Slashdot Mirror
AMD Gives ARM License a Miss, Will Stick To x86
CWmike writes "Advanced Micro Devices has shot down rumors that it is pursuing an ARM license, saying it will stick to developing chips for tablets around the x86 architecture. 'We've made a big bet on APUs, which are x86,' said John Taylor, a marketing director at AMD, referring to accelerated processing units. AMD has been criticized for a lethargic approach to entering the fast-growing tablet market, which is dominated by ARM."
I'm not buying a tablet until it can run MS-DOS and Lotus 1-2-3. Period.
The real path to male liberation
The behemoths that once lay claim to be innovators are starting to drop. I buy AMD chips over Intel for my desktop, but my next tablet will be ARM. As the next generation are pretty much standardizing on Android for the OS (though Microsoft are working hard to make their OS ARM compatible), speed and battery life are going to be two key differentiators. ARM has the clear advantage here. Of course tablet sales are only a tiny drop in the sea of revenue for a company like AMD, but it does seem short-sighted none the less.
Phillip.
Property for sale in Nice, France
Indeed, i dont give a fuck about arm. i have endless supply of software, games and other shit i may need to run on a tablet. until tablets can provide backwards compatibility to provide for that, gtfo. im not paying $400 for something that only presents a 10 inch screen to watch videos on youtube and make facebook updates or use half assed simple widgets. my phone can do these.
Read radical news here
What would an AMD tablet run? Surely they don't think they're going to sell a lot of tablets running Windows 7? What truly mature, tablet-ready OS runs on Intel? (And I'm not talking about Parsimonious Palembang or some other future Ubuntu release -- what's available in, say, June?)
Oliver's law of assumed responsibility: If you're seen fixing it, you will be blamed for breaking it.
For a moment I thought you wrote "trane" -- k5's crack-smoking economic illiterate that makes Paul Krugman sounds reasonable.
This should come as no surprise to anyone who's been paying attention to AMD. AMD has already bet on x86-64 scaling down to tablet form factors; that is, after all, the entire point of the Bobcat architecture. Later this year when Bobcat transitions to 28 nm we'll see if it pans out; even if it doesn't there's always the 20nm transition in late 2012, and that's sure to lower power requirements enough to make an x86 tablet viable.
At the same time, it's obvious that there really isn't any room in the ARM SoC market for new entrants*. NVIDIA is already selling Tegra 2 SoCs for a cut-rate $25 a chip, and those are going into already too expensive Android tablets. The message is clear: the only way to make a profit with ARM chips is in volume, and there's no way a new entrant like AMD is going to ramp to significant volume to even cover production and R&D costs before their own Bobcat architecture has made the transition to 28-20nm and they're basically competing with themselves.
*- Yes, I know AMD wouldn't be entirely a new entrant, as they had an ARM license as recently as a few years ago, which they subsequently sold off, but by this point they'd essentially be new entrants all over again
Oliver's law of assumed responsibility: If you're seen fixing it, you will be blamed for breaking it.
lolARM.
Yea in the 80's it was the shit. Now it's the Shit again, not.
AMD ftw, always, and not in a fanboy sense but in a 6core for less sense.
Intel is as greedy as ATT & Apple, hence their success. Nothing to do with architectures
or ARM, just the advertising. No one gives a shit about ARM.
Here are the facts.
RISC(ARM) - software based
CISC - hardware based
CISC will always be the one to beat, even more so today and in the future. Why? Because we are reaching the limits of clockspeed and transistor size as we know it. How do we make it faster? Hardware. Hardware is what drives everything computing including the "Cloud". No it's not software that made this all possible. The real reason we are able to do what we is because of pioneers like 3COM. Prior to them nothing was really connected. Do you see where I'm going with this? It's about the connections, the hardware, the ability of the hardware to be SMART in how it communicates with other parts of the network, whether it's around the world or 12 nms away on your motherboard. The RISC approach to networks would be to build faster and bigger hubs. You remember hubs, not very smart in where data is transmitted, just blast everything with packets and hope someone gets it. RISC instructions are kind of the same thing. Blast the cpu with instructions until something gets done.
You will never get the speed you need by just trying to fill the pipe with more instructions(software). There is always a limit to that. Why? Because of the hardware limitations. How does RISC get faster, more capable? Improve the hardware. RISC is and forever will be playing catchup to CISC. Oh sure you will see times when RISC has finally caught up with CISC, like it's going to surpass it. What happens? It never does and can't because of the hardware limitations.
That's where AMD's APU comes into play. They are actually improving those hardware limitations. Going beyond the size and clockspeed limitations we are approaching. RISC will never be able to do anything but add more registers and more Hz, but as explained above those days are numbered. Oh but they can you say, they could add this or that, well if they do that it's really not a RISC cpu anymore is it?
Comment removed based on user account deletion
There's no surprise here. AMD (The CPU division anyway) always follows one step behind Intel and only takes an opportunity to jump ahead when Intel does something boneheaded (like in the x64 development.)
AMD has nothing to lose by not jumping into this because AMD chips aren't being used in mobile phones or tablets, and they know the x86 platform is a godamned joke to stick on a tablet. AMD is sticking with where it's doing good: Graphics chips, Servers, Desktops and some Laptops. The tabletPC era using x86 and windows isn't going to happen. It already failed once, and nobody learned from that lesson.
The first para of the article is way off the mark. AMD did not "shoot down" rumours of pursuing ARM. It has bet heavily on APUs, but that does not preclude it from branching out. The relevant passage in the article is this:
"ARM CEO Warren East added fuel to the fire by commenting on the opportunity his company sees for AMD to use ARM processors in the future. Clearly, AMD has signaled that they are going through a bit of a rethink of their strategy at the moment. And therefore ... that presents as far as we are concerned ... a heightened opportunity," East said on a conference call to discuss financial results earlier this week."
So how does the article say "Advanced Micro Devices shot down rumors that it is pursuing an ARM license, saying it will stick to developing chips for tablets around the x86 architecture?" Obviously the editor is an idiot, who probably took 30 seconds to mis-read the article and then spit out some important sounding bullsh*t for the headline. The mindset of the press is so Neanderthal.
Some apps call native library code due to the performance of compiled interpreted code being poorer than native.
These will be compiled for the target device.
You see -- All of the software I need to run is open source, thus it runs on any architecture.
Good for you. Seriously. Good for you. I'm not being the slightest bit sarcastic either. I wish that were the case for more people.
Unfortunately that also tells us some things about what YOUR needs are not. Clearly you aren't a heavy duty CAD user, you aren't an accountant, odds are you aren't a graphics professional (Photoshop), you don't use MRP or ERP either and I could go on. I also very much doubt that all of the hardware you use is open source only. (While possible to do in theory, open source only hardware is very restricting.) Open source is great but there are some types of applications for which the closed source versions remain clearly superior and probably will for the foreseeable future. There is no open source 3D solid modeling software comparable to CATIA or ProE. Hell there isn't a 2D open source CAD package that even matches AutoCAD. There is no open source accounting software comparable even to QuickBooks much less some of the enterprise level accounting software. I use GIMP all the time but there is no open source replacement for Photoshop if you are a graphics professional. There are tons of additional examples. Open source simply isn't the best choice available right now for some software needs. I hope that changes but I'm not holding my breath.
If your needs are fully satisfied by open source software, that is really terrific, but it doesn't describe a huge percentage of the rest of us. I literally could not do my job using only open source software. I use as much as I can but it simply does not exist for some of my needs. (predominantly manufacturing and accounting)