Re: [gentoo-user] {OT} Will ARM take over the world?
On Sat, Dec 08, 2012 at 10:51:15PM -0500, Michael Mol wrote > It's looking promising. Not that I have a horse in the race, but > I very much like ARM's low power consumption. An update to my earlier response. A slasdot article at http://slashdot.org/topic/datacenter/intel-launches-centerton-to-take-on-arm/ > As expected, Intel launched its Atom S1200 System-on-a-Chip (SoC) Dec. > 11. Intel hopes that the 6-watt architecture-developed under the > codename "Centerton"-will allow it to push back against the emerging > ARM microserver market. > > The Atom S1200 is a dual-core 64-bit chip. It's been optimized > for servers and the data center, with features including ECC, > virtualization, and hyperthreading (the lattermost is capable of > delivering four threads per chip). The eight lanes of PCI Express > 2.0, along with a memory controller supporting up to 8 GBytes of > DDR3 memory, help make the S1200 a true System-on-a-Chip. And the usual discussions/flamewrs are at... http://hardware.slashdot.org/story/12/12/12/1712200/intel-announces-atom-s1200-soc-for-high-density-servers -- Walter Dnes I don't run "desktop environments"; I run useful applications
Re: [gentoo-user] {OT} Will ARM take over the world?
On Sat, Dec 08, 2012 at 10:51:15PM -0500, Michael Mol wrote > It's looking promising. Not that I have a horse in the race, but I > very much like ARM's low power consumption. There's an article on Slashdot about Intel's relatively new 22 nm SOC (System On Chip) design... http://hardware.slashdot.org/story/12/12/10/2346257/tsmc-and-global-foundries-plan-risky-process-jump-as-intel-unveils-22nm-soc It's not so much that Atom consumes more power than an ARM SOC, it's that a multi-core Atom, plus a discrete GPU, plus various other discrete components, consumes more power than an ARM SOC. Move the discrete components onto the chip (hence System On Chip), and power consumption goes down. Intel can still wring out a lot of efficiencies. They simply haven't had to in the past. That's totally separate from obvious stuff like fabbing only one core on their low-power-chips. A 2 or 4-core Atom on a smartphone/tablet is the computing equivalant of one person commuting to work in rush-hour-crawl traffic in a minivan or SUV powered by a V8 engine. A single-core chip would be equivalant to a 4-cylinder engine. It'll do the same easy job, but consume less energy in the process. -- Walter Dnes I don't run "desktop environments"; I run useful applications
Re: [gentoo-user] {OT} Will ARM take over the world?
Am Samstag, 8. Dezember 2012, 19:25:55 schrieb Grant: > It seems like ARM processors will destroy x86 before too long. Does anyone > think this won't happen? no two reasons: not enough power does not run x86 software the second one is a real deal breaker. -- #163933
Re: [gentoo-user] {OT} Will ARM take over the world?
On Sun, 9 Dec 2012 13:44:09 -0800 Grant wrote: > > But this is an old, old, old argument. People predicted the demise > > of> mainframes for years when x86 started becoming a quite powerful > > of> cpu.> The current truth is that IBM sell more mainframes year > > of> on year, growth> is more than mere inflation can account for, > > of> and mainframes are just> getting stronger. So x86 didn't kill > > of> the mainframe, instead x86 played> a huge role in making both > > of> stronger. I see no reason to believe the> same story won't play > > of> out exactly the same all over again between x86> and ARM.This > > of> is really interesting. "all over again" is exactly what I > > of> expect to happen, but I didn't realize it happened as you say. > > of> A friend of mine was really into SPARC in the 90's and > > of> complained loudly when x86 grabbed its market share. At least > > of> that was how I understood it. I imagine the same thing > > of> happening with ARM and x86, but maybe I'm jumping to > > of> conclusions? x86 and SPARC is not the same thing as x86 and ARM. SPARC was a RISC processor but in it's heyday was comparable to x86 in terms of computing power. It had one sponsor (Sun) and one user (Sun) and one OS (Solaris, or maybe it was called SunOS back then). x86 had far greater mindshare in general plus it had the killer "feature" - the bean counter was already using it in his desktop and knew SPARC and x86 were quite comparable in some significant ways. He also knew the price difference It's a classic case of a smaller player trying to take on a bigger player directly on it's own turf. x86 vs ARM is not that game at all. ARM is an embedded processor that, whilst it could replace x86 on low-end desktops, really shines in embedded. It won't displace x86 (nor is it trying to), it will carve out new niches for itself, almost exactly like x86 did when mainframes and minis ruled. Where ARM does replace x86, I reckon it will be because x86 was not really a good solution there. For example, Atom vs ARM (that is a valid comparison). I don't think Atom will last much longer - the form factor that really used it - netbooks - is much better served by tablets. The tablet trumps the netbook, and Atom dies when the netbook dies. -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] {OT} Will ARM take over the world?
On Sun, 9 Dec 2012 15:23:28 -0800 Grant wrote: > > > It seems like ARM processors will destroy x86 before too long. > > > Does anyone think this won't happen? > > > > > > - Grant > > > > Yes me. > > > > ARM will take over where it makes sense and won't where it doesn't. > > > > For desktops and laptops, x86 was used simply because there was > > nothing else, so x86 is going to have to fight for it's survival in > > the whole bottom and middle range there. > > And won't the manufacturers be more inclined to use a processor they > can manufacture themselves under license than buy one from Intel or > AMD? And won't end users be more inclined to buy them since they'll > be cheaper? You have neglected to mention the single most important factor of all: inertia Human groups are loathe to change things that already work good enough for something that works better. Intel works good enough. -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] {OT} Will ARM take over the world?
> > It seems like ARM processors will destroy x86 before too long. Does > > anyone think this won't happen? > > > > - Grant > > Yes me. > > ARM will take over where it makes sense and won't where it doesn't. > > For desktops and laptops, x86 was used simply because there was nothing > else, so x86 is going to have to fight for it's survival in the > whole bottom and middle range there. And won't the manufacturers be more inclined to use a processor they can manufacture themselves under license than buy one from Intel or AMD? And won't end users be more inclined to buy them since they'll be cheaper? - Grant
Re: [gentoo-user] {OT} Will ARM take over the world?
> > It seems like ARM processors will destroy x86 before too long. Does > > anyone think this won't happen? > > > > - Grant > > Yes me. > > ARM will take over where it makes sense and won't where it doesn't. > > For desktops and laptops, x86 was used simply because there was nothing > else, so x86 is going to have to fight for it's survival in the > whole bottom and middle range there. > > x86 has already lost the fight in the portable market (phones, tablets) > and truth be told never really got going there. > > For everything else, I don't see ARM making much inroads. There's a hug > server market out there for things with computing grunt and the > software that drives them - x86 isn't going away in that market anytime > soon. > > But this is an old, old, old argument. People predicted the demise of > mainframes for years when x86 started becoming a quite powerful cpu. > The current truth is that IBM sell more mainframes year on year, growth > is more than mere inflation can account for, and mainframes are just > getting stronger. So x86 didn't kill the mainframe, instead x86 played > a huge role in making both stronger. I see no reason to believe the > same story won't play out exactly the same all over again between x86 > and ARM. This is really interesting. "all over again" is exactly what I expect to happen, but I didn't realize it happened as you say. A friend of mine was really into SPARC in the 90's and complained loudly when x86 grabbed its market share. At least that was how I understood it. I imagine the same thing happening with ARM and x86, but maybe I'm jumping to conclusions? - Grant
Re: [gentoo-user] {OT} Will ARM take over the world?
Am Sun, 09 Dec 2012 20:24:58 +0100 schrieb Florian Philipp : > Am 09.12.2012 04:51, schrieb Michael Mol: > > On Sat, Dec 8, 2012 at 10:25 PM, Grant wrote: > >> It seems like ARM processors will destroy x86 before too long. Does anyone > >> think this won't happen? > > > > It's looking promising. Not that I have a horse in the race, but I > > very much like ARM's low power consumption. The way I see it, they're > > only a short list of features away from obliterating x86: > > > > * I'd like to see fast division. > > I keep hearing about how this or that is slow because of ARM's lack of > > strong division. > > > > * I'd like to see a modern baseline of strong instructions. > > x86 kept continually improving in a very fragmented way, but there > > were, from time to time, baseline collections of feature sets you > > could expect all processors to have. i386 represented one. i686 > > represented one. Currently, it's x86_64, which implies not only a > > 64-bit flattened address space and a departure from real mode, but > > also a collection of SIMD instruction sets and other features > > developed between the release of the Pentium Pro and AMD's Hammer > > architecture. > > > > ARM just feels...fragmented. And I don't have the impression I could > > write my code assuming the availability of SIMD (presuming I use > > things like OpenMP to expand my code to leverage it, rather than > > writing processor-specific code. Though OpenCL could very well > > alleviate that issue.) > > > > +1 with regard to fragmentation. What I especially despise is the lack > of a common boot infrastructure. If I'm not mistaken, it is still > impossible to make a kernel that boots on all (or at least a large > subset of all) ARM platforms [1]. > [1] https://lwn.net/Articles/496400/ I'm no embedded developer, but I read recently that the first batch of multi-platform ARM support has been merged in Linux 3.7: http://www.h-online.com/open/features/Kernel-Log-Coming-in-3-7-Part-5-CPU-and-platform-code-1758293.html -- Marc Joliet -- "People who think they know everything really annoy those of us who know we don't" - Bjarne Stroustrup signature.asc Description: PGP signature
Re: [gentoo-user] {OT} Will ARM take over the world?
Am 09.12.2012 04:51, schrieb Michael Mol: > On Sat, Dec 8, 2012 at 10:25 PM, Grant wrote: >> It seems like ARM processors will destroy x86 before too long. Does anyone >> think this won't happen? > > It's looking promising. Not that I have a horse in the race, but I > very much like ARM's low power consumption. The way I see it, they're > only a short list of features away from obliterating x86: > > * I'd like to see fast division. > I keep hearing about how this or that is slow because of ARM's lack of > strong division. > > * I'd like to see a modern baseline of strong instructions. > x86 kept continually improving in a very fragmented way, but there > were, from time to time, baseline collections of feature sets you > could expect all processors to have. i386 represented one. i686 > represented one. Currently, it's x86_64, which implies not only a > 64-bit flattened address space and a departure from real mode, but > also a collection of SIMD instruction sets and other features > developed between the release of the Pentium Pro and AMD's Hammer > architecture. > > ARM just feels...fragmented. And I don't have the impression I could > write my code assuming the availability of SIMD (presuming I use > things like OpenMP to expand my code to leverage it, rather than > writing processor-specific code. Though OpenCL could very well > alleviate that issue.) > +1 with regard to fragmentation. What I especially despise is the lack of a common boot infrastructure. If I'm not mistaken, it is still impossible to make a kernel that boots on all (or at least a large subset of all) ARM platforms [1]. And then, there is the simple fact that current ARMs lack the raw power of an x86 and I guess if you scale them up to the point where they can compete with x86s with regard to computing power per core, there is no point in switching to ARM to begin with. Sure, you can parallelize and make a large array of "wimpy" nodes, but you cannot fool Amdahl's law. And even where you can parallelize nearly 100%, you risk high latency [2, 3]. [1] https://lwn.net/Articles/496400/ [2] http://pages.cs.wisc.edu/~jignesh/publ/nonwimpy.pdf [3] http://research.google.com/pubs/archive/36448.pdf Regards, Florian Philipp signature.asc Description: OpenPGP digital signature
Re: [gentoo-user] {OT} Will ARM take over the world?
On Sat, 8 Dec 2012 19:25:55 -0800 Grant wrote: > It seems like ARM processors will destroy x86 before too long. Does > anyone think this won't happen? > > - Grant Yes me. ARM will take over where it makes sense and won't where it doesn't. For desktops and laptops, x86 was used simply because there was nothing else, so x86 is going to have to fight for it's survival in the whole bottom and middle range there. x86 has already lost the fight in the portable market (phones, tablets) and truth be told never really got going there. For everything else, I don't see ARM making much inroads. There's a hug server market out there for things with computing grunt and the software that drives them - x86 isn't going away in that market anytime soon. But this is an old, old, old argument. People predicted the demise of mainframes for years when x86 started becoming a quite powerful cpu. The current truth is that IBM sell more mainframes year on year, growth is more than mere inflation can account for, and mainframes are just getting stronger. So x86 didn't kill the mainframe, instead x86 played a huge role in making both stronger. I see no reason to believe the same story won't play out exactly the same all over again between x86 and ARM. -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] {OT} Will ARM take over the world?
2012/12/9 Grant : > It seems like ARM processors will destroy x86 before too long. Does anyone > think this won't happen? No, it won't. > > - Grant
Re: [gentoo-user] {OT} Will ARM take over the world?
On Sat, Dec 8, 2012 at 10:25 PM, Grant wrote: > It seems like ARM processors will destroy x86 before too long. Does anyone > think this won't happen? It's looking promising. Not that I have a horse in the race, but I very much like ARM's low power consumption. The way I see it, they're only a short list of features away from obliterating x86: * I'd like to see fast division. I keep hearing about how this or that is slow because of ARM's lack of strong division. * I'd like to see a modern baseline of strong instructions. x86 kept continually improving in a very fragmented way, but there were, from time to time, baseline collections of feature sets you could expect all processors to have. i386 represented one. i686 represented one. Currently, it's x86_64, which implies not only a 64-bit flattened address space and a departure from real mode, but also a collection of SIMD instruction sets and other features developed between the release of the Pentium Pro and AMD's Hammer architecture. ARM just feels...fragmented. And I don't have the impression I could write my code assuming the availability of SIMD (presuming I use things like OpenMP to expand my code to leverage it, rather than writing processor-specific code. Though OpenCL could very well alleviate that issue.) * I'd like to see virtualization be a thing. Productivity and efficiency on x86 *soared* with the compartmentalization that came with hardware-assisted (and therefore cheap! and fast!) virtualization. I haven't heard about the same on ARM, although Citrix is working hard on porting Xen there. Paravirt may well be the first common means of virtualization on ARM... -- :wq
[gentoo-user] {OT} Will ARM take over the world?
It seems like ARM processors will destroy x86 before too long. Does anyone think this won't happen? - Grant