ARM plays hand, reveals 1GHz plans | Drobe.co.uk archives

ARM plays hand, reveals 1GHz plans

By Chris Williams. Published: 5th Oct 2005, 10:11:32 | Permalink | Printable

More mobile, more powerful, more Intel rattling

Semiconductor giant ARM has responded to Intel's plans to introduce low power processors, by revealing its next generation of chips. Dubbed the Cortex A8, the chip is designed to run from 600MHz to 1GHz, which is a massive leap beyond ARM's current low power range of embedded processor cores. The device is set to draw around 300mW of power, and is aimed at future mobile gadgets and consumer electronics that require real time processing of video and other media.

Fabricated using 65nm and 90nm processes, the Cortext-A8 is said by ARM to deliver around 2 DMIPS per MHz, compared to current cores that deliver around 1 DMIPS per MHz.

The new processor core also packs hardware to accelerate integer and floating-point calculations, and the likes of Samsung and Texas Instruments have lined up to license the designs for their own chips. It's scheduled to begin shipping during the first quarter of 2006. Buried in the corporate hype, ARM admits that its new chips will likely include so-called digital rights management technology, which is touted as preventing piracy despite concerns from free software groups.

Although ARM have been working on the Cortex family for years now, the timing of the announcement of the A8 could rattle Intel's plans to use its ARM based XScale processor range as a stop gap, while it works on reducing the power consumption of its x86 families.

RISC OS 4 and 5 can be found running on StrongARM, ARM7, ARM9 and XScale processors.

Links

Cortex-A8 ARM news release

Previous: Fonts website folds under legal threat
Next: Rare Acorn graphics joystick in auction

Discussion

Viewing threaded comments | View comments unthreaded, listed by date | Skip to the end

fpu

nex on 5/10/05 10:50AM
[ Reply | Permalink | Report ]

Fantastic news!

Ofcourse, now the question arises - Will Castle or Advantage6 include these new ARM cores in their forthcoming hardware designs? Are their mouths watering like mine?

This development will propel RISC OS into the future with greater strength and support.

hEgelia on 5/10/05 10:59AM
[ Reply | Permalink | Report ]

superscalar

Cogs on 5/10/05 11:26AM
[ Reply | Permalink | Report ]

Whats a DMIP?

CJE on 5/10/05 11:39AM
[ Reply | Permalink | Report ]

hEgelia: come off it, it ain't gonna happen.

And what's with all the modding down?

Also, doesn't RO4 work on the ARM610 (RPC600) too?

simo on 5/10/05 11:50AM
[ Reply | Permalink | Report ]

CJE: DMIPS Dhrystone MIPS (Million Instructions Per Second)

egel on 5/10/05 11:53AM
[ Reply | Permalink | Report ]

According to the news story on another website, ARM is quoted as saying that it takes 3 years from finishing a processor design to the processor being in widespread use.

I wonder how many of us will still be using RISC OS then, and how many companies will still be actively developing.

demondb on 5/10/05 12:16PM
[ Reply | Permalink | Report ]

There's always one who has to spoil the excitement

sa110 on 5/10/05 12:21PM
[ Reply | Permalink | Report ]

This signifies ARMs first foray in to proper grown up processors, incorparating a superscalar pipleline (the ability to execute up to 2 instructions simulatnously) and level 2 cache (to aleivate problems with the relatively slow memory used in ARM based systems), which will give significant performance benefits. I would expect a 1GHz A8 Cortex to have up to around 3.5x the integer performance of a 600MHz XScale IOP321 used in the Iyonix. The addition of a hardware FP unit would result in to an order of magnetude increase of the speed of floating point operations, of great benefit to OpenGL based 3D graphics.

druck on 5/10/05 12:35PM
[ Reply | Permalink | Report ]

It is still up to the ARM licensees to make real silicon based on this core. Like demondb is saying, ARM is saying this will take 3 years from now on so a lot can happen in between.

joty on 5/10/05 1:22PM
[ Reply | Permalink | Report ]

So we can expect the Iyonix 2 in about 2009 perhaps.

dgs

dgs on 5/10/05 1:37PM
[ Reply | Permalink | Report ]

These processors' stated MHz speed reflect their target function as ultra-low power chips. They could probably be rated for much higher speeds if the tight electrical and heat restraints present in their target market weren't present. What I'm saying is they'd probably at least theoretically 'overclock' very nicely in a RISC OS desktop situation.

Clades on 5/10/05 2:24PM
[ Reply | Permalink | Report ]

In reply to dgs:

Or the A9 2

Col1 on 5/10/05 2:31PM
[ Reply | Permalink | Report ]

Col1: Wouldn't they be more likely to name it the A8?

VinceH on 5/10/05 4:07PM
[ Reply | Permalink | Report ]

Does anybody know why it takes three years from design to silicon?

maikl on 5/10/05 5:09PM
[ Reply | Permalink | Report ]

If it is going to take 3 years from now for this core to be available what does "It's scheduled to begin shipping during the first quarter of 2006" mean above?.

vshears on 5/10/05 5:19PM
[ Reply | Permalink | Report ]

Read properly what demondb said: "According to the news story on another website, ARM is quoted as saying that it takes 3 years from finishing a processor design to the processor being in widespread use."

"Widespread use" doesn't mean three years to become available!

Gulli on 5/10/05 5:33PM
[ Reply | Permalink | Report ]

In reply to Gulli: You are totally correct in this case.

In reply to dgs: Do you really think so, i bet Castle have then already Iyonix 3 in widespread use, may even with RISC OS 5.70 or 6.00

In reply to Col1: Maybe A6 have then not the A9 2 in use, instead of that maybe a A9-Turbo in use, One A9-Home computer with DVD/CD RW and 2 x USB HD and build-in LCD Display.

datawave on 5/10/05 10:07PM
[ Reply | Permalink | Report ]

In reply to Gulli.

I for one had read properly what demondb said.

But I was actually reffering to a post from joty who said "It is still up to the ARM licensees to make real silicon based on this core. Like demondb is saying, ARM is saying this will take 3 years from now"

vshears on 5/10/05 10:25PM
[ Reply | Permalink | Report ]

<shame>IconBar picked this up first... </shame> ;-)

NickyDeB on 5/10/05 10:39PM
[ Reply | Permalink | Report ]

Up to 2MB level 2 cache! :-)

Current x64 (64-bit x86) CPUs only have 1MB.

JGZimmerle on 5/10/05 11:32PM
[ Reply | Permalink | Report ]

vshears: Sorry, forgot to name who I was actually answering which was the exact same comment as you were - yours made good sense to me

Gulli on 6/10/05 12:56AM
[ Reply | Permalink | Report ]

In reply to vshears:

With regards to availability in 1Q2006, I think Chris did not have the story right. The ARM press release states that the core is available for licensing now with most of the tools ready except for the "Artisan Advantage-CE library" (a silicon cells library optimised for high-speed operation and low static and dynamic power consumption) deliverable only from 1Q2006. There is no mention of any production date from any of the licensees even if I would expect them to be fast on the mark given that it gives them a good occasion to grab a part of the XScale market.

andretim on 6/10/05 7:58AM
[ Reply | Permalink | Report ]

This sounds pretty different from existing ARMs, so hopefully it's not too stupid a question to ask if it'll be backward compatible from a software perspective?

SimonC on 6/10/05 9:58AM
[ Reply | Permalink | Report ]

It will be fully backwards compatible with instruction set that we are using on current systems from an application developers point of view, some of the X-Scale only instructions wont be supported, but there is very little non hardware specific code using those. There would need to be a few changes to operating systems for initialising the new level2 cache control mechanism, but the largest change will be due to the differing subsystems of which ever SOC (System On Chip) is chosen, which all new machines use rather than the old fashioned discrete processor and seperate I/O chipset.

druck on 6/10/05 10:41AM
[ Reply | Permalink | Report ]

FAB now how about a real ARM Powered Portable Laptop instead of an A9 MK2 with new mini graphics cards for portables, and RISC OS now being made to work with any graphics card that you can get a driver for, any one work for nvidia 8 ¬) , this may seem far fetched but not if you go to a laptop manufacturer and get the parts and the just make the motherboard to fit, who know they might just make it for you after all a laptop that can run for hours and not over heat, except the Graphics card ? just think a laptop that can boot from a big memory card holding your operating system and software RISC OS, windows CE, linux, MAC OS X = linux , that would make it cheaper all round, just by your laptop and get OS of choice. as far as backward compatible the 32bit compatible software should be as long as it is not directly poking the Vidc20 chip it should work after all there are emulators, mini emulator ? and graphics cards for the riscpc but you should be thinking forward compatible, new better software now that we have an FPU the arm 7500 was ok but just a bit slow for graphics.

runningman on 6/10/05 2:07PM
[ Reply | Permalink | Report ]

Col1: "Or the A9 2"

I'm still waiting for the first A9 to reach general release. My cardboard prototype is starting to look a bit battered!

datawave: "I bet Castle have then [2009] already Iyonix 3 in widespread use."

That seems pretty unlikely. Just as with the original Iyonix, you have to look at what Castle have said they are aiming at, in order to get an idea of how far away any future system is. What they've said so far is that any replacement for the existing Iyonix would have to provide a substantial performance increase (let's say, perhaps double the speed) in order to be worthwhile. Currently offered chip designs - some of which aren't even in silicon yet - don't provide that. So the Iyonix is likely to remain their top end machine for some years to come. But I'm sure the operating system development will continue in the meantime.

dgs

dgs on 6/10/05 2:29PM
[ Reply | Permalink | Report ]

The ARM Cortex-A8 processor is the first applications processor based on the ARMv7 architecture note ARMv7 have an ARM710 in my riscpc and the latest risc os select.

runningman on 6/10/05 2:31PM
[ Reply | Permalink | Report ]

In reply to dgs: Have you actually read this article? The Cortex A8 will provide far in excess of 2x the performance of the IOP321, and is exactly what Castle and STD should be developing their next generation of machines around.

In reply to runningman: the viability of a RISC OS laptop has nothing to do with the processor, suitable processors have been around for years, its getting a case manufactured thats always been the sticking point.

druck on 6/10/05 4:29PM
[ Reply | Permalink | Report ]

druck: "The Cortex A8 will provide far in excess of 2x the performance"

Correct, and my original point was that it won't be available as silicon for quite some time. After which, someone still has to build a system around it - not something that happens in a couple of months.

dgs

dgs on 6/10/05 4:43PM
[ Reply | Permalink | Report ]

@druck: Of course one could do it the other way round: Get a truck load of nice looking laptop cases with screens, keyboard, touchpad, etc. built in, then develop a mainboard that fits into it.

JGZimmerle on 06/10/05 6:43PM
[ Reply | Permalink | Report ]

@runningman: The ARM710 is a ARMv4 processor. That technology is about 12 years old. ARMv7 is brand new. ARM11 processors are ARMv6. Newest versions of RISC OS may make it easier to develop drivers for new graphics cards, but developing such a driver is still a huge task, any driver-development is. At least if you're not stealing or buying the code from someone else.

JGZimmerle on 06/10/05 7:15PM
[ Reply | Permalink | Report ]

Who is rating doing all the rating down on comments that are actually relevant to the article?

JGZimmerle on 07/10/05 4:43PM
[ Reply | Permalink | Report ]

Technical details on the Cortex-A8 at [link]

joty on 08/10/05 3:25PM
[ Reply | Permalink | Report ]

@JGZimmerle: i was just to optomistic, :¬) even arm were getting confused and are now giving new arm chips names like Cortex-A8.

say did you spot that little google advert at the bottom of the page it's gone now, have a look little arm developers mother board FPGA and all that stuff just like the Omega [link]

as the A9 mother board is so small take a pc laptop with a bad motherboard fault and cut out a space to fit the A9's motherboard, with the arm 9 above the fan grill, cut the laptops tracks to the ports, next solder wires to the ports fit ribbon cables drive and a cable to the screen, hope it can drive a laptop display, any one know, should have room in an old bulky model, might not in a new thin laptop if the keyboard is part of the motherboard. good reading.

runningman on 08/10/05 10:50PM
[ Reply | Permalink | Report ]

One of the disadvantages of ARM is that they just design chips, they don't manufacture them, it would be great for small companies if ARM got chips like this manufactured and then sold them at a fixed cost to people and small companies that would not be able to get them, just like AMD even colleges would like this, and if RISC OS could get it's foot in the door there could be a lot of little start up companies using this chip and RISC OS the trouble with RISC OS is that it is to expensive, for mini laptops and such a free copy for research would increase it's use and encourage small manufacturers to use it, if the price was dropped to say £10, this would encorage the larger builders like SHARP seen this, i wonder new version with Cortex-A8 :¬) [link]

runningman on 09/10/05 2:25PM
[ Reply | Permalink | Report ]

In reply to Kevin: Chips are sold in at a fixed cost in batches of 1000, however you need to order a minimum of 10,000 to start manufacture - we don't have that may RISC OS users left. But as I said the processor is even the remotest factor in producing a laptop, whether it cost £10, £1000 or £0.01.

P.S. Please don't hit returns at the edge of the edit area when typing in your message as it then looks a mess, just keep typing and the correct linewrap will be applied automatically.

druck on 10/10/05 09:28AM
[ Reply | Permalink | Report ]

In reply to ZGZimmerle:

The problem with getting a "truck load of laptop cases ..." is:

a) The cost b) where do you store them while designing the motherboard?

AFAIK most laptop case manufacturers want you to buy in such large numbers that the total cost is too high for the majority, if not all, RISC OS developers.

Col1 on 10/10/05 09:49AM
[ Reply | Permalink | Report ]

Please login before posting a comment. Use the form on the right to do so or create a free account.

+++ Message board +++

Read messages or start a new thread

Search the archives

Today's featured article

RiscPC breaks 1.5GHz barrier
Calm down dear, this is a case mod

23 comments, latest by flypig on 13/5/04 1:56PM. Published: 12 May 2004

Random article

Robot pet hamsters arrive for RISC OS
Daft desktop distraction device driver developed

14 comments, latest by davehigton on 8/8/07 3:38PM. Published: 5 Aug 2007

Create a new account
Forgot your password?

Useful links

News and media:
Iconbar • MyRISCOS • ArcSite • RISCOScode • ANS • C.S.A.Announce • Archive • Qercus • RiscWorld • Drag'n'Drop • GAG-News

Top developers:
RISCOS Ltd • RISC OS Open • MW Software • R-Comp • Advantage Six • VirtualAcorn

Dealers:
CJE Micros • APDL • Castle • a4 • X-Ample • Liquid Silicon • Webmonster

Usergroups:
WROCC • RONE • NKACC • IRUG • SASAUG • ROUGOL • RONWUG • MUG • WAUG • GAG • RISCOS.be

Useful:
RISCOS.org.uk • RISCOS.org • RISCOS.info • Filebase • Chris Why's Acorn/RISC OS collection • NetSurf

Non-RISC OS:
The Register • The Inquirer • Apple Insider • BBC News • Sky News • Google News • xkcd • diodesign

Â© 1999-2009 The Drobe Team. Some rights reserved, click here for more information
Powered by MiniDrobeCMS, based on J4U | Statistics
"There has been some confusing and controversial reporting of Microdigital's presentation"
Page generated in 0.0901 seconds.