ARM Stealthily Rising As a Low-End Contender 285
snydeq writes "InfoWorld's Neil McAllister examines how the ongoing rise of netbooks, decline of desktops, and the smartphone explosion are reconfiguring the processor market, putting Intel's Atom processor on a clear collision course with ARM. And here, on the low end of computing, Intel may have finally met its match. Thanks to a unique licensing model, ARM will ship an estimated 90 chips per second this year, and the catalog of OSes and apps available for ARM has been growing for decades, including several complete Linux distributions such as Google's Android OS and Chrome OS when it ships. 'One thing ARM doesn't have, however, is Windows,' McAllister writes, something that could ultimately stymie ARM's plans to compete on the low end of the netbook market. And yet Intel's bet on Windows and its x86 compatibility appeal among developers could backfire, McAllister writes. In the end, it's all about performance. Thus far, Intel has yet to demonstrate a model with power characteristics comparable to those of the current generation of ARM chips, which are fast proving their ability to handle high-performance applications."
ARM/Linux in the Tesla Roadster (Score:5, Informative)
To tie in with an earlier article on the front page: the Tesla Roadster's battery pack management system is ARM-based. It's built around a Philips-LPC2294 with 32 megs of ram and a 1GB U3 Cruzer Micro USB flash drive, running Linux kernel 2.6.11.8-1.3.0.
Re:So, where are ARM netbooks? (Score:5, Informative)
http://www.alwaysinnovating.com/touchbook/
Windows CE and Windows Mobile (Score:3, Informative)
Windows CE and Windows Mobile both support ARM.
There might not be "full-featured Windows" on ARM, but saying there's no Windows at all on ARM is just ignorance.
Low power FTW (Score:4, Informative)
I just purchased a Wikireader, which uses a low power Epson S1C33E07 60 mhz RISC processor, not unlike an ARM. It will run for 90 hours on 2 aaa batteries. And that includes a 240 * 208 capacitive touch screen.
Re:So, where are ARM netbooks? (Score:5, Informative)
I'm frankly a bit surprised. You can get beagleboards and shivaplugs, with pretty credible ARM based specs, for not all that much even in small quantities, and ARM based smartphones are all over the place, so the field seems surprisingly thin on the netbook side.
Re:MAME on ARM in Debian (Score:4, Informative)
http://packages.debian.org/lenny/arm/xmame-sdl/download [debian.org]
I've run Debian ARM distro on an NSLU2. Works great.
The Debian Distro Says Otherwise (Score:5, Informative)
http://www.debian.org/ports/arm/ [debian.org]
I've run it on an NSLU2. Worked perfectly. They've got desktop packages for it an everything.
Ubuntu is has been standing on the shoulders of giants (Debian) for a long time. It's time for you to go straight to the source.
Re:Almost 3 billion chips this year? (Score:5, Informative)
Had you read the article, you'd see where all the chips come from, because it's summarized right below that line on the first page.
Hint: if you have an electronic device that is NOT a desktop or laptop computer, the odds are somewhere around 99 out of 100 that it's using one or more ARM chips. This includes, but is not limited to, cell phones, GPSes, home routers, calculators, and portable gaming devices like the DS.
Re:OS/X? (Score:5, Informative)
Apple's iGadgets are ARM-based and run a variant of OS/X. Of course, ARM also has WinCE, so that kind of balances the karma.
Re:Windows on low-end? (Score:2, Informative)
I actually upgraded from XP Home to Windows 7 Ultimate on my HP Mini Netbook (using an Atom N270). I was very pleasantly surprised at how well it runs. It actually feels smoother than XP and is generally quite usable, even with the slow 60GB hard drive. Then again, I also have 2GB of RAM installed, but the memory usage only went up maybe 100MB. Note that I am a Linux junkie and all of my other computers run Linux.
Re:MAME on ARM in Debian (Score:4, Informative)
$99 + $17 shipping, no tax. There's only 1 supplier in the US at the moment.
And for slashdotters, the devkit is MUCH better than PogoPlug or other 'final' products.
USB -> JTAG adapter. If you fubar it, you should be able to unfubar it.
SD Slot: 8GB card will act as the boot drive. Saving wear on the internal 512MB memory and allowing me to add a ton of other stuff.
I plan on it being my IRC, AIM, Torrent, Usenet, XBMC Serving, HVAC Controlling, 1-Wire Weather Sensing, 5W (max) box.
For kicks I'll probably do some mencoder benchmarks.
FYI: http://computingplugs.com/ [computingplugs.com] is hosted on a Plug. It survived the last Slashdotting. The guy was using it to stream a TV show and it was still only using 40% CPU. He only unplugged it when he didn't know he was getting slashdotted and thought it was acting weird.
Re:OS/X? (Score:3, Informative)
Apple's iGadgets are ARM-based and run a variant of OS/X [sic]. Of course, ARM also has WinCE, so that kind of balances the karma.
It's not a variant. It is the same code that runs on an iMac or MacBook, just trimmed back (e.g., no FireWire or SATA drivers).
The Windows CE has barely any relation to Windows Vista or 7. It's two different code bases, whereas with OS X it's the same code base.
This doesn't really make much of a difference to the end user in most cases, but keeping one code base bug free and thoroughly tested is generally less of a hassle than keep more than one code base the same.
Re:ARM/Linux in the Tesla Roadster (Score:2, Informative)
Supposed to be ported to Alpha as 'Was sold for Alpha' ?
Re:ARM/Linux in the Tesla Roadster (Score:5, Informative)
Between versions 3.5 and 5.0 (Windows 2000), NT was actually ported to several other architectures, including PowerPC and Alpha. None of them were particularly successful commercially, and from what I heard, the Alpha port was killed by in-fighting between Microsoft, DEC, and (IIRC) Compaq. The only non-x86-derived (and on RISC architecture) port of NT currently maintained is Itanium (ia64).
There's no architectural reason why NT couldn't be ported to ARM, and I actually think it would be a good move to replace the WinCE kernel with a ported branch of the NT kernel optimized for smartphones. They could even keep backward compatibility with WinMo by using a WinMo subsystem (similar to the way that NT is compatible with both Win32 and POSIX by way of subsystem - the kernel doesn't directly handle Win32 or POSIX syscalls, they instead both get translated to NT syscalls which are designed to accommodate just about any API). This would also let Microsoft remain relevant on ARM-based netbooks, provided they port the Win32 subsystem (yes, applicaitons would need to be re-compiled, but for many apps that's all it would take).
Re:ARM == Hype (Score:4, Informative)
ARM has ARM mode, Thumb Mode, Jazelle Mode, and ThumbEE mode. FOUR instruction sets. Multiple different floating point unit specs that are incompatible with each other. Crazy page table formats. The architecture spec is over 2000 pages long, for pete's sake!
ARM has a more uniform encoding, but actually has a large number of instructions, and does crazy things like put a rotating shifter in the load address path. Not good from a modern pipeline perspective. You can get around it by breaking up the operation, but then you're getting into complex instruction decode like x86.
I'm not saying ARM is bad. I'm just saying they have no magic. You're right, Intel doesn't either (though they do have manufacturing and an army of engineers to do hand-layout). Nor does MIPS or PPC. But MIPS does make energy efficient cores, roughly as good as ARM. They haven't been as popular as ARM, but they're around.
And I'm certainly not saying x86 is great -- it's certainly not. I don't think it's quite as bad as people make it out to be...
Look, I wish the architecture made a difference. For one, we'd all probably be using Alpha. That was a great, elegant, beautiful processor architecture. For another, I'd have much better job prospects. But it doesn't matter that much. Scalar architectures are scalar architectures. Instruction set makes some difference, but not very much.
Re:Fast is not always best (Score:4, Informative)
Waiting for disk access.
Re:ARM == Hype (Score:3, Informative)
MIPS has been a lot more successful in this space than ARM ever has. Cheap PCs all over China are using MIPS CPUs which rival 1GHz x86 CPUs. <$150 Netbooks have been available for a couple years now, using MIPS chips, a market ARM has been making a lot of noise about, but has only just now entered, and not even near the price point...
See above. ARM has been making a hell of a lot of PR noise, but that's the only thing they've done with any success.
Re:ARM == Hype (Score:3, Informative)
Thumb is just a subset of ARM - so not another instruction set :-) and you haven't really said how a mini-i386 will get that code density. Jazelle etc just let you implement very fast Java apps on a very low power CPU - so what's wrong with that?
There is no such thing as magic - it's all design at some point. ARM had the freedom to invent a better instruction set and they did. It allowed them to make very cheap CPUs at a time when that was needed. Now that processes have improved we can consider x86 but only when wearing very pink spectacles.
So there were problems with floating point but it doesn't seem to have stopped anyone.
I think that you're a bit fixated on CPUs that fit into machines with power cords or heavy batteries and what makes them good - which isn't what makes mobile CPUs good. I think that ARM deserve credit for seeing how important it would become and having exactly what was needed at the right time. They have salespeople (cough) now like any successful company but I have never thought that they benefit from more hype than anyone else.
Re:ARM == Hype (Score:3, Informative)
Having shifts in the address calculation is fine for ARM7 where you're trying to squeeze every possible functionality out of a tiny number of gates, and don't really care about performance. But for even a reasonably high-performance design, you need to have a consistent pipeline.
Probably the most important pipeline is the Decode->RegRead->AddressFormation->Dcache->Writeback pipeline. The latency of this pipeline is critical for performance. ARM has some advantages here: uniform (or, somewhat less so, semi-uniform, a la Thumb2) is easier to decode than variable-length at the byte level x86. Most architectures have an adder in the AddressFormation part (though notably not ia64). If you add two registers (which you can't in MIPS) you probably want to be able to shift by the access size because you're doing something like indexing into an array. So a small left shifter before the adder isn't uncommon, and it's usually about a 4:1 mux in terms of delay.
But ARM allows you to do full rotations in front of the adder. This means you need more levels of logic in front of the address calculation adder, which hurts your memory latency. You can make it a multicycle instruction or split it up into multiple instructions (and many implementations do), but that of course adds significant complexity.
The page table formats are kind of kooky. Most 32 bit architectures choose 4K pages as the minimal page size. 4K L1 translation and 4K L2 translation translates all 20 bits you need. The page tables are a multiple of the page size, which is handy. It's so clean, it's pretty obviously the "right" thing to do.
ARM has a 16KB l1 translation, because they used to support 1KB pages, but no longer do. They have strange attributes that move around the format, which makes it more difficult to manipulate the page table entries. They also have no free bits, which makes it a pain in Linux to keep information like how new or clean the page is.
I will say that the page tables are getting cleaner as they deprecate things like 1KB pages, but they're still pretty painful compared with other architectures.
The Alpha Architecture Handbook [cmu.edu] is a good read, and Alpha is my very favorite RISC. Not that it's magical, either, but it's a lot cleaner than ARM. And it's less than half the length of the ARM Architecture Reference Manual (ARM ARM, which I must admit is a clever acronym).