Eureka unmasks Omega - RISC OS News, Software and Information

Eureka unmasks Omega

Published: 23rd May 2003, 19:44:00GMT Source: drobe.co.uk
By Chris Williams

Page 1 of 1

Shipping? We've heard of it

Correctly touting it as an exclusive, The ARM Club have this week emailed a pre-review of their Omega to their club members. A full in-depth review of Microdigital's otherwise AWOL StrongARM powered desktop offering will be published in the next issue of the club's Eureka magazine, we're told. You'll need to be an ARM Club member to get hold of this.

"A full review of the Omega system will be appearing in the next issue of Eureka, and a more detailed review may be sent out by email before then if we receive more information regarding the machine", ARM Club chairman Toby Smith confirmed to us earlier today.

As you may recall from earlier this month, The ARM Club was the lucky Wakefield show exhibitor to pick up a free Microdigital Omega in the WROCC's prize draw.

The WROCC prize draw chillingly excluded magazines (basically, the free press) from the Omega prize draw and ruled out exhibiting magazines Archive, Acorn User and Acorn Publisher from winning the Omega and being the first to review the machine. However, The ARM Club isn't specifically a press organisation, it's a user group that provides for its members, amongst other things, technical support, discounts and most importantly here, the quarterly printed magazine Eureka. So much for excluding magazines.

First impressions
"I found that it seems to run most software with no problem. The Omega is still a 26 bit machine, unlike the 32bit Iyonix, so there should be no compatability issues with old / non-updated software", The ARM Club preview reads, the highlights of which quoted to us by Toby. The preview sounds like it's not holding anything back.

"Benchmarking showed that, as expected, the processor is roughly 50% faster than a 200Mhz StrongARM Risc PC. The speed of memory writes peaks at roughly the same value as the Iyonix, but memory reads are only about half the speed of the Iyonix due to the use of faster memory in the Iyonix.

"The graphics are currently a little disappointing - moving windows in high resolution modes is noticeably sluggish but I understand that the graphics are not currently being accelerated.

"Overall I would have to say that this does not appear to be a production quality machine, but then I'm not sure whether or not it is meant to be and as MicroDigital have not yet replied to the questions I put to them about the problems with the floppy drive and the graphics system, I am unable to reach a firm conclusion."

We were unable to reach Microdigital for comment earlier this afternoon and our emails never get answered.

Now given that visitors to the Wakefield show were told the Omega would be shipping during the week after the show weekend, we assumed the Omegas being handed out in the prize draw were packaged up finalised products like those ready to be posted to deposit holders. Microdigital's Dave Atkins explained at the show that the Omega's FPGA configuration was not quite finished but then again, they've been saying that for the past two years.

As the machine hasn't arrived at end users' doorsteps, we can only hope Microdigital are racing into overtime to address the glitches in the video system, the lack of USB and ethernet drivers as well as patching up the problems thankfully highlighted by The ARM Club.

Beyond a joke
This is precisely why we need the machine to be in the hands of independent publications, or even still, in the hands of the users who placed their deposits no less than three years ago. There's no need for Microdigital to assume the place of RiscStation, now that RiscStation have quietly sneaked off.

Links
The ARM Club
Microdigital

This article has been linked to, or is available in the following formats:

[Printable] [Digg this] [Blog search]

simo 23/5/03 9:56PM	50% faster than a 200MHz SA RiscPC? So my SA287 is faster than this, my 233T not far off, and Kinetic300's leave it for dust? Doesn't seem right to me, especially for a machine that has [so far] taken 3 years to build. I guess the Iyonix has no real competition then.... -- #include "sig.h"
simo 23/5/03 9:58PM	Oops, thought it said 50MHz, not 50%, still disappointing that it's only like a 300MHz RiscPC though - as that's what the processor does, surely the memory/bus etc. makes the machine seem faster than say a Kinetic? -- #include "sig.h"
mrchocky 23/5/03 10:10PM	The speed results are pretty much as predicted long ago given CPU and bus speed. The 50% speed increase is raw CPU speed, which governs much of the performance. And yes, it will certainly be faster than a (Kinetic) RiscPC on many tasks due to the bigger bus. If you want a fast machine, get an Iyonix. I rather doubt seeing ARMTwister before the end of the year. -- Peter, drobe.co.uk
thegman 23/5/03 10:38PM	Maybe MD are afraid that some members of the press have already made their mind up about the Omega, and that's why they did'nt want mags/press reviewing it before it was available.
jlavallin 23/5/03 10:39PM	Oh dear No Omegas yet.....Mico motherboard replacement is getting futher away. Lets just hope it doesn't fall off the horizon.
mrchocky 23/5/03 10:48PM	tgm: I don't think so. The ARM Club did their very best to do a fair review. A Drobe review would have been similar (whatever anybody might think). Regardless, the numbers have been there to see for a very long time, and MD can only do more damage to themelves by pretending things are other than they are. -- Peter, drobe.co.uk
JGZimmerle 23/5/03 11:18PM	MD are not pretending that things are other than they are. If you (or whoever might have Omega hardware) do not have the very latest FPGA images you should not make comments about the machine's performance. And you have no way of knowing how soon ARMTwister will be with us, so this negative comment is just damaging the market. And there are people for whom the Iyonix is not fast enough in some areas, so to simply tell people they should get an Iyonix does not help either, since most people have different requirements. -- Julian G. F. Zimmerle
mrchocky 23/5/03 11:26PM	Of course I do not know when ARM Twister will be with us. But I am free to make my own personal predition based upon the performance of MD's deliveries so far. You are of course free to make your own (possibly based upon better knowledge). Please don't say it is "damanging to the market". I know you know better than that. Equally, no. I do not know the behaviour of the latest FPGA images. However, clock and bus speed place top limits on performance. If those have changed, then we'd be very interested to hear. If people really want a fast machine, they should get a PC. There is nothing that can change that. My comment is in reference to Simo's comment about speed, and should be taken in that context. -- Peter, drobe.co.uk
thegman 23/5/03 11:39PM	Chocky: We agree on something, I can be a bit of an MD apologist sometimes, probably because I see a lot of promise in the Omega (or at least it's spec). I don't think it's strictly true to say clock/bus speeds put top limits on performance because the Omega (if released with full spec) is supposed to have MPEG/OpenGL in hardware, which should make those two tasks faster than machines with far greater clock speeds. I'm very interested in what Julian has to say, as someone who apparently has close contact with MD. However until the Omega is in the hands of punters, they're in no better position than RiscStation (and that's a shtty position*). So maybe next time you talk to David Atkins, maybe you could ask him for some realistic time scales, I don't think many people would mind waiting 3 months for the full spec machine, but nobody likes to hear 'next week' umpteen times.
mrchocky 23/5/03 11:45PM	We too are very interested to hear what MD has to say - hence our frequent attempts to contact them, which Chris has referred to the in article. Unfortunately what they have said isn't a great deal, and we have only their published specs to go on. Yes, it's certainly true there are expections in any machine - although they technologies you refer to are very specialised and won't affect the vast majority of operation. CPU and bus speed must be crucial factors when considering the final performance of the machine. -- Peter, drobe.co.uk
mrchocky 23/5/03 11:47PM	"exceptions" is of course what I meant. -- Peter, drobe.co.uk
thegman 23/5/03 11:53PM	Sure, but as the 300Mhz SA is'nt going to be able to push around full screen movies by itself, hardware MPEG might be cool. OpenGL acceleration might get some games ported too. I appreciate the difficulty in getting MicroDigital to answer an email (I've tried), I just hope they won't treat customers like they treat potential customers.
Gregor 24/5/03 12:30AM	@Julian: I believe that MD is damaging themselves, unfortunately. Why does not MD answer to emails from The ARM Club who is an owner of an Omega? If you have actual FPGA images better than those in The ARM Club's Omega why does not MD send these images to The ARM Club? The ARM Club has won a prize or in my opinion a gift, but still an incomplete gift. It's very impolite what MD is doing!
Gregor 24/5/03 1:04AM	Sorry for posting my comment twice! I use an Windows-PC for emails and internet. I suggest that Microsoft's Explorer has sent this posting later once again because it was still cached. [Removed duplicate comment - Ed]
JGZimmerle 24/5/03 1:10AM	Well, the ARM Club should not have produced an article on their Omega, because the press was excluded from the price draw. So if they did consider themselves to not belong to that group, they should have stuck to that and not change their mind later on. As you can see on MD's website there will be a support area for owners of MD machines soon ("owners club"). AFAIK the first machines have been delivered to customers already. On the subject of limitations on speed: There are several buses in the Omega wich run at different speeds. If you were refering to the speed of the RAM, then we all know there is more to this than the clockrate (wich is actually lower on the Iyonix, but uses DDR technology wich does not* double the speed). There are many ways in wich a chipset can optimise RAM usage, caches being the most common one. -- Julian G. F. Zimmerle*
MrChimp 24/5/03 1:42AM	Julian G. F. Zimmerle = Muhammed Saeed al-Sahaf. "The machines have been delivered already. Delivery is an abstract concept. Your technical arguments are irrelevant. They are in a state of hysteria. Losers, they think that by producing facts and trying to distort the feelings of the people they will win. We shall still crush them at the gates. They are slitting their own throats. We are fully in control. My feelings, as usual, Omega... far superior. Much like my shellsuit." "We are winning".
Spriteman 24/5/03 1:55AM	Poor taste
NeilWB 24/5/03 9:05AM	*In reply to Julian:* I don't see any problem with the ARM Club reviewing the Omega. They are not associated with the media. Would you have felt this way if the review had been more positive? As long as the review is honest, based upon facts, and free from speculation, then I don't see that MD has anything to complain about. You say that AFAYK the first machines had already been delivered to customers, but I kinda remember you saying pretty much the same thing at Christmas. As an MD insider, perhaps you can explain why the Omega is still unfinished 5 months after we were told that outstanding delays were due only to contractual issues. I'm very curious. Regards, Neil
mrchocky 24/5/03 9:07AM	I agree. Tony's comment was not required. *In reply to Julian:* This is a very odd position to take. The ARM Club have done precisely what the vast majority of RISC OS want to see - hard information about the machine, even in its unfinished state. The Iyonix suffered no less scrutiny before and after it was released. Try as I might, I cannot find your points about bus speeds any more than hand waving. It doesn't matter how many busses you add, the CPU can only do so much. If you, or someone else, would like to write a technical analysis of the situation then Drobe would be very happy to publish it, and we can discuss the matter further. -- Peter, drobe.co.uk
imj 24/5/03 10:30AM	So all you care about is speed, is it? If it's not a 600Mhz XScale you immediately write it off as useless, nevermind the (supposed) great graphics, PCI interface, etc? I'm in no way defending MD because I think their choice of shared graphics memory is plain stupid these days - the CPU will be starved for RAM access in high resolutions, but you do need to consider the whole machine when comparing against RiscPC and Ironix. Oh yeh, and it doesn't run an outdated version of RISC OS
mrchocky 24/5/03 10:34AM	There's no implication that it's "all about speed" nor about writing it off cos "it's not a 600Mhz XScale". The only question is that the claims about Omega are questionable at best. If the functionality of RISC OS 4 = an outdated version of RISC OS, then there's not much than can be done. -- Peter, drobe.co.uk
Martyn Fox 24/5/03 10:52AM	If a 300 MHz StrongARM Omega is 50% faster than a 200 MHz RiscPC, this would seem to imply that the RiscPC wasn't held back by the speed of its bus after all. Either that or that the Omega has some similar constraint. Martyn
MrChimp 24/5/03 11:31AM	*In reply to Spriteman:* Poor taste? The shellsuit? This really isn't the place to discuss sartorial elegance.
mrchocky 24/5/03 11:53AM	*In reply to MF:* I don't follow your reasoning. The RiscPC certainly was held back by its bus, and the Omega certainly will benefit greatly from having a sensisble sized bus. An unqualified speed quote will generally just be a meansure of CPU speed (benchmarked of course). The point I've been trying to make is that except in some unusual circumstances (e.g. dedicated processor), the CPU speed dictates the final theoretical speed of a computer. Busses/memory speed and other items impose certain restrictions on that in practice (and in a RiscPC it's a bit sucky), which things like caches can help alleviate. It might help if I said that bigger busses help the machine go "less slow", even though that's slightly nonsense when analysed too closely. *In reply to Tony:* enough, thanks. -- Peter, drobe.co.uk
AMS 24/5/03 1:38PM	Well at least someone has finally managed to get Omega reviewed (phew). It perhaps would have been more helpful if MD had made the machine available to Drobe and ArmClub or whoever to review with the caution that "This is pre-release hardware" and that would have satisified curiousity and yet had that proviso in just in case. The results (at least from this article) seem not surprising. It does still sound like a machine that is a bit away from being in a production state. Performance wise it's better than a typical SA RPC, but (and I am only going on what's in the article above) the question is it sufficiently faster to justify the price. Without real benchmarks and finalised hardware this is a hard one to call. Part of the problem (as it were) of Omega is that because of the use of FPGA's it will always be possible for someone to say that "oh my versions faster because I have a later FPGA build" type excuse. The problem is there are limits to the speed improvement that can be had from tweaking the FPGA VHDL - fundemental limits will still exist (the 66MHz I/O bus on the SA for one, the RAM speed for another). Besides by this stage (the point of giving machines away as prizes) one would hope the FPGA's configuration would be more or less settled by now - and that all that should be going on is system software tweaks. At least it's nice to see some progress has been made to replace the RPC, but I would suspect that it will be a while yet before finished Omegas will be available. Keep plugging away guys..... -- Annraoi McShane,
sebgate20 24/5/03 8:42PM	It seems to me that MD have lost their chance. They seem to have spent three years faffing about and Castle have taken that opportunity. Can the Iyonix plays MPEGs?
diomus 24/5/03 8:44PM	The Iyonix can play MPEG2 videos, of course. There's a 32bit KinoAmp. Cineroma can also play MPEGs too and is 32 bit, but still unreleased. Chris, drobe.co.uk
JGZimmerle 24/5/03 10:44PM	Well, the RiscPC can also play MPEGs, but you really have asked the wrong question, wich should have been: Can the Iyonix play full-motion widescreen DVD resolution MPEGs at full speed? The Omega will soon, with it's hardware-accelerated DCD/iDCD and the FPA. The (minor!) video problems on the ARM Club's Omega have been resolved in the latest FPGA images. And if you used a TFT monitor then you would not have seen it at all. The floppy drive problem is almost resolved. And most people don't use their floppy drives anyway these days. I think it is unfair that the ARM Club's publication got a head start by winning an Omega, while the other publications have to wait for the official release. -- Julian G. F. Zimmerle
Spriteman 24/5/03 11:19PM	Julian, define 'soon' -- Spriteman.
The Doctor 24/5/03 11:28PM	I must point out that at the Wakefield show, we were told by David Atkins that the Omega would ship 'next week'. He 'must' have known it wasn't ready and therefore wouldn't ship.
Use the forum for more comments on this article