Slashdot Mirror
AMD Announces First ARM Processor
MojoKid writes "AMD's Andrew Feldman announced today that the company is preparing to sample its new eight-core ARM SoC (codename: Seattle). Feldman gave a keynote presentation at the fifth annual Open Compute Summit. The Open Compute Project (OCP) is Facebook's effort to decentralize and unpack the datacenter, breaking the replication of resources and low volume, high-margin parts that have traditionally been Intel's bread-and-butter. AMD is claiming that the eight ARM cores offer 2-4x the compute performance of the Opteron X1250 — which isn't terribly surprising, considering that the X1250 is a four-core chip based on the Jaguar CPU, with a relatively low clock speed of 1.1 — 1.9GHz. We still don't know the target clock speeds for the Seattle cores, but the embedded roadmaps AMD has released show the ARM embedded part actually targeting a higher level of CPU performance (and a higher TDP) than the Jaguar core itself."
OK, RISC is good and all, but they're claiming performance well below the competition at a higher cost per flop in watts. And it's going to need everything recompiled.
Seriously, WTF happened? The Opteron 2300s were very good and extremely competitive, but AMD decided to burn that to the ground for really no reason. It's not their engineering that kills them, it's their execution. I know of no other company that can piss away 5-6 years of R&D and then claim over and over that the inferior new stuff is so great. It's mind-blowing.
Jaguar is for tablets and seems to be designed for price point and not speed. That's why they are comparing it with the ARM stuff and not using an Opteron 6386 as a comparison.
A57 r0p0? So... no way they're going to production with that. Expecting quite a long gap between early samples and production parts (like shipping systems 2015). Will be interesting to see how this compares to Denverton.
"breaking the replication of resources and low volume, high-margin parts that have traditionally been Intel's bread-and-butter.": Intel has served us sort-of well, only. That's one thing. I also have friends who worked there and found it unhappy. I also am personally unhappy that Intel broke Nat Semi's forward looking CPUs and that they worked on the standard salesmen's plan. ARM is a good idea, and is winning. Kudos. Good luck also to AMD.
Back then the mobile space hadn't exploded yet.
I believe you've read it wrong. Basically, AMD actually traveled back in time to develop the first ARM processor.
$
What exactly does ARM offer here that x86 doesn't?
An FX-8150 has a specInt_rate of 115. I've never seen an 8350 but it should be around 130-ish, just like an Opteron 6212.
thegodmovie.com - watch it
Intel == x86. When x64 came alone, Intel tried to hang on to Itanium, which failed.
Intel will not go the ARM route, because Intel does not control ARM. Intel likes to have a monopoly (or in reality a slight duopoly with AMD).
If Intel went for ARM, Intel would have to wait for ISA improvements like SSE/NEON from ARM and would be limited to improving manufacturing, microarchitecture and SOC integration. And the (for mobile) important latter, Intel is Not Good.
Is this another nail in the coffin for the x86 architecture? Is it realistic to expect Windows/Mac OS X for ARM in their desktop versions in the near future? (Linux is already there). Of course x86 won't suddenly disappear, but may become "legacy". Intel should start moving on the ARM front
Actually x86 IS efficient for for something completely different. The architecture itself is totally unimportant as deep inside it is yet another micro code translator and doesn't differ significantly from PPC or Sparc nowadays.
x86 short instructions allow for highly efficient memory usage and for a much, much, much higher Ops per Cycle. This is just that big of a deal that ARM has created a short command version of ARM opcodes just to close in. But then this instruction set is totally incompatible and also totally ignored.
Short instructions do not matter on slow architectures like todays ARM world. These just want to safe power and therefore it fits in well that ARM also is a heavy user of slow in-order-execution.
A nice example, increasing a 64 bit register in x86 takes ONE byte and recent x86 CPUs can run this operation on different register up to 100 times PER CYCLE, all commands to be loaded in THREE to EIGHT Cycles from memory to cache. On the other hand, the same operation on ARM takes 12 bytes for a single increment operation, to load some dozend of these operations would take THOUSANDS of clock cycles.
And now you know why high end x86 is 20-50 times faster than ARM.
"Life is short and in most cases it ends with death." Sir Sinclair
I would have thought AMD would have a licensing clause as part of the sale of the Imageon (Adreno) to Qualcomm in case they ever decided to re-enter the market.
I believe you've read it wrong. Basically, AMD actually traveled back in time to develop the first ARM processor.
No - its making a food processor for cannibals. The design brief was that you should be able to process a whole arm.
The microserver market is still less than half a percent of the server market and most of that is x86, not ARM. That's probably why Calxeda went bust.
We've been using Java for desktop and embedded work for the last 10 years.
The codebase is over a quarter of a million lines. We've never had an upgrade
break our code.
We had one bright guy who decided to put an ARM in instead of the Power PCs
we were using. Well, it turns out a very complex driver for a very complex
comms chip used bitfields all over the place to describe the hardware
registers. And bitfield interpretation depends on endianness. ALL the
register and control structures that used bitfields had to be edited. It
was a nightmare.
Your CYCLE is an instruction cycle.
On an ARM that would be a CLOCK cycle as most instructions execute in one clock cycle.
Intel wastes 100 clock cycles.
Plus all the overhead of huge caches, huge instruction buffers, translation tables, reordering processing.
ARM doesn't need all that.
The ARM designers haven't even begun to speed up the implementation yet, and already have threatened the low end Intel based servers.
They're calling it the Opteron A. Seriously, AMD? That won't be confusing, when Opteron can now mean ARM or x86_64. AMD's processor naming scheme is already confusing, and they just decided to make it more confusing. Idiots.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Until ARM gets PCI, ACPI, UEFI, and other standardized hardware platform design equivalents, I don't want to see any more ARM hardware being built than I have to. The most severe problem with ARM devices is that the source code released for them bit rots and they become too difficult to maintain once interest in the platform wanes. In many cases the platform just never really quite picks up enough steam. Often these SoC devices have closed-source binary firmware blobs that are mandatory for even basic functionality. The Raspberry Pi even has this problem.
Until ARM hardware provides software with a common bootstrap method and a common hardware information presentation method a la ACPI or PCI, every ARM platform requires a unique kernel for that platform. This is the biggest obstacle to keeping ARM devices relevant beyond a few years after release, and while the current state is apparently okay for mobile phones since they're replaced constantly, any general purpose ARM computer must have a critical mass of users or a major corporation actively supporting it to stay up-to-date. Chromebooks have Google, RasPi has a massive user base, and routers have the Open/DD-WRT communities, but that's about it. Even mobile phones bit rot within 2-3 years due to the exodus to newer phones, and the WonderMedia platforms are largely stuck with older kernels that are difficult to get working.
tl;dr: ARM really needs the kind of standardization x86 PCs have to remain relevant long-term.
It's a quip-- of course I know that ARM stands for Amalgamated Regional Militia.
http://www.geoffreylandis.com
AMD SkyNet!!!!
Former AMD/ATI employee here, sorry the first AMD arm processor was about 5 years ago. It was an imageon A250 processor, done by the now defunct handheld division in toronto, assets sold to qualcomm.
Details here http://en.wikipedia.org/wiki/Adreno
Why would not Intel, with its fabs and process nodes, produce best ARM around?
Well it does already as Altera's Stratix 10 would use Intel's 14nm fabs.
http://newsroom.altera.com/pre...
But why not Intel itself?
4wdloop
Or define "winning" as having a higher revenue, higher margins, more R&D, and more fab plants, when comparing any single other company to Intel.
We are talking about x86 vs. ARM so this should be about total revenue rather than single companies.
But if you want to talk about companies, Qualcomm has the same market cap as Intel and has similar margins. And unlike Intel they are not sacrificing 1.5% of their margin this year to buy their way into the mobile market.
The headline would have been more clear if they had included one more word. "AMD announces ITS first ARM processor" would have said it unambiguously.
It's an awkward sized board so it's a matter of replacing the chassis as well, plus new disk controllers would be needed as well. A bit of power for another year or two is cheaper than that. So your advice, although obvious and thought of years ago, is useless.