chance

[hairyfeet]

Uhh...what great chips? The Thubans were good, but everything based on Bulldozer just blows through power while having terrible IPC, thanks to having shared integer and floating point units. If they were to be honest the "modules" would be treated as single cores with hardware assisted hyperthreading, because the benches show that is a hell of a lot closer to what they are than to true cores. Hell since the release of BD they don't even have a single slot anymore on Tom's Hardware "Best Gaming CPU" list whereas they used to pretty much OWN everything under $200. Hell look at how badly their new chips rate compared to even their old chips, with not only the X6 but no less than TWO of the X4s, the 980 and 955, scoring better than their new FX 8120. So I'm sorry, this is coming from someone who has been building AMD exclusively since i heard about the OEM bribery, but the new chips? Just not good.

And sadly ARM isn't gonna save them either, they are too late to the game and from the looks of it ARM simply isn't gonna scale while keeping its lower power budget. Just look at how companies like Nvidia, that have been sinking a ton into ARM, are having to use ever more cores to get the performance up, it just doesn't scale. And since Intel has the fabs they can get to the lower sizes quicker, and their chips are frankly getting lower powered all the time. A 55w Ivy will frankly curbstomp a 125w Piledriver and with servers while there are some loads you can run without the IPC frankly there are a LOT more loads where you'll need that IPC and AMD just doesn't have it, and with electricity costs and cooling costs? It really don't look good for AMD, damned I wish it weren't true but it is what it is, AMD is in REAL bad shape right now.

Re:AMD might stand a chance

[Sable Drakon]

Great chips? AMD hasn't have a truely great chip since the AthlonXP and maybe the Phenom 2. And even then, their Intel counterparts gave them a beating.

Re:AMD might stand a chance

[Anonymous Coward]

AMD no longer has a fab of their own, as of two years ago(?). I believe they are currently using TSMC for most of their production.

Re:AMD might stand a chance

[Sable Drakon]

AMD can't even keep up with a single competitor, so just how are they going to compete against 3+ vastly more nimble companies in the ARM space. AMD is making yet another last gasp at trying to stay relevant.

Re:AMD might stand a chance

[Dahamma]

Note the article says ARM *server* processors. In that market, GPUs are totally irrelevant, power usage is secondary to performance, and price of the CPU is a distant third.

Any ARM CPU is at least an order of magnitude behind the current x86-64 server CPUs. Not to mention the additional work required to support multiple ARM CPUs on a motherboard, and even convince the major server manufacturers to build an ARM-based server in the first place. Good luck AMD, though you won't need it since even luck won't help you here...

Re:AMD might stand a chance

[TheEyes]

Maybe the new direction is going to be heterogeneous computing. We're already seeing AMD and Intel combine x86 and a GPU on one die; maybe AMD will try to combine everything and have a couple of ARM cores for low-power tasks, a couple of Bulldozer modules for more intensive tasks, all combined with their GPU.

Re:AMD might stand a chance

[symbolset]

Bingo.

Re:AMD might stand a chance

[ducomputergeek]

I just find it ironic that Apple could very well be going back to RISC after not even a decade of being on x86. Even more ironic given the amount of work Apple contributed with Acorn back in the 1980's.

A MacBook with ARM chip wouldn't surprise me. After all the iPad, iPhone, and iPod are all arm chips already.

But then again I bought this MBP earlier this year as well as parallels and Windows 7 Pro because I do enough development work on multiple platforms that i do need to test against windows as well as use Visual Studio now and then. And this gives the flexibility to stay in Unix 90% of the time and then boot bootcamp using either parallels or even reboot directly into Windows 7 if I need to do some heavy work in VS.

Re:AMD might stand a chance

[Max Littlemore]

Note the article says ARM *server* processors. In that market, GPUs are totally irrelevant, power usage is secondary to performance, and price of the CPU is a distant third.

That is as has been, but I'm wondering if this is not a strategic move on their part. Perhaps They are thinking of large clusters of low power ARM cores that kick in as the workload demands with some kind of clever way of sharing resources (Freedom Fabric?). With the global political landscape the way it is, that could be an important point of difference.

Reducing energy consumption is now the "in thing" and will continue to grow in purchasing decisions as financial incentives to reduce carbon emissions grow. If a server can be run using a single voltage PSU instead of needing different voltage on different rails, there are likely to be energy savings over x86. If that server can idle at less than a Watt and then ramp up in small increments as demand requires, that might also yield an overall advantage.

Sure, for serious continuous load applications, it's probably not the best, but for a lot of cloud type applications I can see this as being useful for renewable energy supplied server farms.

I'm just speculating of course, and I fully agree GPUs are irrelevant, but I think the idea that power usage is always secondary to performance has reached it's use by date.

Re:AMD might stand a chance

[Max Littlemore]

I wouldn't want to tow an 18 wheeler with a Nano, but I also wouldn't want to deliver a small package to an inner city address with an 18 wheeler. Horses for courses (or pidgeons, horses and elephants for their respective courses if you will).

Re:AMD might stand a chance

[ppc_digger]

Note the article says ARM *server* processors. In that market, GPUs are totally irrelevant, power usage is secondary to performance, and price of the CPU is a distant third.

Power usage translates directly into heat, so if you have a CPU that takes 10x less power you can cram 10x more of them onto the same server.

Re:AMD might stand a chance

[Anonymous Coward]

My memory might be failing me, but wasn't AMD the reason that Intel stopped the pointless GHZ bumping? And didn't they beat Intel to a proper multicore architecture? And weren't they before Intel when it came to the APU?

Not to mention that Intel's graphics are still pathetic compared with nVidia and AMD.

That doesn't really sound like a company that's grasping to stay relevant, that sounds like a company that's pretty much defining the course of future technologies. Even if at each step they aren't as fast.

Re:AMD might stand a chance

[Dahamma]

That is as has been, but I'm wondering if this is not a strategic move on their part. Perhaps They are thinking of large clusters of low power ARM cores that kick in as the workload demands with some kind of clever way of sharing resources (Freedom Fabric?). ...
If that server can idle at less than a Watt and then ramp up in small increments as demand requires, that might also yield an overall advantage.

After 20 years of Wintel I finally caved to try a Mac. The new MBP Retina is insanely fast CPU-wise for the same battery use over my old laptop (not quite as much GPU-wise vs my Windows desktop, but we are not talking GPUs). And when I'm not doing much it uses very little power. That's with a 4-core (hyperthreaded to 8) i7.

Intel has already come up with solutions for standby & lower power scenarios - when doing nothing there isn't much difference. When doing "a little bit" ARM definitely wins, and when loaded the ARM isn't even in the equation since it can't come close to the loads of the high end x86-64 CPUs. And it also depends on the application, as you say. For large data-intensive applications the power draw of the motherboard, RAM, HDDs, etc. surpass the CPU anyway so it's diminishing returns.

Anyway, I do think there is a market for low power servers. But it's currently a really low margin, barely profitable market, and will be until power costs are higher than hardware costs for the advantages it brings. And does anyone really think Intel is not capable of adapting to the market at that point to maximize their profits yet again?

Re:AMD might stand a chance

[lightknight]

Not without the top-level chip designers their previous CEO nuked. They may be the Chicago Bulls in name, but the player lineup has changed.

Re:AMD might stand a chance

[lightknight]

Then they should take some of their idle Sales / Marketing / Business guys, have them fly over to {country}, and let them spend some time charming the other foundries into not only giving them the capacity they need, but doing so at an excellent price. At the very least, it will give them something to do.

Re:AMD might stand a chance

[TheRaven64]

Given that Intel is trying to wind down its StrongARM line it inherited from DEC, AMD may see the ARM line as a place where it can finally be top dog

Intel isn't trying to wind this line down, they sold it outright to Marvell two years ago. Even then, they were pretty anaemic. XScale was the P4 of the ARM world: twice as high a clock speed as everyone else but a much lower instruction-per-clock. It's an ARMv5 implementation, which seems painfully archaic today (especially given the lack of FPU, which even most ARMv6 implementations have).

Re:AMD might stand a chance

[makomk]

The Thubans were good, but everything based on Bulldozer just blows through power while having terrible IPC, thanks to having shared integer and floating point units. If they were to be honest the "modules" would be treated as single cores with hardware assisted hyperthreading, because the benches show that is a hell of a lot closer to what they are than to true cores.

Errrm, all of the integer units are dedicated and the shared floating point units still give each core as much floating-point resources as on the previous generation of AMD chips even if every single core is using floating point 100% of the time. If AMD hadn't screwed up on the engineering side, it'd be a really great design.

Re:AMD might stand a chance

[Paul Jakma]

StrongARM? Didn't intel sell that to Marvel years ago?

Re:AMD might stand a chance

[samoanbiscuit]

It's been said before on this thread, but I'll say it again. AMD remaining solvent while competing against Intel for 30 years is a lot more impressive than most people realize, especially considering they competed using Intel's own ISA. It's too soon to tell now, but it's reasonable to expect that AMD (being in Intel's weight class) could plausibly compete with most of the current ARM manufacturers. I'd certainly expect their 64 bit server chip efforts to be a lot more interesting than what the cell phone chip makers have been putting out from a performance perspective.

Re:AMD might stand a chance

[funkboy]

This is almost certainly for a SeaMicro-based architecture. The GPU might be mildly irrelevant in this market today but will continue to gain importance as more tasks transition to being executable via OpenCL & its cousins.

What you are looking at is a small box densely packed with lots of cores. Another flavor will likely come as a box with a few weak ARM CPUs used to control a large quantity of GPUs for HPC applications.

The thing that will make or break ARM in a SeaMicro style chassis is whether they can get a successful OpenStack port to it done in time for the launch. My guess is that they will, as OpenStack development is going gangbusters at the moment.

Re:AMD might stand a chance

[drinkypoo]

Note the article says ARM *server* processors. In that market, GPUs are totally irrelevant, power usage is secondary to performance

Begging the question. Is power usage actually secondary? Not for many kinds of workloads, which are storage-intensive. For SOME servers it doesn't make sense. For OTHER servers, it clearly makes sense; people are already using ARM-based servers. Perhaps you should consider a little self-education.

Re:AMD might stand a chance

[drinkypoo]

Given that Intel is trying to wind down its StrongARM line it inherited from DEC, AMD may see the ARM line as a place where it can finally be top dog. It has the expertise to give Broadcom, TI and Samsung a run for their money.

Unless AMD has hired some of the bright names of ARM (and maybe they have, I've not been following such) they have essentially zero chance to come up to speed on ARM quickly enough to challenge any of the entrenched players any time soon.

Taking a really big drink from the hypothetical Kool-Aid, I could see ARM64 processors being used as x86-64 replacements in palmtops and laptops

Stop seeing that. It's not really plausible. There's no reason to do it, either, and there never will be unless ARM kicks x86's ass sometime in the nebulous future.

Re:AMD might stand a chance

[drinkypoo]

If AMD hadn't screwed up on the engineering side, it'd be a really great design.

I thought Bulldozer's reason to exist was to enable high clock rates. That doesn't seem to have panned out. Is that what you're talking about, or something else?

Re:AMD might stand a chance

[neokushan]

The new MBP Retina is insanely fast CPU-wise for the same battery use over my old laptop

You're comparing old technology to new technology and proclaiming the new technology to be better. I am shocked.

Re:AMD might stand a chance

[Type44Q]

They're not on par with Intel on die size

This statement may have been true... back when they actually owned some fabs. :)

Re:AMD might stand a chance

[Anonymous Coward]

Note the article says ARM *server* processors. In that market, GPUs are totally irrelevant, power usage is secondary to performance, and price of the CPU is a distant third.

The Jaguar supercomputer is being upgraded into the Titan supercomputer, by adding GPUs. Power consumption is one of the main drivers here. A 64-bit ARM could really flourish in that space.

Re:AMD might stand a chance

[Kartu]

I think 64 bit ARM doesn't have to be a "server processor". We already have smartphones with 2Gb of RAM. It won't take long until 32 bit CPUs run into memory addressing problem.

Re:AMD might stand a chance

[Kartu]

So Athlon 64 that wiped the floor with Intel's offering in all regards, doing more and consuming less, doesn't count as "since Athlon XP" or doesn't count as "great chip"?

Re:AMD might stand a chance

[Anonymous Coward]

I think AMD only exists because Intel let them. Intel knows it would have antitrust issues if AMD were to die, so it's in Intel's best interest to have a barely-solvent competitor. I'm sure Intel could price AMD out of existence if they wanted to.

Re:AMD might stand a chance

[drinkypoo]

I'm sure Intel could price AMD out of existence if they wanted to.

Because I'm not familiar with Intel's budget, that makes me wonder if Intel can exist at all without being able to throw around massive piles of cash.

Re:AMD might stand a chance

[Mashdar]

power usage is secondary to performance

Performance per watt is the real key for most server applications. It becomes a question of how much crap can you fit in a room before you can no longer cool the room. So, to be fair, performance and power usage are equally weighted. It's also very hard to compare current ARM products with current Intel x86 products because of the differences in process. The Intel process advantage is huge regardless of whatever you do with the architecture.

Not to mention the additional work required to support multiple ARM CPUs on a motherboard

I also don't see why an ISA change would really require many changes in memory interconnects, etc.

Re:AMD might stand a chance

[samoanbiscuit]

Of course Intel could price AMD out of existence. But why would they? What could they possibly gain by doing that? A few years of shitty profits, then the courts stepping on their necks when AMD goes belly up. Intel has a history of doing vicious and illegal things when AMD has an advantage (bribing OEMs during K8 vs Netburst), letting their superior marketing and industry connections work for them when AMD is on par (Wintel coziness, Intel compiler being standard for so many proprietary apps, etc) , and taking the pedal off the gas when AMD is behind, which has been the case in the high performance X86 game since 2008 or so. ARM is putting pressure on Intel in the mobile arena, but not on the server, and not on the desktop/laptop, and they (Intel) might have a good answer for mobile parts soon, if their roadmap delivers what it promises.

Re:AMD might stand a chance

[Sable Drakon]

AMD remains alive for two reasons. Graphics cards, which are frankly the only thing keeping AMD afloat at this point, and Intel's sufferance. Intel allows AMD to limp along because they know they'd fact Anti-Trust if they were to just go forward and acquire/liquidate them. The ARM market is quite a bit more crowded and the players in it would have no such restrictions against gobbling up a much smaller company.

Re:AMD might stand a chance

[citizenr]

If AMD hadn't screwed up on the engineering side, it'd be a really great design.

I thought Bulldozer's reason to exist was to enable high clock rates. That doesn't seem to have panned out. Is that what you're talking about, or something else?

so 1GHz higher clock rates compared to x6 Phenoms doesnt count?

Re:AMD might stand a chance

[K. S. Kyosuke]

Except that servers and supercomputers are two different things. E.g., the former ones run banking applications, the latter don't. Somehow I don't think that AMD don't have black hole simulations in mind when they think of 64-bit ARM servers.

Re:AMD might stand a chance

[hairyfeet]

But the IPC is soooo terrible that you'd need a 2GHz boost to beat the X6 on IPC and the problem as always when it comes to bumping up the MHz is power and heat and the BD/PD/SR design is a power pig, not the thing to be when everyone is talking green and worried about rising power and cooling costs.

Lets face it...its AMD's netburst is what it is. Just like Intel they focused on only one metric and got a chip that OCs well but with such a low IPC that frankly on all but the absolute highest SKU you HAVE to OC just to get anywhere near a decent IPC out of the chip and the heat and power just go crazy when you crank the BD design. If you'll look at the link I gave in the under $200 chips frankly THREE of the previous chips beat the FX 8120, including not one but two quads...that ain't good. Frankly the only way they've been having BD beat the Thubans is by pairing FX 8 against Thuban X6 and even with a supposed 2 core advantage it can't get more than a few percent ahead and in many cases it falls by a few percent.

Anyway you slice it the Bulldozer design is just a bad call, it was approved by a CEO that did a slash and burn on R&D and engineering and expected computer layouts to save them and instead it gave them a power sucking hog.

Re:AMD might stand a chance

[Dahamma]

Comparing old technology to new technology and proclaiming it's better was the whole POINT, sarcastic dumbass.

Higher performance with lower power usage from Intel chips over time - hence "Intel has already come up with solutions for standby & lower power scenarios." It shows that Intel is perfectly capable of adapting to market demand, which with recent Macbooks and Ultrabooks includes power efficiency.

Re:AMD might stand a chance

[Dahamma]

Think about it - storage intensive applications use lots of server-grade 7200-15k RPM HDDs which along with the required cooling will dwarf the power usage of the CPU. Not saying there isn't a place for ARM-based servers at all, but your example was dumb. Self-education, indeed.

Re:AMD might stand a chance

[Lonewolf666]

It wiped the floor with the Pentium 4. But not much later, the Core2Duo from Intel showed up. Since that time, Intel has dominated the high end on desktop x86. The Phenom II was not bad, but it could never quite catch the best Intel CPUs.

Re:AMD might stand a chance

[samoanbiscuit]

What makes you so sure the ARM manufacturers are going to be able to "gobble AMD up"? AMD has been in the CPU business for a lot longer than most of the ARM chip makers, and they spent that time competing against a much larger and nastier opponent. Also, these ARM chip makers have little to no experience making non-mobile high performance server chips, which is what AMD has announced they're making. In my view the ARM market is more crowded with weaker competition, a good thing for AMD's chances, not a bad thing.

Re:AMD might stand a chance

[Sable Drakon]

Given how AMD has mismanaged itself into it's current state against Intel, are you sure that competition in the ARM market is inherently weaker?

Re:AMD might stand a chance

[drinkypoo]

Not saying there isn't a place for ARM-based servers at all, but your example was dumb. Self-education, indeed.

Here's a little more education for you: I didn't even give an example. HTH, HAND.

Re:AMD might stand a chance

[samoanbiscuit]

Given that AMD has committed to designing ARM chips for servers, something that it's managed to beat Intel at several times in the past, and it's ARM competitors have little to no experience with, are you sure it's ARM competitors are inherently stronger? There's always the chance AMD will fumble, but there's also a chance it can show the ARM makers what competing in the x86 market for 30 years can teach a company.

Re:AMD might stand a chance

[Dahamma]

Not for many kinds of workloads, which are storage-intensive.

Workload: variable. Storage intensive: specific example, implies lots of storage aka HDDs.

I suppose education won't help when you don't even read your *own* comments...

Re:AMD might stand a chance

[Sable Drakon]

Given how AMD's turned out over the last decade, I don't hold out much hope for AMD being able to hold their own in the ARM market. They may have beaten Intel to ARM, but AMD is a weak little fish swimming into the sharks. What AMD lacks is the ability to research, develop, and fabricate. They've been ridding themselves of the ability to function as a CPU manufacturer for the last 5 years, and you expect them to stay alive among competitors that aren't dependent on others to fabricate their hardware? That's a laugh.

Re:AMD might stand a chance

[samoanbiscuit]

What are you talking about? Only Samsung has it's own fabs as far as I am aware of, and TI has given up on the smartphone market. Most of them use TSMC, which AMD already has a relationship with since that's what manufactures their GPUs. Once again, I think you're seriously overestimating the ARM pool's strength. Qualcomm and nVidia, the two ARM makers making the most waves recently, are both fabless.

Re:AMD might stand a chance

[Anonymous Coward]

I believe they are currently using TSMC for most of their production.

That depends on what you mean by "most".

GPUs are exclusively TSMC, but that's because ATI always used TSMC before being acquired by AMD, and continued to use TSMC after being bought.

On the other hand, most AMD CPU production is still at GlobalFoundries (AMD's old fabs), especially the "fast" stuff like FX series CPUs. I think they have a few low end CPUs on TSMC now, but all the high profile stuff is GF.

There's a lot of inertia here. Major design projects can take years, and retargeting them at a different fab midway is not necessarily a great idea.

Re:AMD might stand a chance

[cheesybagel]

Intel's fabs weren't that efficient. It is just that Intel has a lot of fabs all around the world.

Re:AMD might stand a chance

[haruchai]

I know that Intel stomped AMD in number of fabs but I also thought that they were getting much better yields when the manufacturing process got below 65nm and particularly so at 32nm.

Re:AMD might stand a chance

[cheesybagel]

Yes AMD and other fabs using SOI (e.g. IBM) had manufacturing issues in the beginning. However AFAIK the problems were solved later on. One issue AMD had was that with only one fab available they had to be able to switch production for different products using the same assembly line without significantly hampering production. However if you look at other companies which don't use SOI like TSMC they also have had a lot of issues lately.

Fingers Crossed!

[jerquiaga]

I'm hoping AMD does something to stay relevant. If they were to leave the market (or effectively leave the market by selling super low volume), then there's nothing to keep Intel honest.

Intel

[Citizen of Earth]

Intel will be doing the same thing in 3... 2... 1... Just like missing the 64-bit era with Itanium, it is missing he mobile era with Atom.

Re:Intel

[NidStyles]

Intel, they just went for their license last week. AMD has had their license for 4 years now. Who want's to make bets on who is going to win this race? AMD has won all of the previous ones.

Re:Intel

[Anonymous Coward]

I wouldn't be surprised at all if Intel had a team working on ARM ISA designs as a contingency plan, but I highly doubt they'd transition to ARM unless x86 was facing virtual annihilation. They're well aware that if they start releasing ARM chips, the whole industry will much more quickly transition away from x86. There's no way they would willingly destroy their extremely profitable, high-margin x86 business.

Re:Intel

[asliarun]

Intel will be doing the same thing in 3... 2... 1... Just like missing the 64-bit era with Itanium, it is missing he mobile era with Atom.

What are you even talking about? Since when did Intel miss the "64 bit era" as you put it? Sure, Itanium was a failure and Intel sunk billions of dollars trying to make it work. However, Intel could afford that mistake and still continue chugging along. As things stand today, Intel absolutely dominates the 64 bit market. In fact, except for Intel, AMD, and the IBM Power chips, there is no other game in town as far as 64 bit is concerned, and in this market, Intel probably has 80% or 90% market share, and has the best performance and performance per watt numbers.

So, I'm not sure which 64 bit era you are talking about, and how Intel missed it.

As far as Atom is concerned, yes, Intel is struggling quite a bit. However, Intel is trying to scale down its power consumption while ARM is trying to scale up its performance. Sooner or later, the two shall meet and it will be a very interesting battle. I wouldn't write off Intel so soon yet. In fact, the upcoming Clovertrail based Windows 8 tablets should be a very interesting launch. Take a look at the Thinkpad Tablet 2 for example. It should be a very interesting tablet for corporate customers or for users who want x86 along with Windows 8 Pro along with 3G and LTE mobility and a full-size USB port and with 8-9 hrs battery life.

I'm not saying Intel will win or lose, and it needs be relentless in improving power efficiency to even be a viable alternative to ARM. However, to say that Atom has already lost the race is a bit premature.

Re:Intel

[Dahamma]

Who want's to make bets on who is going to win this race? AMD has won all of the previous ones.

I assume you are joking, right? It's not a sprint, it's a marathon. Being first to market means nothing, it's winning the market. And Intel is crushing the 64-bit processor market right now.

Re:Intel

[Citizen of Earth]

Intel went for IA-64 and it was a complete failure. Ultimately, it was forced to adopt the AMD-64 instruction set. That's what I mean -- Intel missed the boat and the 64-bit instruction set it uses isn't even its own. Since adopting AMD-64, it's dominated the market space. If it wants to get anywhere in the mobile space, it will need to fold its current Atom strategy and go all-out ARM. Until it does that, it's Itanium all over again.

Re:Intel

[ChunderDownunder]

One of the bugbears of the ARM platform is the absence of mature, complete FOSS drivers for the embedded GPUs. e.g. PowerVR (proprietary), Mali (lima), Tegra (proprieatry), Adreno (Freedreno).

I could see Intel going the other way - keeping ARM at a distance but licensing its HD Graphics GPU to SoC manufacturers at minimal cost on the condition that they use Intel's factories to fabricate them.

(Just speculating - have no idea what % of a Sandy Bridge CPU's power draw is due to the graphics core(s))

Re:Intel

[asliarun]

Intel went for IA-64 and it was a complete failure. Ultimately, it was forced to adopt the AMD-64 instruction set. That's what I mean -- Intel missed the boat and the 64-bit instruction set it uses isn't even its own. Since adopting AMD-64, it's dominated the market space. If it wants to get anywhere in the mobile space, it will need to fold its current Atom strategy and go all-out ARM. Until it does that, it's Itanium all over again.

Okay, I get what you were trying to say earlier. Fair point too - because AMD64 was a vastly superior design and more importantly, a vastly more pragmatic design compared to what Intel was trying to shove down people's throats. Goes to show what hubris can do.

I'm not 100% sold on your recommendation of Intel dropping Atom and adopting ARM though. x86 is still very attractive to corporate clients and others who value legacy support and enterprise support. Business upgrade cycles are often very slow, and the only reason why iPads have even made inroads in the corporate market (and they are quickly doing so) is because there were no x86 and Windows alternatives. That's again why I say that the new wave of Atom Clovertrail Windows 8 tablets would be a very interesting battle in the business sector.

The other alternative - which is Intel adopting ARM would be a bad strategic move for Intel. Or perhaps I'm being too conservative in my thinking and perhaps it is a really bold idea.

Re:Intel

[citizenr]

One of the bugbears of the ARM platform is the absence of mature, complete FOSS drivers for the embedded GPUs. e.g. PowerVR (proprietary), Mali (lima), Tegra (proprieatry), Adreno (Freedreno).

Yes, this must be the biggest obstacle OEMs face when deciding on a platform for new phone/tablet!

Future for AMD

[drhank1980]

I hope going with ARM is successful for them; maybe enough to get them to try to make something to compete with the Tegra in the mobile space eventually.

Yep, I knew it...

[NidStyles]

I called this three months ago after they announced that they would likely be going in a new direction. Not here obviously...

Re:Yep, I knew it...

[Anonymous Coward]

[citation needed]

Back to Imageon?

[mozumder]

Are they bringing back the Imageon line now? (ARM + ATI GPU on die?)

Or the past

[mozumder]

Consider that they used to sell Imageons (ARM CPU + ATI GPU), which they sold off.

Let's hope this time it works out for them. Power optimization is now important, unlike the Imageon days..

use cases?

[Aryeh Goretsky]

Hello,

When it comes to servers, I use comparatively few (a small lab with a few rack's worth that used for research projects) at work, so I'm wondering what sort of tasks these would be useful for? It sounds like they'll run RHEL and other Linux distributions, but even after looking at the second slide in this presentation, it's unclear to me advantage this would be to a a small business, or, in my case, a small department in a larger organization.

Is this new CPU/server line intended only for the enterprise? If so, what would the "trickle down effect" be for small groups like my own? Also, why would someone want to throw out their investment in existing hardware (including whatever talent they might have at programming and maintaining said hardware) for a design that's relatively proprietary?

Regards,

Aryeh Goretsky

Re:use cases?

[CAIMLAS]

Yeah I'm guessing this is geared more towards people who do the whole vendor support thing, and they've got a handful (or at least one) person dedicated to maintaining specific equipment (eg. linux servers vs. switches). Homogenous is key, but high thread count will also push ARM advantage here, because you could fit (say) 8 of these small systems with multiple CPUs each in a single 2U without much issue, and still leverage your SAN storage.

You'll probably see them in low-end "server" devices too, I imagine. Eventually. Unless you're running stuff that'll compile and run on ARM, it probably gives you no advantage unless it's worth recompiling/running it on something with a GPU. If Windows 2010 were available for ARM, running a terminal server cluster might be useful, for instance (depending on how big your dept is).

With AMD's experience in x86, if anyone can push ARM to the server/desktop it's them. If they can standardize on things like the 'BIOS' for their own products, I'm sure many smaller companies will follow. (You're already seeing a degree of standardization eg. with integrators like Samsung, but that's now where AMD is going with this, exactly.)

For SMB - say, IT shops with fewer than a dozen identical systems now, or special use cases needing GPU (eg. scientific computing), this is a non-starter, at least for now. Think high flung web shops, cloud computing, and the like.

Re:use cases?

[Anonymous Coward]

Webservers, if there's lots of cores and they have the ability to turn on or off cores to save power...see the Sun T2/T1.
There are some workloads that just need a core, not a great core, and they need to scale the cores/threads economically. ARM is good for that.

Re:use cases?

[TheLink]

And see how successful the Sun T2/T1 was. By the time they launched their crap Intel was beating them on performance/watt and performance/watt/$$.

If AMD are really going ARM for the server market it's desperate clutching for straws.

It has to be some other market or they are committing suicide.

Re:use cases?

[symbolset]

For IT shops with fewer than 24 typical servers, what AMD might do in 2014 is not relevant. You would not be interested in trying this thing until it was field proven for three years. Even if it arrived on time (not AMD's strong suit) and it was nerdvana, that's 2017 before you're racking it. More likely the first version is quirky and your pilot starts two years later. But let's say 2017, for giggles. A typical 2 socket rack server can now be configured with 32 2.7GHz cores, 768 GB RAM, and terabytes of SSD million-IOPs storage and it comes with dual 10Gbps FCoE.

By 2017 even with outstanding business growth your server needs will be met by at most three geographically separated VM hosts, or cloud hosts. You are quite literally done at that point, as server performance will outpace your business need faster than the server warranties expire. Most small businesses are already there. If the extreme growth in mobile performance continues the logarithmic path of the last few years, by 2017 you might be able to replace all those servers with a mobile device. In 2017 100 Gbit dual port CEE onboard NICs will be standard, and we'll be laughing about magnetic media's long reign. "Our 128GB drives were as big as a deck of cards and could do 140 I/Os per second on a good day - and we liked it."

I've been a nerd for a long time - since the late 70's. The pace of these changes is stunning and unprecedented. For most people the capacity of one server is already far beyond their need. The second and third are just for redundancy. And they don't cost much, in constant dollars, compared to the servers of yesteryear. The software licensing can still be a burden if you're into paying for permission to use software rather than (or worse in addition to) paying for software support, but that is a different issue. Frankly I never did understand that strange turn that software took for a while.

Re:use cases?

[Anonymous Coward]

The T line has actually sold rather well, and if you could link to a source where Intel was beating the T-series on performance/watt on the specific workloads they were designed for, please do share.

Re:use cases?

[Anonymous Coward]

Your post and website are straight out of 1990. Where have you been the last two decades Aryeh?

Re:use cases?

[Anonymous Coward]

If they really sold well Sun wouldn't have been in such financial trouble.

Of course they are better on specific workloads. The problem is they suck so badly at other workloads that's why Intel's offerings are superio. For the T-series friendly workloads Intel's stuff is not far behind, for other workloads Intel's stuff is much better. Just a little change to your workload and the T stuff sucks for you- e.g. if SSL acceleration somehow breaks, or the application/usage changes and somehow the instruction mix is now different.

How many corporations are going to buy T for the very specific workloads, and then buy Intel servers for the other workloads they have? Most will just buy Intel servers.

Re:use cases?

[cheesybagel]

Actually were it not the T2/T1 Sun wouldn't have a processor design they could sell in a server. T2/T1 was meant originally for entry level servers with only a couple of processor sockets and they worked well in that scenario. The UltraSPARC V, Rock designs for high-end servers flopped and that, compounded with a move towards Linux on x86 by the rest of the market, is what killed Sun. The T1 design was biased towards integer heavy multi-threaded applications such as web serving or running Java server programs but the later designs are better all round processors. The design concept was interesting at the time.

Re:use cases?

[cheesybagel]

They did sell well but only the T2+ and latter processors supported multiple socket motherboards. There is not a lot of margin selling in the low end server market. Sun's problem was that UltraSPARC V and Rock designs flopped and this left them with no high-end processor and no high-margin product line. The T-series designers were originally hamstrung to work on single socket processors in order not to overlap with the high-end product which never happened.

Can we see

[gcnaddict]

x86-64 and 64-bit ARM on the same chip?

I can see this being a remarkable selling point for Windows devices if both ARM and x86 code can execute on the same device without emulation.

Re:Can we see

[slashmydots]

When they control the heck out of any apps on any tablet or device, why bother to open it up to multiple types of apps?

Re:Can we see

[Ostracus]

Actually I see this eventually rendering "for X architecture..." irrelevant. The more important "for what OS" will be dealt with using VMs. The need for more cycles to pull all this off competitively will mean we finally find a use for all those "solutions looking for a problem" engineers have been dreaming up.

Re:Can we see

[drinkypoo]

I disagree. I think that would be foolish for customers. It would be great for AMD though, because people would be upgrading left and right (or overspecifying) to make sure they don't wind up limited by one or the other.

It would make more sense for AMD to finally invent a system which can take asymmetric processors linked via HyperTransport, so that you can plug one amd64 or ARM processor and then however many amd64 or ARM processors you like after that.

Re:Can we see

[Anonymous Coward]

I don't see the point. Presumably you could JIT ARM 64bit instructions to x86-64 and achieve reasonably good performance, the other way around not so much. What benefit would having both instruction sets bring when most legacy binaries are x86-32?

Re:Can we see

[gman003]

Why not x86-64 and ARMv8 on the same *core*?

Every x86 chip on the market is some secret, internal RISC design with an x86 translator in front of it. I do not believe it would be terribly difficult to redesign the translator unit to accept ARM code as well, although getting it to perform as well as x86 does may be challenging. With a decent design and some clever firmware, you could probably make it boot as either ARM or x86 depending only on a BIOS setting, and change cores on the fly.

Re:Can we see

[Anonymous Coward]

Sounds like you're reinventing Transmeta.

This is interesting

[banbeans]

x86/AMD64 is overkill for many server functions.
It will be interesting to see if chips appear optimized for different functions.
For example hardware sql accelerators or massive i/o for file serving.
Since many hardware raid controllers are nothing but ARM cores anyway it would be interesting to see multiple cores, some used as RAID controllers and some more advanced cores for the os and file serving with a 10GB lan controller all on one chip.
Add power, drives and Ram and have a killer file server.

Re:This is interesting

[NidStyles]

I will have to agree with that. It seems the way to go in the future anyways with Windows grinding out onto ARM cores now. Sure the *NIX crowd will be capable of doing this, but with the number of Win users on server side being able to run ARM based machines, it's telling that the market will be more flexible and apt to adopt technology of this nature.

Re:This is interesting

[fm6]

It's overkill if you have precisely one hardware server per function. That's becomming increasingly rare.Nowadays, a "server" is most often a VM that doesn't need exclusive access to the physical CPU.

Re:This is interesting

[Anonymous Coward]

I thought of this too
Two 10 GBE controllers and two FC controllers. use what you want.

Re:This is interesting

[bloodhawk]

The problem won't be the hardware, it will be the same problem that IA64 had, hardware is great but without the server software designed for it they are just expensive paper weights.

Re:This is interesting

[Anonymous Coward]

Fortunately just about anything that runs on Linux is available in an ARM binary too.

Re:This is interesting

[Anonymous Coward]

Much of this stuff lives and dies by vendor support, especially in the enterprise. if AMD want to make a viable market they will need more than just the current Linux ARM binaries.

Re:This is interesting

[Anonymous Coward]

"Add power, drives, ram"??? Think SOFTWARE, retard. Solid software to drive that. You are talking mainframe sophistication. Linux is not even in the same galaxy.

Educate yourself about what you can do with a mainframe. They use coprocessors everywhere, that's how they outperform anything else on Earth. But it doesn't work magically because you throw the silicon together and stir it in a pot. The level of software sophistication here is unknown to 99% of you.

And we wonder why the world doesn't work. It's because it's full of gullible idiots who will believe anything presented with their crippled-la-la-land view on just a fraction of the ecosystem needed to make something really work.

Re:This is interesting

[jabuzz]

Really I have an IBM Storewise and SVC that beg to differ.

Welcome to the club

[Taco Cowboy]

Welcome to the club, AMD !

Unlike the X86 community, there are so many more competitors in the ARMs camp - companies such as TI and Broadcom from USA, Samsung from Korea, Hitachi from Japan, Allwinner from China, which produces $7 ARM-based SoCs.

AMD, you can't even compete against ONE company in the x86 arena - Intel.

Are you sure you can complete against the whole slew of them, this time??

Re:Welcome to the club

[fm6]

Your facts are off two ways. First, going up against one big monopolistic company is a lot harder than going up against a lot of small ones. (Do you think it's easier to fight an elephant or a bunch of guys who are also fighting each other,) Second, they've managed to survive in the x86 market for 30 years. I think that counts as competing.

Re:Welcome to the club

[TubeSteak]

Second, they've managed to survive in the x86 market for 30 years. I think that counts as competing.

The OP has a point.
AMD has abandoned the high end CPU market to Intel.

AMD's brand new, 8-core, flagship CPU, is competing with Intel's 4-core i5 chip.
And despite being clocked higher, it loses to the i5 in almost(?) every single-core test.

I know AMD pioneered the multi-core field, but they've gotten left behind.

Re:Welcome to the club

[lightknight]

Indeed. I am trying to grasp, somewhat desperately, the events that must have taken place inside AMD headquarters when the CPU design team said they wanted to do hyper-threading. Having seen how badly Intel got knocked around when they did it, and the fact that for the price of duplicating a fair amount of the CPU, you are still only occasionally eking out a slight performance gain...and sometimes, a performance loss, their strategy doesn't make sense. What was so hard about welding two Phenom II X6's together, using the hyperlinks already present in the CPU design, and calling it a day? Knowing full well that Intel wouldn't be able to compete with that design (they've been core adverse compared to AMD), being happy that all of the cores were full cores (who'd complain?), and that they'd be a hot item for system builders everywhere. Sure, some of the gaming websites like to barf about how single-threaded performance still matters, on some games that no one cares about (the GPU, of course, mattering a lot more than the single-threaded performance of a CPU here), but to take the advantage of having 6 full cores, and trade it in for 8 half-cores...was this some idiotic attempt at market segmentation? Did some moron in a suit have a brain fart, and think "we can't have 12-core Phenom IIIs, it will cannibalize our Opteron server sales"? Fire his ass, and cut the strings on his golden parachute on the way out.

For the life me, I just can't fathom how they turned a major market advantage, with the CPU design practically on the design table already, with a popular and critically acclaimed design, and decided that f*ck it, we're doing so well here, let's go for a lobotomy, and compete on Intel's turd with an unproven half-assed design. Let's go from a full-core design that everyone complements, to some terrible half-core design that nearly killed Intel at some point. Seriously, who is commanding AMD such that they were in their nappies when the whole Intel hyper-threading business was going down (which every half-decent tech knows about), and how did they get boardroom approval?

The proper response, of course, was not the Business School of Failure's attempt at mandating some perverse product differentiation, which bears as much similarity to surgery as bludgeoning a person to death with a hammer, but through true, non-crippling differentiation. Phenom IIIs get 12-cores, and the latest SSE instructions + something that the boys down in the instruction lab cook up; Opterons get larger caches + more cores + special server instruction sets that mean something concrete, even if it means implementing hardware Apache threads; that's on top of the SSE3 stuff and so forth. Would companies buy Opterons over Phenoms if one had hardware accelerated support for web services over the other? I believe the survey would say hell yes.

As for the GPU stuff, the low-cost, low-power stuff is nice for chump change, but it's a fierce market with many competitors. What you want, what large companies no doubt want, is the ability to slam in GPU-daughter boards, to add 10 or 20 7970 GPUs on a single board (preferably with sockets, which drives up the cost a few cents, but also taps into the smaller markets, where you may buy 4 GPUs now, and 6 later), so that they can drive those large super-computing projects that already make use of these GPUs, but do so more efficiently.

As for gaming, the more stream processors, I imagine, the better. When in doubt, double them, as it will give Intel and Nvidia something to curse over.

Re:Welcome to the club

[Let's All Be Chinese]

Your argument doesn't stack up.

First you say they're bringing an 8 core chip to compete with a 4 core chip. Fine. Then you complain the cores cannot keep up 1:1. So you're expecting AMD's chips to be twice as good as intel's to be able to compete.

That, of course, is rigging the test, and so is dishonest.

One could also say that with single cores not much worse than the competition, but double the number of cores, and a lower price to boot, you get better value. Moreso if you can make good use of the double number of cores.

And that's before considering that single-core benchmarks are entirely unrepresentative for multi-core performance thanks to various tricks like turbo core and turbo boost — that aren't 1:1 comparable so you'd have to do full, sustained benchmarks on all cores simultaneously to find out which delivers the most sustained instructions per second.

Meaning that AMD's offering takes more marketing footwork, but technically is not all bad. Not at all.

Re:Welcome to the club

[Anonymous Coward]

To be fair, he has a point. Per core performance is often at least if not more important than the number of cores in a cpu. Games are slowly moving towards use of multi-core architectures but they also frequently have a lot of interdependencies between different parts, making multi-core very difficult. The result is that some things like sound and networking each get their own threads but the brunt of the game simulation still happens on a single core that may or may not be a bottleneck in the entire system, depending on individual core performance.

Re:Welcome to the club

[segedunum]

I don't have mod points but I am equally as puzzled. AMD haven't had that many opportunities over the past few years (none at all really) but that was certainly one.

Sadly the systems I work on are all Intel because we do a great deal of report and post-processing on data and that requires CPU grunt and running as much as we can in parallel. Had AMD done this they would have been under consideration. Hyper-threading makes very little if any difference to us really, it's all about getting as many full cores on as possible.

Re:Welcome to the club

[unixisc]

But it's not a lot of small ones - it's companies like TI, Broadcom, Qualcomm, NVIDIA, Freescale, Samsung, Hitachi, and so on. Each much larger than AMD.

Also, there is nothing about ARM that inherently makes it more powersaving @ the same performance level than other RISC CPUs, be it SPARC, POWER, MIPS and so on.

Re:Welcome to the club

[TheRaven64]

It's a relatively small club. Note that both the headline and summary are wrong. AMD has not licensed a processor design, they have licensed the right to make their own implementation of the ARMv8 architecture (which isn't just a piece of paper, it includes access to ARM's rich set of regression tests and assistance from ARM engineers when requested on both the hardware design and the supporting software). I know of three other companies working on ARMv8 designs. For ARMv7, I think there is basically only ARM with the Cortex series and Qualcomm with the Snapdragon (which is a massively hacked-up Cortex A8, with a completely redesigned FPU, a better interconnect, and some other improvements, but not a complete independent implementation). Compare this with the ARMv4 and ARMv5 situation, where StrongARM and XScale were complete independent implementations. ARM has intentionally delayed producing their own ARMv8 design to give other companies a chance and promote more competition. This worked very well for x86 during the '90s, when Intel, AMD, Cyrix/IBM, IDT, and others were all pushing out compatible products at different market segments. In the ARM world, because they all have to go through the same set of conformance tests, compatibility should be even higher.

Re:Welcome to the club

[TheRaven64]

I am trying to grasp, somewhat desperately, the events that must have taken place inside AMD headquarters when the CPU design team said they wanted to do hyper-threading. Having seen how badly Intel got knocked around when they did it, and the fact that for the price of duplicating a fair amount of the CPU, you are still only occasionally eking out a slight performance gain...and sometimes, a performance loss, their strategy doesn't make sense

Perhaps they looked at IBM or Sun's implementation of SMT instead. Adding a second context to the POWER series added about 10% to the die area and gave around a 50% speedup. If you have multithreaded workloads (especially on a server) then it can significantly improve throughput for two very simple reasons. The first is that when one context has a cache miss, the CPU doesn't sit idle, it can let the other core work. The second is that it makes branch misprediction penalties lower, because if you're issuing instructions alternately from two contexts you can get the instruction that the branch depends on a lot closer to the end of the pipeline than before you need to make the prediction. This also helps with various other hazards, so you don't need so much logic for out-of-order execution to get the same throughput.

Re:Welcome to the club

[TheRaven64]

Also, there is nothing about ARM that inherently makes it more powersaving @ the same performance level than other RISC CPUs, be it SPARC, POWER, MIPS and so on.

I can think of several things. For Thumb-2, there is instruction density. MIPS16 does about as well as Thumb-1, but it is massive pain to work with. AArch64 doesn't (yet) have a Thumb-3 encoding, but one will almost certainly appear after ARM has done a lot of profiling of the kinds of instruction that CPUs like to generate. Even in ARM mode, the big win over the other RISC architectures is the it has fairly complex addressing modes, so you can do things like structure and array offset calculations in one instruction on ARM or 3-4 on MIPS. For AArch32, you also have predicated instructions. These make a big difference on a very low power chip, because you don't need to have any branches for small conditionals. For AArch64, most of these are gone, but there is still a predicated move, which is a very powerful version of a select instruction and lets you do mostly the same things. With AArch32 you have store and load multiple instructions, which basically let you do all of your register spills and reloads in a single instruction (the instruction takes a mask of the registers to save, the register to use as the base, and whether to post- or pre- increment or decrement it as two flags). With AArch64, they replaced this with a store-pair instruction, which can store two registers, and has the advantage of being simpler to implement (fixed number of cycles to execute).

Re:Welcome to the club

[LtGordon]

Do you think it's easier to fight an elephant or a bunch of guys who are also fighting each other

Better yet: would you rather fight one elephant-sized duck or 100 duck-sized elephants?

Re:Welcome to the club

[Kartu]

Get your facts right please, there was a bunch of x86 manufacturers, so argument "only one competitor" doesn't stand, they were there and they all died, unlike AMD:
http://en.wikipedia.org/wiki/List_of_x86_manufacturers

Now, regarding Intel, the company that has bleeding edge fabs, as simply "one competitor" is a demagogy.
x86 is also so much more complex than ARM. AMD in ARM world would be like heavy weight boxer competing with a bunch of 60kg guys.

Re:Welcome to the club

[Anonymous Coward]

I guess you take the words of Intel fanboys literally. No, the Bulldozer architecture is not hyper-threading. No, it does not mean only a slight performance gain and especially not a performance loss. I recently made 3 microbenchmarks on an Opteron 6234 (Bulldozer too). I measured the negative effect of sharing some circuits in a Bulldozer core. This negative effect varies from insignificant to small (3%, 13%, 25%). I run the same two threads on the two cores of a single bulldozer unit vs two cores on separate units. Intel hyper-threading brings 30% more performance - in the best case. The bulldozer core pair brings 75% more performance - in the worst case. How can you compare them? They are not in the same league.

The funniest benchmark was the floating point. The most frequent complaint against the Bulldozer architecture is that two cores share a single floating point unit. AMD should tell one million times that yes, they share a single floating point unit, but that is a 256 bit wide unit, which can be split into two 128 bit parts. And what is the size of the usual floating point number? Not 256 bit, not 128 bit, but only 64. In reality I measured that the two cores in a single unit processes floating point instructions almost at full speed. The negative effect of circuit sharing was only 3%, barely measurable. How ironic.

Re:Welcome to the club

[Anonymous Coward]

Note that both the headline and summary are wrong. AMD has not licensed a processor design, they have licensed the right to make their own implementation of the ARMv8 architecture (which isn't just a piece of paper, it includes access to ARM's rich set of regression tests and assistance from ARM engineers when requested on both the hardware design and the supporting software).

Despite having the right to implement their own architecture, rumor has it that AMD will be debuting with just a lightly modified licensed ARM design, due to being late to the party (and having an engineering department in chaos due to cutbacks and churn). If this is the case, they are going to get thrashed by the other companies doing custom designs.

Re:Welcome to the club

[Anonymous Coward]

For games and desktop use, yes, current applications are predominately single-thread, single-process. But for *server* chip, it's a different story. Many server applications *are* not only capable of utilizing a multi-core cpu, but pretty much embarrasingly parallell.

Re:Welcome to the club

[Anonymous Coward]

What AMD is doing isn't SMT. SMT requires minor transistor overhead to allow multiple threads to use the un-used execution resources of the pipeline that would otherwise go unused. AMD's modules are full cores for most purposes.

Re:Welcome to the club

[TheRealMindChild]

to take the advantage of having 6 full cores, and trade it in for 8 half-cores...was this some idiotic attempt at market segmentation? Did some moron in a suit have a brain fart, and think "we can't have 12-core Phenom IIIs, it will cannibalize our Opteron server sales"?

No, it has a lot more to do with technological/practical limitations and planning for the future. Two 6 core Phenom IIs, welded together, would use 200-250W. That isn't even close to ok. Nextly, their strategy it to modularize parts of the CPU, so going forward, integrating say, a GPU on die (done already), or maybe a few hundred ARM cores, it will be a no-brainer, because it's "already in the CPU design".

Re:Welcome to the club

[marcosdumay]

All that is right, throughput increases, but you'll still reduce the single thread performance.

I'm aware that AMD replicated a fuller set of the core on their hyperthreading architecture than Intel did. But even then, hyperthreading always means highter throughput but lower thread performance.

And if they really did get 50% more throughput by using 10% more area, they lost the perfect opportunity to come out with a 8 cores 16 threads model. Why did they kept it at 6 cores?

Re:Welcome to the club

[K. S. Kyosuke]

That's a rather narrow-minded comment, isn't it? The ARM core is just a small piece of the puzzle for the server solution. Those other companies you mention are doing 32-bit ARM mobile SoCs and ARM microcontrollers, whereas AMD is attempting to do a 64-bit large-memory-enabled multicore server solution with extensive communication fabric, in an area they have a lot of expertise in (Opterons, HyperTransport etc.) You really believe that the other companies are significantly ahead in this area?

Re:Welcome to the club

[Kjella]

Meaning that AMD's offering takes more marketing footwork, but technically is not all bad. Not at all.

I guess I'm looking at it, and yes it needs more work. Fewer and faster cores are better if not all the cores are fully loaded, choosing the benchmarks where all 8 cores are perfectly balanced is cherry picking. Intel makes better all-rounders at the same price. As for price and value, if the price doesn't match the value people feel it gives them the prices come down. The FX-8150 started as a $245 chip with Intel's 2500K at $216. Now the FX-8350 sells for $199 while Intel's 3570K sells for $225. Why are you now paying $50 less for a CPU that's much, much better while the competition has increased their prices? Because the FX-8150 delivered terrible, terrible value and that's engineer fail, not marketing fail.

Re:Welcome to the club

[Let's All Be Chinese]

I guess I'm looking at it

Sorry to have to disappoint you there.

I'm looking for a new desktop, and I found I had to re-assess my shortlist when I realised that the usual single core benchmarks not merely are equally cherry picked, but also are factory rigged in hard-to-compare ways. Notice that the balanced MP cherry picking is your contribution; I just said to not focus on single-core benchmarks only as there's 4 or 8 cores available. My requirements mean that MP is not unimportant, but sheer gaming prowess is.

What with the price, you ask? Perception. intel is sitting pretty so they can (and have) put prices at very neatly balanced gouching points whereas AMD has had a thorough trouncing in the press so they have to provide more perceived value this time round. The bulldozer failure may or may not have been alleviated with thread affinity mods in the OS, but even if they'd fix that now, the damage has been done. Point is: It doesn't need to be an engineering fail to require fixing through pricing drops.

Also note that AMD does drop prices over time, and intel pretty much hasn't for a while. intel can get away with keeping the prices up even with newer and presumably better parts appearing at the same marketing and pricing slots, AMD cannot. That's not all engineering, that's intel sitting pretty and AMD threading water.

Re:Welcome to the club

[fm6]

OK point taken. But even those big guys are midgets when compared to Intel. And if AMD can carve out a tiny niche in an Intel-dominated market....

I don't get what you're saying about ARM versus other RISC chips. ARM is already the last man standing in the low-power space (how many phones use MIPS?) just as x86 is the last man standing in desktop and data center markets.

Re:Welcome to the club

[Lonewolf666]

In theory, yes (I'll take your word for it). But in practice, something went wrong. In actual benchmarks, the four-module/8-core FX8150 could barely beat the fastest Phenom II X6 "Thuban" 1100T, despite being made in 32nm instead of 40nm like the Phenom II.
Using your metric of "75% more performance - in the worst case", the FX8150 should have performed like a 7-core Thuban, not counting any advantages from the improved manufacturing process. As it is, it did perform much like a 7-core Thuban. Except that in some benchmarks the Thuban actually won.

A Thuban 6-core in in 32nm would probably perform on the level of the FX8150. A Thuban 8-core in in 32nm might even beat the "Piledriver" FX8350.

So one really has to wonder what happened at AMD. Maybe the anonymous AMD ex-employee told the truth about a premature switch to automated design. (see http://www.insideris.com/amd-spreads-propaganda-ex-employee-speaks-out/)

Re:Welcome to the club

[Anonymous Coward]

I thought they would have had separate threads for rendering, user input, physics, network and disk streaming.

Re:Welcome to the club

[viperidaenz]

Welding two Phenom II X6's together means it consumes twice the power and must dissipate twice the heat.

AMD only kept up with Intel in the years gone by because Intel made a bad decision in adopting their still-born Netburst architecture. As soon as they threw that away and went back to the P3/Pentium M base of the new Core series, AMD got left in the dust.

Intel will always be a process node ahead of everyone else. Unless you have a significantly better architecture you can't compete with that.

Re:Welcome to the club

[lightknight]

200-250 Watts is perfectly fine in my world. I have an 850 Watt power supply in my main machine, I think it can supply the needed energy.

As for the thermal envelope, which you will probably want to discuss next, I must admit my lack of knowledge here. Would a liquid coolant system be capable of displacing that much energy? I'd like to think yes, and you are correct in that that would be a minor price bump for the processor (it ships with the coolant clip attached, but you still need to invest in the rest of the coolant system). But then, I am already (like many others) using a liquid coolant system on my FX-8150, so the point is somewhat moot.

What would please me more is enabling multiple sockets for a Phenom III. I'd like to be able to buy a board with two or four sockets, especially if the motherboard manufacturer does it right. That means I want those unnecessary overclocking features, I do want as many PCI-Express lanes as is even remotely possible (12 of them would not be enough), and I want as many RAM sockets as you can possibly fit on the board. Frankly, I am absolutely surprised that Asus hasn't stepped up to the plate, and started offering motherboards that would fit 256 GBs of RAM. F*ck the low-cost, low-margin, low-performance world that the tech industry has a perverse hard-on for these days. People will always pay for performance, then don't give a sh*t how much heat it dissipates as long as it freaking delivers! You only catch flack when you put out a terrible design that your competition uses to make their decade old products look good in comparison.

And more SATA sockets. Six is not enough. I have a Thermaltake Armor Full-Tower case, and I can easily fit a dozen hard drives in it. And you get us that PCI-Express 3 stuff that Intel has, and make sure all the USB ports are USB 3.0.

Who would buy this motherboard / processor? Only all the professionals that make up a huge portion of the desktop / server markets. Don't want it to cut into your server offerings? Do something similar for your server CPU offerings...and for the love of Deity, do something about those server motherboards. They are just...disasters in terms of features.

As for ARM, as I said, the market is already over-competitive, and they are having trouble sticking to the ARM instruction set while offering something different (to gain a competitive advantage). Who cares if your ARM processor has an extra two or three instructions if programmers don't want to use it because it breaks their code on devices using a processor from one of the other manufacturers? It's a fools game, just money down the drain. Oh, but ARM is going to replace x86? It's the future? Have you seen how badly these things scale? How you been paying attention to how much performance is being lost in the name of saving a little energy? By the time they are competitive with x86 offerings, they will be seconds away from drawing as much energy, or more, than x86 processors. You've reinvented the wheel!

The GPU on die stuff is cute, but honestly, who cares? The performance is fine for flash-based games, and Portal 2 if you don't run it with high settings, but load up Crysis on Ultra-high or perhaps some Skyrim, and you're not doing so hot. It's essentially a 'nice-to-have' for those long plane rides where you are just tapping out a document or watching a DVD, but that's it! Any, and I mean any, tablet manufacturer or laptop manufacturer that wants to kick the snot out of their competition just needs to work in a mediocre dedicated graphics card from two generations ago, and it will wipe the floor with the competition's integrated crap. It's essentially a video card for people who do not know what a video card is, and will purchase a new machine when they realize the current one can't be upgraded / can't do what they want it to do. I.e. it's just a waste of money. It saves the OEM money in the short term, and possibly the consumer, but 6 months later when the consumer realizes they need something a little more powerful, they need to rebuy

Re:Welcome to the club

[lightknight]

Honestly, I don't care that it would need to dissipate twice the heat. Many of us who upgraded to FX-8150s are already running liquid coolant solutions.

And compared to the i7, which is hot enough to cook an egg under a tight load, AMD has some room to play in before people begin cracking jokes about needing a black-hole to control the heat.

Re:Welcome to the club

[TheRealMindChild]

200-250 Watts is perfectly fine in my world. I have an 850 Watt power supply in my main machine, I think it can supply the needed energy.

There has never been a chip to consume that much power because it would catch fire. It isn't speculation. The highest wattage CPU made was 140W. There is a really good reason for that. You can carry away all of the heat unless the cores are laid side by side, but then you would have a processor the size of your motherboard.

The GPU on die stuff is cute, but honestly, who cares?

Imagine a world, where your video card has a direct link to your CPU (and memory). It is here! It is more than cute, it is efficient and effective. The short of it, however, is they aren't putting super-magically-delicious GPU cores in them... yet.

You also are totally missing the purpose of having an/many ARM cores in with the x86 cpu. It's ultra-low power stupid cores that you can throw a multitude of tasks at, without throwing away x86 compatibility. The Amiga is a great example of how this would work.

Re:Welcome to the club

[viperidaenz]

The Phenom II X6 is a 95W processor. When was the last time you saw a 190W processor? All the power needs to be transferred through the same surface area on the top of the chip and send in through the same number of pins. You'd need a new socket, a larger CPU size, and new, expensive motherboards to handle the power delivery.

Re:Welcome to the club

[lightknight]

I haven't seen a 190 Watt CPU, but I have seen a 150 Watt CPU.

Core 2 Extreme QX9775 Yorkfield XE (45 nm) 3.2 GHz 150 W

From Wikipedia: http://en.wikipedia.org/wiki/List_of_CPU_power_dissipation_figures

Re:Welcome to the club

[viperidaenz]

Why do you think the QX9775 uses a server grade (LGA771, it is the only non-Xeon to use it) socket instead of the consumer version of time (LGA775) like its lower power friends the QX9770 and QX9650?

Re:Welcome to the club

[Anonymous Coward]

I guess you take the words of Intel fanboys literally. No, the Bulldozer architecture is not hyper-threading.

You don't need to be a "fanboy" to see that the Bulldozer architecture is, in the abstract, similar to Intel hyperthreading. With a twist.

No, it does not mean only a slight performance gain and especially not a performance loss. I recently made 3 microbenchmarks on an Opteron 6234 (Bulldozer too). I measured the negative effect of sharing some circuits in a Bulldozer core. This negative effect varies from insignificant to small (3%, 13%, 25%). I run the same two threads on the two cores of a single bulldozer unit vs two cores on separate units. Intel hyper-threading brings 30% more performance - in the best case. The bulldozer core pair brings 75% more performance - in the worst case. How can you compare them? They are not in the same league.

You do need to be a fanboy to claim Bulldozer is supar awsum by looking only at relative gains rather than actual performance. When you do directly compare performance, you'll find that generally one Intel Sandy Bridge core outperforms one Bulldozer module. Using less power. And less die area. Great scaling going from one to two threads on a module doesn't mean as much when the baseline 1-thread performance isn't very good...

The funniest benchmark was the floating point. The most frequent complaint against the Bulldozer architecture is that two cores share a single floating point unit. AMD should tell one million times that yes, they share a single floating point unit, but that is a 256 bit wide unit, which can be split into two 128 bit parts. And what is the size of the usual floating point number? Not 256 bit, not 128 bit, but only 64.

I hate to break it to you but you're being more than a bit ignorant here. The point of "128-bit" SSE and "256-bit" AVX floating point units isn't to do 128-bit or 256-bit FP precision. SSE and AVX actually do not support 128- or 256-bit precision at all! Instead, they support dividing a 128- or 256-bit machine register into multiple IEEE single precision (32-bit) or double precision (64-bit) fields. When a CPU executes a "256-bit" AVX floating point add, it's doing either eight 32-bit or four 64-bit adds in parallel.

In reality I measured that the two cores in a single unit processes floating point instructions almost at full speed. The negative effect of circuit sharing was only 3%, barely measurable. How ironic.

What's ironic is that you think you understand anything about this stuff.

Re:Welcome to the club

[Anonymous Coward]

And compared to the i7, which is hot enough to cook an egg under a tight load, AMD has some room to play in before people begin cracking jokes about needing a black-hole to control the heat.

Bolding mine. There isn't one single chip called an i7, it's a huge range of different things. There are ULV mobile i7s which use a mere 17 watts, and at the other end of the spectrum is the i7-3960X, at 130 watts. Today's mainstream desktop i7s are 77 watts or less.

Besides, you're operating on a very out-of-date meme. It was the Pentium 4 chips Intel was selling over six years ago which were super hot compared to AMD's CPUs. Ever since Intel switched over to Core 2 and retired the P4, they've been ahead on absolute power and power efficiency.

Here's an example. Today, AMD's best FX series desktop CPUs use 140 watts, and they are generally behind the performance of Intel's 77-watt i7 CPUs. When you account for the fact that 77W i7s have an integrated GPU (which uses some of 77W power budget) and AMD's FX CPUs don't, Intel uses less than half the power to deliver equivalent or better CPU performance.

So the reality is exactly the opposite of what you say: AMD doesn't have any room to play in on this front. They're not just a little behind, they're a lot behind.

Re:Welcome to the club

[Anonymous Coward]

And that's before considering that single-core benchmarks are entirely unrepresentative for multi-core performance thanks to various tricks like turbo core and turbo boost — that aren't 1:1 comparable so you'd have to do full, sustained benchmarks on all cores simultaneously to find out which delivers the most sustained instructions per second.

Er. You do know that when people do perform "full, sustained benchmarks on all cores simultaneously", AMD's 8-core processors are basically a wash with Intel's 4-core processors? As in, they win some, they lose some, you wouldn't go far wrong with either. Except... if you ever run software which does not or cannot scale to 8 threads, AMD doesn't have a chance. It loses by giant margins.

Meaning that AMD's offering takes more marketing footwork, but technically is not all bad. Not at all.

Disagreed. Technically it's really, really bad that they need twice as many cores, about twice as much power, and much more die area to match Intel's parallel throughput, and it's a total failure that they fall way behind on anything which can't saturate all 8 cores.

You're right that they do have to do a lot of marketing footwork to sell this, though!

(Actually, you're not even right about that. Most of the "footwork" AMD does to sell these chips is to live with very low margins compared to Intel in order to attract customers with price, since the performance is lacking.)

Re:Welcome to the club

[Anonymous Coward]

You don't need to be a "fanboy" to see that the Bulldozer architecture is, in the abstract, similar to Intel hyperthreading. With a twist.

I would say that kind of "abstraction" is extreme simplification, and I would expect less of that here on Slashdot.

You do need to be a fanboy to claim Bulldozer is supar awsum by looking only at relative gains rather than actual performance.

You completely missed my point. Which was measuring exactly that, the relative performance loss caused by the shared circuits. There are many absolute performance benchmarks on the web, I do not intend to repeat them. On the other hand there are many claims and false ideas (like comparing it to hyper-threading), but not much factual data about the effectiveness of the Bulldozer core-pair architecture.

When you do directly compare performance, you'll find that generally one Intel Sandy Bridge core outperforms one Bulldozer module. Using less power. And less die area.

On the other hand it is priced proportionally higher.

I hate to break it to you but you're being more than a bit ignorant here. The point of "128-bit" SSE and "256-bit" AVX floating point units isn't to do 128-bit or 256-bit FP precision. ... What's ironic is that you think you understand anything about this stuff.

I am definitely not an expert, that is why I measure things. Believe it or not, I heard about SSE etc., but I somehow doubt, that more than 6 out of the 12 cores want to execute SSE etc. instructions at the same moment in my database and application server. Yes, I should also measure that, but actually I am perfectly satisfied with the performance of this Opteron 6234 server. If you are more interested in this subject please contribute your own benchmarking results.

Re:Welcome to the club

[cheesybagel]

All these cores are superscalar and the present Intel cores in this class have more execution units per core than AMD. The caches in the Intel processors are also much better performing. However the performance of the AMD Bulldozer processors isn't as bad as some people think, especially if you compile your own binaries and run Linux.

Re:Welcome to the club

[cheesybagel]

AMD used to have processor node parity back in the 180 nm node. In fact they even had copper interconnects before Intel. The issue is they lost their advantage and now they don't even own their own fabs. The funny thing is AMD's ex-CEO and one of the founders Jerry Sanders used to say "real men have fabs". He attributed AMD's ability to survive against Intel due to their low per unit costs.

Re:Welcome to the club

[viperidaenz]

That was the end for node parity. Intel moved to 0.13 the next year and has been a node ahead ever since. The cost per unit was the reason after my P3, I bought AMD CPU's for several years. Never did like the P4.

Cheaper, Lower power, More Cores, Faster-ish

[Anonymous Coward]

They are cheaper, so expect cheaper servers. They use less power, and drastically less power when idling, this means less heat, and less cooling, and better packing densities. You can put more cores in a chip, more chips on a blade, and more blades in a rack and the cooling is still viable. So in practice that means your bang per buck is far better, because you can afford more cores and can cool more cores.

Faster- ish, faster because more cores, ish because some apps need a lot more processing on one core and don't scale well as cores scale.

For small servers, well as long as you don't use any of the MS stuff, you won't even notice the change. A lot of file and print servers run with Arm now, so there won't be much change there. If you're an MS shop with all your stuff tied to Microsoft. Well, good luck with that, I don't use MS servers so I don't know what your options are.

Re:Cheaper, Lower power, More Cores, Faster-ish

[marcosdumay]

I don't use MS servers so I don't know what your options are

The options are simple to understand. People that use MS servers can either 1 - Pay or 2 - Cry. Also, they get to choose often.

ARM64 + Hypertransport = Interesting Outlook

[Uzull]

In fact AMD has an amazing technology portfolio. Having graphics chip (ATI Division), the hypertransport technology and AMD64, we can expect some interesting developments

Re:ARM64 + Hypertransport = Interesting Outlook

[Spy Handler]

I remember when AMD bought ATI many years ago... everybody (including us Slashdot posters) were saying what a bone-headed waste of money that was.

Now everybody's saying AMD is really fucked except for one bright spot which is its graphics division....

Re:ARM64 + Hypertransport = Interesting Outlook

[drinkypoo]

These are not contradictory statements.

AMD could be really fucked today because they put too much effort into graphics, and not enough effort into CPUs. Only time will tell if their graphics will save them. I suspect they won't unless they learn how to write drivers that work.

Re:ARM64 + Hypertransport = Interesting Outlook

[Kjella]

Well I still think it was a bone-headed waste of money, they bought a company that did okay and is still doing okay but they haven't really gotten that much synergy out of it. The world still runs on exclusive CPU or GPU solutions, HSA is still a bit pie-in-the-sky. Instead they could have put that money into CPU design and/or processing tech and let ATI/nVidia/PowerVR compete for delivering chipsets/IGPs or the graphics part of their APUs. With Intel putting the thumbscrews on nVidia, they could have been rather favorable to helping AMD but not when AMD owns ATI. Biting over more than they can chew is a leading cause to AMD being in the situation they're in now.

Re:ARM64 + Hypertransport = Interesting Outlook

[cheesybagel]

Even if they wanted to buy ATI they could have done it for much less. The deal was ruinous for AMD. In fact that deal put a lot of people from Intel Capital in court and ex-AMD CEO Hector Ruiz only didn't get sued himself because he agreed to testify as a witness in the case.

Originally designed for mobile phones???

[Anonymous Coward]

ARM architectures are considered more energy-efficient for some workloads because they were originally designed for mobile phones and consume less power.

Fuck no. The ARM1 was released in 1987 as a coprocessor for Acorn's BBC Micro. They were designed for low power operation because the engineers were impressed with the 6502's efficiency. There weren't any significant mobile phone deployments until 18 years later in 2005.

Re:Originally designed for mobile phones???

[Anonymous Coward]

Almost. The first ARM1 was produced in 1985. This was used in the BBC micro coprocessor to design the ARM2. The first ARM2 silicon was produced in 1986 and the Archimedes computers, which ran on the ARM2, were released in 1987. I've still got my A310.

But yeah, it had nothing to do with mobile phones.

Re:Originally designed for mobile phones???

[serviscope_minor]

They were designed for low power operation because the engineers were impressed with the 6502's efficiency.

Nope. They were designed for low power so that they could use cheap plastic packaging instead of expensive ceramic packaging.

Re:Originally designed for mobile phones???

[Anonymous Coward]

You got the part about 1987 and Acorn right. The first ARM processor went into the Acorn Archimedes - which came much later than the BBC Micro. (Though the previous poster was correct that the BBC Micro coprocessor interface (the "Tube") was used as a prototyping platform for initial bring-up).

ARM as a separate company from Acorn had its first big design win in the Apple Newton (as Apple was instrumental in spinning ARM out of the carcass of Acorn/Olivetti). Apple wanted a low-power RISC processor, and at the time, essentially most the other RISC players were targeting high-end workstations (e.g. Sun SPARC, SGI MIPS, etc.) or high-end embedded (AMD's 29k and intel's i960). So low-power, handheld applications are in ARM foundational DNA.

But the poster above claimed ARM had little mobile presence before 2005? Crazy. Nokia used ARM cores in their mobile phones at least as early as their widely popular 6110 handset circa 1997. Texas Instruments ARM-based OMAP SoCs were widely used in Nokia's & many other vendors' handsets throughout the early-to-mid 2000's. These primarily use the ARM7, ARM9, and ARM11 cores.

The latest ARM Cortex A- series cores have been available since around 2005, and these are widely used in smartphones and tablets today (e.g. Cortex A8 in iPhone 3GS, Cortex A9 in iPhone 4, iPad 2, etc. Cortex A15 just starting to show up in the Google/Samsung Chromebook.)

Re:Originally designed for mobile phones???

[am 2k]

I read an article a while ago that stated that the ARM processors were so efficient by accident. They started from scratch with the design, not having the experience of Motorola, IBM, Intel and AMD of what a fully-fledged processor requires, and so it became a very simple one. This happens to be an important element for power efficiency.

Re:Originally designed for mobile phones???

[MrNemesis]

Much like everyone else says, they were designed more for simplicity than anything else, and extremely low power consumption was an unintended side effect. Of course they were going for low power so they could use the cheap housings as mentioned above, but the frugal amounts it did actually eat were unintentional.

There was an article on The Register some months ago on ARM development history (can't seem to find it now), and if it's to be believed they were investigating a series of mysterious crashes in the prototype ARM CPU, and in debugging they found the power on their dev kit wasn't actually connected to the chip - it was running entirely on leakage current and if there weren't enough 1's going into the chip to provide current, it wouldn't have enough power to run.

Re:Originally designed for mobile phones???

[Anonymous Coward]

They were so efficient due to two factors...

1) they wanted to use a plastic package for low cost, so could not dissipate a lot of heat.

2) they accidentally achieved a lower power budget than they were trying for to accommodate the plastic packaging.

Re:Originally designed for mobile phones???

[Anonymous Coward]

You're correct. Originally the low power was an accident and completely unintentional. It had bugger all to do with "ceramics" at least at the beginning.

Re:Originally designed for mobile phones???

[matmota]

There was an article on The Register some months ago on ARM development history (can't seem to find it now), and if it's to be believed they were investigating a series of mysterious crashes in the prototype ARM CPU, and in debugging they found the power on their dev kit wasn't actually connected to the chip - it was running entirely on leakage current and if there weren't enough 1's going into the chip to provide current, it wouldn't have enough power to run.

I think you mean this article, "ARM creators Sophie Wilson and Steve Furber Part Two: the accidental chip":
http://www.reghardware.com/2012/05/03/unsung_heroes_of_tech_arm_creators_sophie_wilson_and_steve_furber/

Re:Originally designed for mobile phones???

[mikael]

I was told ARM didn't have access to the big design tools and libraries that the other companies had. They had to design every logic cell by hand, so by coincidence fewer transistors mean lower power consumption.

I guess it's like C++ where you could build an object from the inheritance of other classes and have a whole load of functions you don't use, or just write a custom class with just the functions you need.

Android Java Server

[Anonymous Coward]

Pretty please.

Very funny - desktop isn't the high end

[dbIII]

While you may have wandered in from the territory where the beige box is called the "hard drive" and the screen on the desk is called "the computer", there's people that work with the hardware, and some of that hardware is Xeons, 16 core AMD cpus, sparcs etc etc. Even though I'm only at the cheap end of that stuff I don't mistake the desktop for the "high end".

Premature Optimization in translating x86-ARM

[girlinatrainingbra]

They need a cute small name for the ARM cores to go with the bulldozer for the large cores: maybe call them Cats or Toros or Deeres, more lawn-mowery than bull-dozery. :) Anyway, why translate X86 to arm when you could recompile directly for ARM? Isn't that where GNU and Linux (or GNU and HURD) could show the advantage of free-software-open-source-software's source code access allowing for the direct or cross-compilation of the source code into binaries which run on the ARM?

I would think that taking pre-compiled X86-binaries and then translating {pre-compiled and optimized for x86} code into ARM code is a waste of time and a lot of premature optimization.

Re:Premature Optimization in translating x86-ARM

[neokushan]

How would that work in a situation where you have both x86 and ARM cores on the same system?

From what I've read, I definitely get the impression that AMD is doing some kind of modular system, whereby their APU cores can be coupled with either ARM or x86 variants. I'm not sure if that also includes ARM+x86 coupling, but that seems to be the point of "fabric" - it's a universal interconnect of some kind. However, how would that work in real life? What would be the advantage of it? ARM for seriously low power, x86 for when the CPU has a lot of work to do and the APU for anything that requires GPU-style parallel computing? A trifecta of computing where there's a core for every possible need?

NOTironic that Apple could very well be going back

[girlinatrainingbra]

Well, considering that they made the jump for the PowerPC architecture to the x86 architecture because IBM/Motorola could not provide a low power version of the G5 PowerPC chip to be used in the portable space of laptops, it doesn't seem ironic at all that Apple might consider using a low power consumption chip in the laptop or portable space at all. It almost makes darned-good-sense.
.

And considering what they'd been doing with Pink / Taligent in keeping a parallel universe of development of their codebase always going on the x86 architecture while publicly showing only PowerPC development, they've probably got a skunks-work factory team somewhere that's already been running ARM-based IOS or even ARM-based OSX for a year if not for years...

so now we know...

[slew]

So now we know what Jim Keller is back at AMD to do...

on a 1 billion trans die, = 500 cores?

[cheekyboy]

If they can fit 200-500 cores per die, and then have 8 dies per server, that will kick ass.

Since most server tasks are long in terms of seconds. And dont require 2800 bogomips per thread.

Re:NOTironic that Apple could very well be going b

[Anonymous Coward]

they've probably got a skunks-work factory team somewhere that's already been running ARM-based IOS or even ARM-based OSX for a year if not for years...

iOS is ARM-based. Always has been since first release.

TSMC might stand a chance

[Anonymous Coward]

TSMC has already demonstrated with ATIs 7xxx series GPU that they can do 28nm chip sizes. They should be able to apply that to AMD CPUs and remain competitive with Intel on that front.

Re:TSMC might stand a chance

[Anonymous Coward]

Intel's already at 22nm, and there's more to a process than just its geometry. (as in, Intel 32nm was significantly better than anyone else's 32nm.) Don't expect TSMC 28nm to be a savior.

MIPS64

[FithisUX]

For me it would make more sense if they followed the MIPS64 path. But.. its their money.

Re:MIPS64

[drinkypoo]

The market chose ARM over MIPS because MIPS stalled. You can still buy a SuperH core, but why would you do that? Same for MIPS. Unless you really don't need performance and you're getting a really great deal, it's a bit difficult to fathom. All the interest is in ARM, so that's where the talent is developing. And since newer stuff tends to be built on a lower process it's not just faster, but also lower-power which is what everyone and their mom (literally) is demanding today.

MIPS didn't keep up with ARM. Unless they can actually leapfrog it and be _faster_ (and a benchmark or two doesn't prove much) there's no reason to go that direction. Unless, of course, they can somehow drastically reduce power consumption, but it's difficult to imagine them pulling such a trick out of their hat now.

Re:NOTironic that Apple could very well be going b

[Anonymous Coward]

But also Darwin based, just like Mac OS X

tough times

[ssam]

AMD are having tough times in several sectors of the CPU market (server and low end desktop seems ok). some company resort to lawsuits to cling on, very glad to see diversification instead. GPUs only get so far for HPC, a large number of simple cores on a die will be applicable to a wider set of tasks.

ARM will succeed for servers

[Required Snark]

ARM is going to be a significant part of future server space. This issue comes up every so often on Slashdot and I always see the same reaction: x86 is the one true architecture and nothing will displace it. That's not a technically based opinion, it's a religious dogma.

When you're on the client side of the network, it makes no difference what's on the server side. It could be a giant room full of hamsters and abacus. As long as the results come back fast and correct, you shouldn't care. That's the way the internet was designed. Heck, that's why it'd called the Inter-Net. Inter networking between different processor platforms.

Intel is a one trick pony. Besides the evolution of the x86, they have never fielding an architecture that had any staying power. Anyone remember the i432 or the i860? The current standard x86-64 architecture was defined by AMD, not Intel. Itanium got that moniker because it was accurate. The only reason that the Itanium is alive is because of a civil suit by HP.

What Intel is really really good at is putting gates on silicon. They did not succeed on architectural grounds, but by having the best implementation of a clunky architecture. They were always able to succeed by using more gates at a lower price then the competition.

ARM is an architectural rival to x86. Intel won with the x86 because they could cram more gates onto silicon. They loose this advantage against ARM because ARM requires less silicon to do the same job. This translates to lower power usage, which is getting more and more important as time goes on. Other foundries can compete even if they are trailing Intel in processes capabilities, and they want to be in this market. As does AMD.

ARM also benefits from being the dominant architecture for the smart phone/tablet sector, which means that there is a large community of developers and all the software one could ever want. An ARM-centric ecology exists, and it applies to servers as well as client software. Linux/GCC/MySql are happy on ARM, so any open source server software is easily available. And Microsoft has shown they are ready to run on ARM as well. It's not a risk from a software point of view.

It's not that Intel/AMD x86 is going away, but ARM will also be a player. And we should all be glad about it, because AMD being less competitive with Intel is the road to monopoly, which means increased prices and a stagnant CPU sector.

Re:ARM will succeed for servers

[ledow]

I once saw a 1U rack that contained something like 16 ARM boards (the entire board, networked together with a switch, powered from individual cables, with disk interfaces over some custom central channel). It cost less, used less power, and did more in the same amount of space. It was a bit homebrew-esque (despite being a professional product), but the advantages were rife.

I was sorely tempted to use it just because, as you say, server-side doesn't matter for most things. And with that sort of basic setup, you could deploy one app per board and not worry about having enough power for all the tasks combined (firewall on one, http server on another, mail on another, auth on another etc.). And keep a board or two in reserve for spares / failures and it would last longer too. Rather than having one monolithic machine and hoping it never goes wrong, have a bunch of them in a box (yes, the box can still go wrong, of course) each modularised to do one task and then adjust / failover accordingly without anyone even knowing there's more than one computer serving their request or even what architecture they are).

And software support? Well, you just install the same version of Linux on them all, then install one of those multi-SSH utils that executes the same command on all the servers (hell, you could have one board do nothing but monitor and repeat all commands onto the others).

If I were to put a 1U rack in colocation, that's how I'd do it. Who cares what's underneath so long as you have power enough to do everything and OS-support enough to run it. And with ARM, OS support isn't a problem.

£500 to the first person to supply a 1U filled to the brim with Raspberry Pi's (or equivalent) with appropriate "single point" Ethernet access and mains power input on the exterior and some kind of central storage for them all.

Re:ARM will succeed for servers

[am 2k]

£500 to the first person to supply a 1U filled to the brim with Raspberry Pi's (or equivalent)

I think the Parallella board would be perfect for this, much better than the Raspberry Pi.

Re:ARM will succeed for servers

[ledow]

Not really.

1) Kickstarter. Sign of a project doomed to failure when it concerns hardware, really. Especially where they are talking on the scale of producing hardware boards with en-masse dozens of cores on them from a few hundred thousand dollars.

2) No OS support - it seems to be a number-cruncher with an ARM-controller, not a generic computer with lots of software already ported. Nobody will rewrite their software to take advantage of it unless it's MADLY to their advantage (i.e. number crunchers, not generic machines). Sure, it has a C compiler but so does a Z80. That means nothing if they can't be turned to generic-purpose cores that anything can use without having to change the way they work / thread.

3) $99 (not guaranteed, they don't even have the tooling for it yet!) for the prototype which barely has two ARM cores as the power house and 16 of their custom (unsupported) cores for doing some floating-point. Not worth it compared to, say, a $25 Raspberry Pi en masse.

I'm not a RPi fan, to be honest - I have one, and I've commented enough on here about its shortcomings. But they *are* delivering hardware (not perfect, but it's usable if you know about the problems with it), it's quite simple, bog-standard, cheap, works, there are half-a-dozen Linux distros with optimised code available for it already and it "just runs" Debian ARM programs without needing to recompile to take advantage of its hardware.

This, in comparison, isn't/doesn't.

There's a point at which the sacrifices to squeeze another core or reduce power by another watt, or decrease size by another mm will cost you ordinary users even if they don't cost you hardcore, specific-purpose hardware geeks out to save everything they can.

Hell, the RPi is being bought by/for kids so they can play on it - it's really that well rounded, supported and available.

Re:ARM will succeed for servers

[am 2k]

Not really.

1) Kickstarter. Sign of a project doomed to failure when it concerns hardware, really. Especially where they are talking on the scale of producing hardware boards with en-masse dozens of cores on them from a few hundred thousand dollars.

That's just prejudice on your part. There are many amateurs doing Kickstarter projects they don't fully understand, but there are also some professionally done ones on there. The hard task of backing something there is to find out which of these two types the creator is.

2) No OS support - it seems to be a number-cruncher with an ARM-controller, not a generic computer with lots of software already ported. Nobody will rewrite their software to take advantage of it unless it's MADLY to their advantage (i.e. number crunchers, not generic machines).

It supports OpenCL, which is the standard for this kind of thing across many device types. Of course, if you're talking about web servers and databases, you might have a problem.

Actually, I can't think of anything in a web server that could benefit from OpenCL... In databases, there might be a lot of things. The core issue with those processors is that it's only really efficient when you operate on a rather small dataset that can be kept local memory.

Of course, as you said, there's nothing off-the-shelf here, you need to adapt the code to that kind of architecture. The cheap trick to spread the workload on multiple cores by simply handling every request on its own thread doesn't work with it. The point though is that you still aren't stuck on a single device when adapting it, since OpenCL is supported for many architectures, including the Intel and AMD processors. They also benefit from data locality due to their caches, although it is not that huge of a requirement.

Re:ARM will succeed for servers

[marcosdumay]

They loose this advantage against ARM because ARM requires less silicon to do the same job.

They lost this advantage because Moore's law is comming to the end. The cost of shrinking the feature size is exploding (even in semicondutor industry's terms) and the difference between generations are getting smaller (there were 3 generations where Intel couldn't half the feature size already).

Re:ARM will succeed for servers

[Anonymous Coward]

A few companies are doing that:

Code think

Calxeda

Dell

Re:ARM will succeed for servers

[viperidaenz]

Why not just stick a bunch of $7 Allwinner SoC's on a board and call it a day? Cheaper more powerful and denser than a bunch of Raspberry Pi's

Re:ARM will succeed for servers

[Anonymous Coward]

You're conflating less silicon for the same job with smaller lithography. ARM's benefit isn't because of smaller lithography, which is the only reason new x86 generations are still able to eke out improvements. ARM's benefit is that it doesn't that smaller lithography, because it doesn't have to drastically increase transistor count to improve.

Re:ARM will succeed for servers

[evilviper]

x86 aways wins because anythig useful is put into an x86 core, wholesale. From FPUs to SIMD to virtualization to GPUs. If ARM had any advantages, Intel or AMD would be copying them, and putting them into x86 chips for all to use.

That's how x86 CPUs went from CISC to RISC, and added all the above features, and more.

ARM is a clear example of the MHz myth in action. Their DMIPS/MHz is VASTLY lower than any x86 chip, and their overall MHz/clock speed is much lower, too. ARM has nothing to offer for servers. AMD went the same way with Geode years ago, and went nowhere, too. And even if ARM had some real, technological advantage.... If x64 has taught us anything, it's that backwards compatibilty is worth trillions of dollars to companies, and piss-poor emulation is not a valid substitute.

Re:ARM will succeed for servers

[cheesybagel]

The x86 overhead is less significant than a lot of people think and backwards compatibility is the hallmark of all successful computer architectures.

sorry, I typo-d my statement

[girlinatrainingbra]

oof! My mistype. I meant to say "x86-based IOS or ARM-based OSX" to complement the pre-existing ARM-based OIS and x86-based OSX.

(poetically getting the theme-scheme a+d and c+b to complement a+b and c+d

Use cases: AMD leverage ARM+Radeon

[DrYak]

Also, AMD can spit out some interesting use cases where it can find a nice empty niche market:
by leveraging their built-in GPUs.

Not simply putting a low powewred mobile GPU (AMD radeon's Qualcom Adreno cousin, PowerVR, etc) something more high-powered (some of their own low power radeon designs):

- Coupled with a multi-core ARMv8 CPU, Can be useful for netbooks with good graphic performances (the same kind of market after which Nvidia is running with their own Tegra series).

- Some numerical loads can benefit from a low power CPU core and a decent GPU.
When building server farms, the performance-per-watt matters, and that's why ARM is starting to eat on the x86 territory.
With a low-power ARM and decent GPU, AMD's creations can also end-up as low-power GPGPU nodes with a crazy low energy requirement. (I've seen ARM + custom parallel chips combo appearing. There might be market for something like this).

So there is definitely some place for a multicore ARM 64bit by AMD.

Best part for us geeks? AMD is likely to keep their open-source policies.
So you can expect documentation pushes to coreboot&co (for the ARM part) and to opensource radeon&co (for the GPU part).
And this is a racidally different situation than the present ARM situation, where most of the graphics core are closed.
Either completely closed (PowerVR)
or undergoing some reverse engineering (Mali/Lima)
or getting some docs (Adreno's common ancestry with Radeon helps)
or being an opensource farce (Broadcom's Videocore is basically a software 3D engine running on a dedicated core. Their "fully opensource driver" is a thin layer which uploads the firmware to the dedicated core and then forwards the OpenGL ES calls as-is).

Currently, no nice opensource graphics in the embed world, unlike Intel (and AMD).
But if AMD keep their opensource trend, and Nvidia keep their promises (after the Linus' "Fuck you scandal" they ended up promising to open-up a little bit more), we're going to see an interesting battle of "ARM with big GPUs" in the embed linux world.

re: how would that work

[girlinatrainingbra]

One possible way to make that work: when compiling to generate the Application Bundle for a specific program, simultaneously generate two sets of code to include in the bundle:
1 - x86 code
2 - ARM code

Then, depending on power-resource utilization requests by the OS or directly by the user, the executing instance of the application can be migrated from one of the types of cores (eg x86 core) to one of the other types of cores (eg ARM-cores) by copying over (a) the current instances variable values, (b) the current instance's stack as a stack pointer = SP, and (c) a pointer to where the current instances program is within the context of the program code as the next instruction pointer (execution pointer.

Implementing (c) above may require creating an array of equivalent program break-points and entry-points for the x86-compiled code and the ARM-compiled code, meaning that the code to generate the bundle would compile the code for both architectures and then generate the bijection map between the two code-bases.

Probably: they want to do both x86 and ARM?

[Erich]

What other company could make a processor that does both x86 and ARM? Windows 8 that runs both ARM and legacy x86 apps? I could see that as being pretty differentiated. Their GPUs are on par with nVidia, and they have better processor microarchitecture.

Re: how would that work

[Anonymous Coward]

You cannot realistically do this unless you define a common ABI and virtual ISA and compile to that. There isn't a 1:1 mapping of instructions, registers, stack, and heap values for optimized code between these two ISAs or really any two ISAs.

In the end, you'd probably really need a virtual ABI and ISA and then JIT compiling interpreters that could share analysis state and transfer the running interpreter between hosts while minimizing the re-start time to get the hot sections coded to efficient host machine code again.

That Spot Isn't That Bright Either

[EXTomar]

I've been waiting for years for the AMD + ATI merger to payoff but there are parallels with what is going on with graphics as with processors. This was supposed to be a synergy that would lead to optimized and amazing PC technology for games but has that happened? Especially after the last machine I built for myself both AMD and ATI have been slipping in quality.

I have my complaints with how Nvidia but their cards have gotten reliable and each new generation has had not only the standard performance increases but new features as well. Features like Adaptive VSync are a "why didn't they have this before?" kind of feature that I can't find if ATI supports yet. Meanwhile ATI has been...around. I have noticed that hardware and driver quality and stability differ from card to card and game to game.

I can't figure out what happened. I used to think AMD got distracted with ATI causing them to lose focus in processor but I wonder if something deeper happened in management that caused company to wander.

A lil correction

[Anonymous Coward]

You do not license ARM-Architecture based designs from "ARM" you license them by Acorn. Deeply sorry having too correct you, probably the statement was from AMD. However nice little Barcode you have there Slashdot. Seems like a joke on the current QR hype.

Regards, the redfox

Memory ordering may be a problem

[harryjohnston]

I'm imagining they're mostly intending to sell these to folk currently running open source Linux software on x86. I wonder how much of the existing software has subtle memory ordering bugs that are hidden by the lenient characteristics of the x86 ISA? If ARM servers become thought of as unreliable - even if the cause is properly identified as software bugs - folk may be unwilling to use them.

On the other hand, the big players (such as Google) should be able to fix their own code base in test environments before pushing out to production, and if they see advantages in power consumption and density they could probably keep AMD in business all by themselves.

For those that aren't familiar with memory ordering issues, Jeff Preshing has quite a good blog. See for example This is why they call it a weekly ordered CPU.

Re: how would that work

[Anonymous Coward]

give this man a job, sounds like what apple did after moving to x86; sounds good.

transparent swithcing between low powered arm and high powered x86, amd gpus are very awesome; having them present on arm and x86 would provide a solid gaming platform

Windows 9

[Anonymous Coward]

RT is no more