sa110
5/10/07 12:59AM |
Hopefully this new technology will mean faster RISC OS computers in the future. Quad core would be nice too. Now all we need is multiprocessor/core support in the OS. Perhaps if ROL start looking at this now, we me have an OS to drive this new design when it becomes available. |
stevek (+1.0)
5/10/07 6:49AM |
The most exciting thing here isn't the quad core stuff because it is pretty unlikely that RISC OS would be able to take advantage of it without a massive amount of work. The floating point maths bit however is exciting. Anyone who has tried to encode an MP3 will understand why...and that isn't the only thing that would benefit!
|
druck (+3.0)
5/10/07 9:07AM |
In reply to sa110:
don't distract ROL, they have the A9 to finish before looking at anything else.
Support of multi-core is fare more likely to come from the ROOL initiative, not with symmetric multiprocessing as that's pie in sky territory for RISC OS, but rather people taking various parts of the OS and using the other cores as off load engines. For example shifting the network stack on to another core, which is of particular benefit when using computationally expensive https encryption. Much of the PCI I/O subsystems such as disc and the USB stack could be placed on another, leaving one core to run the main part of the operating system and applications, and another available to share by custom written applications which do a lot of non interactive operations, an MP3 player for instance.
But even using the single core, there will be a substantial speed increase from the both the clock speed and the additional memory bandwidth which will be available to be able to feed all 4 cores. FP would also be a bonus, and would open up float heavy APIs such as OpenGL.
Trouble is 2010 is a long way away. I've just upgraded from a 2.1mPix to a 7.1mPix camera, and I need a machine thats 3x faster than the Iyonix before then to keep my productivity at the same level. |
arawnsley
5/10/07 10:45AM |
(dons flame retardent suit) druck - there are already plenty of people running RISC OS (approx) three times faster than the Iyonix and happily using 7mPix cameras with RISC OS apps. You don't need to wait till 2010, you know! But I know it matters to you how you run RISC OS, but if you can get over that, there's plenty of RISC OS speed available. It just takes the right combination of parts and correct optimization/configuration.
Back on topic, it is good to hear of new ARM chips, because the Acorn heritage means we all have a soft spot for the processors. A lot of IT press has been pushing the x86-for-all approach recently as the power requirements of x86 technology have tumbled, but it is good to see ARM fighting back. |
druck (+1.0)
5/10/07 11:15AM |
VPRC still does not perform 3x faster for computationally intensive Photodesk operations on the top of the range dual core intel here at work, and if I were to get one of those myself, I would be using the full native power with non RISC OS software.
I'm afraid it either has to be a genuine ARM based system running RISC OS and only RISC OS, or if I have to buy an x86 box with the effort of maintaining another OS, that's what I'll be using most of the time. VRPC will eventually become relegated to just retrieving the odd old Impression document or some such.
ARM based RISC OS hardware is the only thing that makes me interested in writing software outside of working hours, otherwise I'll spend the entire time inside of FireFox. |
jess
5/10/07 11:23AM |
In reply to Druck:
I'm curious, is it the Arm processor or the fact the OS runs natively? ie would you prefer ARM Linux (with ROX or ROLF) or an RO port to x86? |
arawnsley
5/10/07 11:54AM |
With respect druck, please don't try and compare your own results with a self-built/off-the-shelf PC and those of a RISCube. I know you think you can, but just to give an example, a chap brought in a fairly high end core2duo system for me to work on 2 weeks ago. It ran about a quarter of the speed of a similar RISCube due to the selection of parts and configuration. Obviously it will depend on the program, and I don't know the insides of photodesk, but I've had computational tests running over 5 times faster than an Iyonix. Clearly it will depend on exactly what the software is doing (just as an Iyonix varies wildly), but it is perfectly possible. Anyway, I know that for you (and many others) only a native ARM cpu is acceptable, and that's cool, but for other people, it is still possible to stick with RISC OS and have high performance. Take the guy who received his RISCube this week - his most-used computational operation dropped from 3minutes (RiscPC) to 3 seconds. He has his machine configrued so he never needs to look at Windows. |
Piers (+0.9)
5/10/07 2:24PM |
Um... Is it worth pointing out the Cortex has an different instruction set? There's no way RISC OS will run on it without significant changes - probably at least as large as the 32-bit project. |
jmb (+5.2)
5/10/07 2:44PM |
In reply to Piers:
By my reading, it's just the ARMv7-M variants of the Cortex family which don't support the ARM instruction set, running only Thumb2. Both the -A and -R variants support ARM, Thumb and Thumb2. |
Piers (+2.0)
5/10/07 2:56PM |
In reply to jmb:
Ah, fair enough. I hadn't realised there were two variants of the Cortex (M and A), so read the wrong one.
I do wonder how many instruction sets a single embedded processor needs 
|
Revin Kevin (+2.0)
5/10/07 4:01PM |
Nice to see ARM chips getting faster.
Hopefully this will lead to new ARM powered RISC OS machines appearing. |
bluenose (+1.0)
5/10/07 7:31PM |
In Reply to AW
So exactly where are you going to buy this new ARM processor from then as it won't be avilable until around 2010 by all accounts.
As to the ROOL initiative it seems that some more components will be available for the SEShow and that it will include the !Printers elements that will allow for some work to be done it to overcome some Postscript limitations. |
AW (+2.0)
6/10/07 1:09AM |
In reply to Bluenose:
the reason for advocating genuine RISC OS hardware is more than sentimental. It's the very fact of having a distinct alternative to cosmetic variations on a PC box. There have been other ARM options posted on this site faster than the Iyonix. |
killermike (+1.0)
6/10/07 3:09AM |
A 1ghz processor will seem very slow by 2010. Particularly on an OS that can't fully take advantage of the multiple cores. By that time, the emulated solutions will be even further ahead.
I understand people sticking with the ARM powered desktop machines for the time being, but in 2010, emalation will be about 10 times ahead of the real hardware.
Even a more advanced form of emulator that could use the features of the underlying OS would be more of a step forward for RISC OS. Imagine how quick a version of photodesk could be if it did some of its processing using the native build of imagemagick but controlled from RISC OS.
I'm of the opinion that media/3d acceleration or expansion into other form factors and roles (PDA, set-top etc) would deliver far more benefit to end users than a small speed increase by 2010. |
bluenose (+1.0)
6/10/07 10:02AM |
In Reply to AW
Totally agree with you on the native hardware front. I've got an Iyonix and would love the option of a faster version but crucially I would also say that we need the software to take advantage of that faster speed. There have been a couple pof faster processors mentioned before including one from Intel.
In Replay to Killermike
What I would advocate is a linux kernal that boots up a basic system and ARM Processor emulator and then fires up RISC OS. That way we can get to use off the shelf components. Perhaps specially adapted versions of RISC OS software could then run virtualulised versions of native Linux applications for the items we miss.
All of this of course won't happen if people don't carry on buying RISC OS native hardware or software. |
JGZimmerle (+1.0)
6/10/07 11:33AM |
May I point out, that soon energy-hungry computers will become a lot more expensive to run, as the world runs short of its primary energy source: petroleum. This is also true for their manufacturing. So computer systems like RISC OS, wich use small, efficient chips, will have a real advantage.
For more info look at http://www.lifeaftertheoilcrash.net or watch the movie "A Crude Awakening". |
Cogs
6/10/07 2:31PM |
In reply to JGZimmerle:
a) Oil is generally not used to generate electricity, so high oil prices do not necessarily push up electricity prices. Gas prices are more relevant, at least in the UK, but while gas prices to the consumer have more than doubled in the past 10 years, electricity prices have gone up below the rate of inflation.
b) The efficiency of the processor is only part of the equation. If you use a CRT monitor with an Iyonix instead of an LCD monitor, you probably use more power than a budget PC with an LCD. Another big saving would be to buy an efficient PSU. A typical ATX PSU is only 60%-70% efficient, but if you pay more you can get one that is fanless and up to 90% efficient. And one can always downclock ones PC in the BIOS, if obsessive about power consumption, and it would still probably be quicker than an Iyonix.
But all this may prove irrelevant. More and more people now want laptops, and there isn't an ARM laptop (A4 aside), nor is there any prospect of one in the future.
Actually the refrigerator is the biggest power consumer in the home. My fridge and freezer are B rated, which I assumed was good when I bought them, but they probably run at around 250% of the electricity usage of A++ rated ones! |
jess
6/10/07 2:59PM |
In reply to cogs:
Do you have any good recommendations for ATX PSUs? (one that is still high efficiency at Iyonix power levels.) |
mripley
6/10/07 6:53PM |
Laptops are the way to go! I have no intention of purchasing a "desktop" computer again. So inconvenient and they require you to be hidden away in a study room somewhere. A laptop on the other hand can sit on your lap in the front room and accompany you wherever you travel. Most people have a wireless router these days so your computing can travel with you. I now a new found enthusiasm with Linux on my lap  . Oh forgot mention all my music and video sits in that study on served drives wirelessly accessed. |
epistaxsis (+1.0)
6/10/07 7:33PM |
In reply to mripley:
Guess you're not actually using your laptop on your lap.
They get so hot that the term "laptop" is now a misnomer - especially the ones that actually have CPU cooling vent on the underside! (we have a few of them at work - classed as non-lap laptops...)
For transportable RISC OS computers I'll stick to my A9 thanks |
Cogs
6/10/07 8:55PM |
In reply to Jess:
I haven't tried it, but for an Iyonix you could probably get away with using a PicoPSU (http://linitx.com/viewproduct.php?prodid=10916) in combination with a laptop power adapter. I think an Iyonix is likely to consume 50 to 70W in general usage (disclaimer: I've not bothered to test this). Therefore a 120W PicoPSU would seem easily well specced enough. You should be able to pick up a PicoPSU & suitable laptop power brick for about 50 quid.
For a typical X86 desktop PC though, it's just not likely to provide enough power to cope with peak loads. The only fanless ATX PSUs I've managed to track down above 200W all cost 90 quid or more, and that put me off. |
timephoenix (+1.0)
6/10/07 11:23PM |
What a shame hardware development is so capital intensive. Otherwise we could form an open source cooperative like ROOL for hardware... </dream>
The question still remains - what is the biggest gap for RISC OS today: software or hardware? About a year ago I would've said software, now I'm beginning to think both. |
stevek (+2.0)
7/10/07 12:38AM |
RISC OS will never compete with PC's on raw processing power. So we need to think outside that square and position RISC OS in places where it can compete.
People general don't by a computer for the operating system or the hardware, it is usually for the applications running on the hardware / OS combination. For RISC OS to survive it needs to offer something other systems don't.
Lower power is definitely a key capability of ARM processors and this is why they are commonly used in PDA's and mobile phones. To leverage that capability RISC OS companies need to target markets where low power is a requirement e.g. providing wireless web browsing and email etc for boats, mobile homes, camper vans. Perhaps a modified A9home?
However if power saving is the best we can come up with I am worried. Look at the premium price you are paying to save money on the power bill...We need to have a much stronger arguement that just RISC OS savings money on your power bill. I agree this is part of the package but the full picture must be looked at.
For RISC OS to succeed we need a number of things from what I can see:
- Faster processors - MHz do matter
- Floating point maths - to improve multimedia performance, Open GL etc
- Faster OS development - so the boat isn't missed
- Support for modern hardware - printers, USB2, SATA disk drives
- Improved Internet Applications - Flash, Javascript etc
- Applications that differentiate RISC OS in a positive way e.g. the way ArtWorks does.
- Better marketing
- Positive user base
- More sales...
In the marketing world people will tell you that it requires far less effort to make a sale to an existing customer than it is to gain a new one. Something needs to be done to keep people on board or everyone is just flogging a dead horse.
|
JGZimmerle (+1.0)
7/10/07 8:11AM |
In reply to Cogs:
True, oil does not account for the majority of electricity generation, however it is used to meet short demand-increases (consumption-spikes). But what do you think will happen when oil-prices start sky-rocketing? IMHO cunsumers will start to switch their oil-powered devices to electricity, which might very well create a severe price-increase in electricity as well.
Also much electricity is generated from natural gas, which is expected to peak just a few years after oil. And oil is also used to build and maintain power-plants and to obtain their "fuels" (like uranium or coal).
Finally great amounts of oil and oil-derived chemicals are used to make computer chips, so computer chip prices should increase with the chips size. ARM chips are very small and so a lot more of them can be made from each wafer than x86-chips.
But all of this may be irrelevant, if the oil-shortage should lead to a collapse of the global economy and we might have to obtain the power to use computers our selves, for example by using the photovoltaic cells on our rooftops, if you were fortunate enough to obtain some before the oil-crash, or a hand/feet-powered generator. In this scenario I am glad I have an a9home and a flat-screen which I could rig to use natural light instead of the built-in back-light. |
AMS (+1.0)
7/10/07 10:35AM |
killermike wrote>"A 1ghz processor will seem very slow by 2010. Particularly on an OS that can't fully take advantage of the multiple cores. By that time, the emulated solutions will be even further ahead"
Part of the motivation of putting multiple cores on the same chip (or at least in the one package) is that each does not need to be clocked at multi-GHz speeds to acchieve overall reasonable performance (many PC laptops today using CoreDuo's are clocked at 1.6GHz - in the past single core PC's were being clocked in the 3.5->4GHz region). Emulated solutions (in speed terms) don't have much of an advantage and that's not likely to change. If faster single or dual core Intel xScales (c. 1.2GHz) or these ARM9 based multi-cores (c. 1.0GHz) appear they'll improve performance of the native systems by a lot more - I don't frankly see much change in a two year timeframe on x86 performance (other than adding 3 (AMD) or 4 cores (AMD & Intel)) and that *won't* necessarily translate to much emulator speed increase as that is largely clock rate bound which may not change much.
The key issue is we need a new computer based on faster ARM cores (be they quad - in 2 years or dual xScale now). The number of cores is probably less significant than the (a). Clock speed and (b). The Cache size. So a dual Core xScale now at 1.2GHz may be a better platform than a quad core 1.0GHz ARM9 in 2010.
In either event one version of RISC OS (ROOL's release of RO5.xx) offers an open option for upgrade, Castle/Iyonix have experience of manufacturing whole ARM based computer systems they'd be the logical ones to push this. ROL could do OS development instead - but have a number of mountains to climb (i). Their OS is largely intended to run on machines identical to RiscPC's, (ii). Emulated to look like a RiscPC or (iii). Have restricted expansion options [hence requiring less driver/hardware related OS development] e.g., A9Home. For them (ROL) to produce a multicore aware RISC OS would be a challenge for those technical issues thrown up and considering their other vendor and customer committments. On the one hand if they produce an OS that can make native RISC OS run faster than emulated that could hit one of their big licensees. If they abandon their RiscPC base to concentrate on a multicore system then they alienate their subscription customers.
I wouldn't absolutely rule out ROL doing something in the multicore area - but it would be yet one more speculative demand with so many others demanding attention (RO for A9, Select updates etc.,) and that just makes it difficult for them.
It would also mean them trying to persuading a hardware vendor to push into a full computer design with multicore capability. That would not be a typical embedded design (like A9) and might not appeal to Ad6 whose focus is embedded - so who would build such a mytical ROL inspired computer? Short of ROL themselves commissioning a new machine I can't see this happening, that having been said I'd be happy to be proven wrong.
If (and it's a big *IF*) RISC OS appears on new hardware it'll more likely to be a derivative of Iyonix running RO5.xx. In either event IMHO something needs to be done soon regarding an Iyonix replacement and multicore (even dual ones like Intel were touting - never mind quad quad A9's) may be the ticket. |
nijinsky (+1.1)
7/10/07 1:26PM |
Dons flameproof suit...
To Me RISC OS is an operating system. It runs RISC OS apps. For me that is the end of the story. If it runs on ARM, x86 or indeed is powered by a hamster on a wheel.
What I simply want is a faster desktop experience and a more stable experience. If that means running on emmulation via WinXP then fine. If that means porting the OS to x86 then fine. If that means sticking with the situation that my image files wont load at a respectable speed if at all then I won't use RISC OS. Therefore, for me the OS has to run on cheap mainstream hardware at the fastest possible speed. FYI My image files are 50-500Mb/image. YES 1/2 a Gig.
Cheers
Bob |
Cogs
7/10/07 3:33PM |
In reply to JGZimmerle:
We are going waay off topic. But I'm going to be naughty and reply anyway because I'm bored.
I don't share the hysteria that predicts oil prices will "skyrocket" any time soon. We may see a gradual doubling or tripling in price over 10 years, but no economy rich enough to be buying computers will suffer from that. It's the third world that will have to forego cars. This is already happening.
As for a switch to electricity from oil, I don't see that happening. An investment would have to be made up-front that would negate the long-term advantage of switching from oil. Ie: It wouldn't work out cheaper than oil, even at a higher price, because of the costs of new battery powered vehicles.
More likely is that people are forced to cut back on their mileage, or switch over to biofuels. The people who do buy new cars will aim for more fuel efficient models (hybrids?), but the average age of the cars on the road is more than 10 years, so even a switch over to hybrids would be gradual.
Noone using heating oil would consider moving to electricity either. The losses at the generation end will always make it more economical to just burn at home whatever the power stations are burning.
If gas prices rise too much, the energy companies will switch back to coal. There's plenty of that to last for decades.
Bringing us slightly more back on topic, let's look a the cost of running my PC. Let's say it draws an average load of 150W and I use it 10 hours a day. That would currently be costing me something like £50 a year. Even at 10 times the cost for electricity, it's not going to bankrupt the economy on computer use alone. |
JGZimmerle (+1.0)
8/10/07 6:30AM |
@Cogs: You clearly did not read the article I linked to. All the reasons you list above why peak-oil might not be an issue have been taken apart in that article (and each one in more depth in many other places) already, so I'm not going to repeat it here.
As the richest and most powerful people and some of the biggest companies in the world prepare for peak-oil, I don't see why I should ignore the problem. So me and my family are preparing for it, and that includes having very-low-power ARM-based computers available.
Finally, if peak-oil should really turn out to be a non-issue, we will still be better off, as most of the things we do to prepare for it, make good financial sense anyway. |
| 2 comment(s) are below your moderation threshold. Login to view them. |
| Use the forum for more comments on this article |
ARM set out its next-generation processor design stall this week by unveiling a four-core CPU clocked at 1GHz. But chips featuring the Cortex-A9 technology will not arrive until 2010.