Slashdot Mirror
Build Your Own Crusoe-Powered Computer
jonmason00 writes "Just checked the Transmeta webpage, and discovered that they are now offering a Crusoe TM5800 System Development Kit. It's a bit expensive ($995) and you gotta register before you can buy one, but they need your support." How about an Astro development kit instead? :)
Transmeta is a fucking business for God's sake, not Greenpeace.
They came up with a crappy business plan at the wrong time, and natural selection is taking care of it. Why try screwing the process up?
Sorry, this might be an unpopular view because Linus works there, and all, but no company needs my support if they've got a product I find useful for a price that is competitive with other companies, and I won't waste my money supporting someone else's unsucessful ideas, just because they've got cool technology or cool employees.
Or, build your own Intel- or AMD-based computer that's cheaper, faster, and uses less power.
Transmeta had a great goal when they started four years ago to reduce power use in their chips and allow for code-morphing, but it's now 2002 and mobile Intel and AMD chips are faster and use less power. And don't even get started on desktop CPU comparison...
Intel and AMD have both signed on to Microsoft's Palladium program. We need a chip maker who hasn't succumbed to this yet.
A crappy marketing strategy is no reason to write off an innovative technology [and yes, for once I believe the word is used rightly here.] The lower power consumption specs don't hurt either.
Use ISO 8601 dates [YYYY-MM-DD]
And, just for the record, this product isn't one I will be buying. It is way overpriced for what you are getting. A comparable barebones motherboard + CPU based on Intel or AMD could be had for an order of magnitude less money without requiring any kind of signup deal.
It's not our responsibility to keep the company afloat. I'm all about supporting software projects that obviously serve the open source movement, but a business faces risks when playing in the processor arena. I am happy to pay money for projects like the free blender foundation. Transmeta is responsible for getting their processors put into retail products, we are not responsible for donations.
Sir Timbly of Cannatuna, offical Knight of the Heptagonal Table
I bet even the mobile versions of the aforementioned processors are cheaper and faster.
--sdem
What's with this "you gotta register"?
Why can't businesses be content with getting my money?
Oh, because they're not selling commodities anymore, they're turning their customers into commodities to be sold.
Shall I supply my blood type and a DNA sample and a piss test and fingerprints in addition to my social security number and home phone and mother's maiden name?'
Get out of my life! Sell me your product and go away! I am not a number, and I don't need to be in your database!
Opinions on the Twiddler2 hand-held keyboard?
or you could just use via's mini-itx multimedia platform for around $100 (board, cpu, network, sound & video).
mini-itx.com
I can see the attraction for a kit like this. For those of us who might want to put together some type of home theatre pc or car pc something _like_ this would be nice. With the low power/heat requirements, it would lend itself nicely to such uses. IF it were a lot smaller. It's just too big. Plus they are just competing with all the other microatx mb's out there. This is the feature set that I think would actually be attractive:
- make it much smaller, all small as feasable, don't worry about making it a "standard" size
- chuck the floppy/parallel, keep one serial for programming
- add 2 firewire ports for camcorders/ipods
- sell a 12v converter for car/boat apps
- allow the sound/video to be upgradable somehow (logic on sodimm?) to allow upgrading.
- of course, drop the price.
If they had done these things, then I would be most interested since I have several projects that are begging for a formfactor smaller than microatx, but yet doesn't force me to use crap audio/video.
Hell if you're going to go for a niche, then GO FOR THE NICHE, don't come out with a "me too" product that has very little to distinguish itself from the competition.
I mean, I understand they are trying to target the "non-desktop" but were exactly does paying $1000 for a development kit prevent someone from having to come up with a custom motherboard fab for a set-top or other embedded device? Where exactly is paying $1000 (and I really want to know this, there might be an answer) help you in writing code for embedded device on a processor that is suppose to be x86 compatible?
I'm really missing the point here... But IMHO, it's more likely they are leveraging the geek market to help promote other mobo manufactures to build transmeta boards, so they can sell more chips in the long run..... And prolong the time it takes before they reach the fate that Cyrix did (bought out, low market share, nitch low-budjet/low-preformance x86 compatible chips).
Similar to the way that an Athlon 1600+ is faster than an older 1600mhz Tbird, the VIA C3 processors are nowhere near as fast as equivalently clocked Pentiums or Athlons. I'm highly skeptical that a 900mhz C3 is close to the performance of a 900mhz Crusoe, which by most accounts performs as you would expect a near-1ghz processor to do.
But man, Transmeta has totally missed the boat by not making basic, affordable computers available to hobbyists. FlexATX and C3-driven Mini-ITX boards are enjoying the kind of hobbyist popularity that helped put AMD on the map a few years ago. This $1000 "developer board" is too little, too late, and too much freakin' $$$!
// I will show you fear in a handful of jellybeans.
It's now, what, 4 years later and they still only have x86. Shame on you guys for lying to me like that!
I don't think they lied, it all has to do with reality. The reality is that the cpus have not sold in the quantities that they would have liked. In order to spend money on resources to come up with these other "personalities" you have to get the money from somewhere. I can't imagine that it would be cheap to develop such a thing and make it performant enough to justify the expense. Like the original poster of the article said, their hurting and it sure doesn't make any sense to pour money into supporting other chips that don't have nearly the market as the intel chips anyway. plus one of their claims to fame, low heat dissapation, isn't as big of a deal with the PPC since the chips run at lower clock rates and use less power in general anyway. Do cut them some slack, nice idea, just no money in it, sounds familiar.
I don't know...I was just trying to address a lot of the misdirected bashing in the other comments. If Transmeta is trying to enter this market, there's no reason for them to engineer a 200MHz CPU when they can throw in one of their more powerful models. Obviously there will be a point at which their Crusoe can handle more than a StrongArm, which allows for more features or the addition of unanticipated ones.
Sort of OT; but why isn't the EPIA series, especially the 6000 model with 800MHz C3 not gaining more attention? It runs fanless, has TV-out on the board, can take 1GB of ram and has everything you could need on the board, throw it into
one of these which happens to have an external PS and a front USB port and was pretty much designed with the EPIA 6000 in mind. I've harped on it before, but why isn't this the preferred RYO PVR platform?
The form factor fits right in with your vcr, and the case comes with a pci-riser card to fit that AMD AWIP card for TV-in, and excellent 3D. Nearly silent and the price is right (the Mobo/Case costs around $195). You could make a fantastic PVR/PC-compatible-gaming-consolesque system for around $500. $195 +$100(AIW) +$75(80GB eide hdd) +$60(1GB pc133 ram) +70(DVD/CD-RW comb0 drive). Add it up. And that's a lot of overkill on the ram, but hey, it's cheap. So why isn't this a story? Because it's not coming from Intel, AMD, or Transmeta? All of this "digital convergeance" is leading to more and more embedded arenas, if we can make a box like that at retail prices, think of what we could prototype for our own foray into the embedded market...
put the what in the where?
I really liked the promise of the Crusoe. I thought it was very clever technology.
:)
The problem is that the marketing was terrible.
Crusoes were targeted at laptops. Problem: In a world where I can buy a Celeron 1500 laptop for USD 1000 or less, you'd better do something to impress me if you wanted me to buy a Crusoe. You can make it cheap, which you didn't (the only Crusoe laptop I see locally sells for ~USD 1800). You can make it last a long time on the battery, but I can get similar battery life for less money by throwing an external battery on the back of said $1000 Celeron. You can make it perform well, but you didn't.
Make it competitive, or make it cheap. Transmeta (or perhaps more likely those who made products out of Transmeta CPUs) chose neither.
I was surprised you didn't see the Crusoe landing on the corporate desktop too. There are plenty of places with 300-400-500MHz machines that are adequate (because they don't feel the need to play games). I would think that a Crusoe-based box would be ideal for that sort of role (as a replacement for dead boxes, or to put analogous boxes on more desks)... with limited cooling requirements it would be quiet, not belch out tonnes of heat, and save maintenence hassles by removing moving parts (fans).
I also assumed you'd get a lot more flexibility with the "morphing" technology than you actually did. When Transmeta made the necessary documentation to exploit this stuff hard to find, they cut off their nose to spite their face. Why doesn't that "new Amiga" someone promises every few years run on a Crusoe set to emulate a 400MHz 68060? Why don't you see one-off Crusoe boxes designed to simulate obscure and otherwise irreplacable hardware. Hell, why don't you see them set up for student projects at universities to emulate the architecture they want to use as a teaching tool?
What did get exploited was that 'Web Pad' deal which never materialised. Now, we have this 'Tablet PC' thing... which will probably immediately bloat its way out of the Crusoe's performance range.
I had hoped that someone would buy Transmeta and do something useful with the technology. Now, I hope that when Transmeta finishes circling the drain, the local dealers start dumping Crusoe laptops cheap.
It's just like a fascist dictatorship, without the punctual rail service!
Yes, Ditzel would want to show Transmeta as pandering to the Windows market - after all, how can a HW manufacturer stay in business if it doesn't?
But remember: there will be a fast-growing market for non-DRMed CPUs. If Transmeta play their cards right [get better marketing and business managers, dammit!], they can show the PHBs of the world why it is a bad idea to invest in hardware that they do not have full control over.
Thanks for the informative post.
Use ISO 8601 dates [YYYY-MM-DD]
Do they play DVDs? Set-top box with integrated DVD and PVR functions.
Crusoes are really tablet/laptop boxes, though. This is intended to allow people to kludge together a laptop rather easily - hence the daughterboard which supports a battery.
Or a server board, to be honest - when you have dozens/hundreds of servers, small power savings add up in a power budget quite quickly.
People here keep suggesting the VIA Eden platform as being "just as low power" - bull. Transmeta processors are extremely power efficient. Just because something runs fanless doesn't mean it uses very little power - it just means that it has a power dissipation that can be handled by a heatsink only.
I doubt the Eden platform is actually lower power than the Crusoe platform. Running fanless simply means that the processor has a low power dissipation - that is, W/mm^2 - power per area. It doesn't necessarily mean that it actually uses low power.
That, and the Crusoe devkit is basically designed for a laptop. You can't buy a charge controller/keyboard/touchpad interface for the Eden platform.
Not to mention the fact that you get schematics, as well. For $995. That's cheap.
It's a pity that transmeta didn't do this a few years ago when it first launched. A lot of small-time developers like me were looking to incorporate Transmeta products into internet appliances, handhelds and wearables. All we (the little guys) got was a load of snobbery; if you didn't have a billion dollars in the bank, they didn't want to talk to you. It didn't matter how good (or bad) your idea was, it just mattered how many millions of units you had the potential of financing and manufacturing per month.
Whatever happened to sowing a thousand seeds and seeing which one bloomed? Whatever happened to grass roots, passion, and evangelism? It wouldn't have hurt them to at least share the specs and allow anyone to develop innovative product demos. Transmeta needed all the friends it could get when it was getting started, but their fat-headed sales reps felt they knew all and they were wise enough to hand-pick the few partners that would carry them to success.
I've seen too many products die because of policies like this. "We don't have the resources to support the little guys thus engaging them might tarnish our reputation" "We don't want shoddy mom-and-pop shop products ruining the name of our fine product" I'm sorry, but even the big players turn out some flops, and sometimes the most visionary products come from holes in the wall.
It's a development kit, you dumb fuck.