Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. ×
Transmeta

Crusoe To Be Used By Netwinder, IBM, NEC, Others 89

theGEEK writes "Rebel.com will be making Netwinders with the crusoe chip from Transmeta. In related news, Fujitsu, Hitachi, IBM and NEC will all be showing off notebooks using the Crusoe today."
This discussion has been archived. No new comments can be posted.

Crusoe to be Used by Netwinder, IBM, NEC, Others

Comments Filter:
  • What's the point? The major point about emulation is to tap into that unbefuckinglievable large pool of software for the x86. Alpha? What software? Dragonball? Get serious. You don't need fancy hardware to emulate it faster than any existing hardware implementation. The Dragonball is _slow_.

    Don't forget that what's interesting about the Crusoe is not just that it emulates another ISA, but that it optimises and reoptimises at runtime. That's a whole new ball game and it's just getting started.

    Anon

  • When you add in the cost of electricity for an always on computer, the Netwinder pays for itself in about two years compared to a Intel-based PC. It uses only about 7 watts average and 15 watts max.

    I'm really curious how this will compare with a Crusoe-based Netwinder. No fan has to help. And no multimedia should mean that the Crusoe will run pretty cool.
  • With the 486, they went from 25 mhz to 33 mhz, which was good!(~30%) The Pentium went from 50 to 55 to 60 to 66 mhz.(~10%) But now that the chips are running at 700 - 800 mhz(~4%), a 33 mhz boost is just stupid. Intel is playing with their customers.

    I won't buy a new CPU until I can really see a speed boost in it. AMD will allow that to happen faster.
  • I was at PC Expo yesterday and got a chance to see one of the IBM laptops running on the Crusoe chip. It seemed pretty good; the speed varied from 333MHz to around 600 or so (of course, depends on what you're doing). The part that I was really interested in is that the laptop is as light and "sexy" as the chip - they claimed that one of the laptops was going on about 7 hours, though the IBM representative claimed that 8 hours was the holy grail so to speak. Heck, it beats the 2 hours MAX that my current laptop provides. Remember, that the chip is low power, but that is only a fraction of the equation for battery life, so while the IBM laptop might get 7 or 8 hours, someone else could build something far more efficient.

    Oh, and speaking of fans, Transmeta had by far the "coolest" expo gift at the show: portable mini-fans for cooling yourself while walking around the Expo (which didn't seem to have very good air conditioning for some reason).

    --

  • I Love this intel now has to run for its money they have to compete with Transmeta
    and the best part is Linus Torvalds works there!
  • Less Heat eliminates need for a processor cooling fan. eliminating the need for a fan makes the package smaller. see also #1.

    If less heat also means no fans, i.e. low noise, and a neat small case that easily fits onto any desk, I'd seriously consider a Crusoe based desktop PC once the one I have now becomes obsolete.
  • Actually, Microsoft made a lot of effort trying to sell their software on other platforms too, up until a couple of years ago.
    Windows NT3.51 came in a version for x86, MIPS, sparc, Alpha and PowerPC. NT4.0 still comes in a version for the Alpha architecture, but is no longer supported.
    Microsoft ported their most popular apps, and gave complete compiler suites away for the Alpha architecture trying to get companies to support the Alpha-NT platform. Together with digital they _did_ write an emulator for Alpha-NT, that enabled you to run x86 binaries, with various degrees of success (some apps were faster, but most were slower than on the competing x86 machines, and stability suffered).
    To no avail, no one was interested. Not extremely surprising in retrospect.

    You seem to forget that unlike Apple, Microsoft doesn't make PCs. They only make software. Because the design of PCs is not controlled by one company, but to a very large degree by the consumer ("Can I afford the hardware? Will I have to buy new software or can I use what I already have? Will I be able to keep using my old printer/soundcard/logo controlled lawnmower?"), no one can simply force the PC architecture to switch to a different processor.
  • No, I've tried it, and as I mentioned, x86 binaries ran slower than on comparable x86 systems, and stability suffered (native apps did ok, x86 apps tended to crash, makes the regular PC user feel right at home).
    I also had a full visual studio suite running. Most sanely written win32 applications would probably only have needed a recompile on the Alpha-NT box, to

    Emulated x86 on a different platform was just never viable, and MS knew it, because the price performance ratio of the x86 PC can not be matched. It might have worked if the Alpha-XL series had been able to run PC software several times faster than a PC, but this wasn't the case.
    So people who chose for the Alpha weren't interested in x86 application software, and they were far better off running Digital Unix or linux.

    How can a new processor be successful when the OS on 95% of the world's computers only talks x86 ??

    I almost agree with this, but I'm convinced it's the applications than matter, not the OS. The popularity of the PC architecture has accellerated the development of this architecture so much, that experimental new architectures could never compete in terms of price/performance ratio.

    To replace what everybody's using, just being a bit faster is not enough. It has to be a lot faster, cheaper, more reliable, you name it.
  • Cut & paste remains something better left to the professionals :)

    Most sanely written win32 applications would probably only have needed a recompile on the Alpha-NT box, to build a native alpha NT version. How many 3rd party developers used that to port their apps?
  • you will have to excuse my ignorance, but what is
    so obsolete about the core that Sun uses? are sparc's not good procs?
  • What about electronic paper as display.. wouldn't that drop the price and power consumption?

    I think it could still happen.
  • I have no idea if it's possible with Crusoe, but I would love an SMP box based on Transmeta chips.
    As far as I know, it's not possible with the current crusoe's, due to the cache or memory logic being on the chip itself. Someone around her probably knows for sure though.

    I guarantee 8 or 16 Crusoe 700s will beat any two PIII 8xxs (1GHz PIII is not certified for SMP).
    Depending on the OS... Since 90% of the world is Win 9x, and MOST (not all, i know) of the market for SMP machines consists of servers, a cheap, low volume 8 or 16 way SMP platform wouldn't really be that commercially viable.

    Of course, for various reasons, the consumer market is anti-smp. Intel would rather sell you one
    super-expensive chip than two cheap chips.


    You'd add an incredible about of cost at the motheboard level by making it capable of handling 16 CPU's... When you're looking for performance, I'd think that people would be better off with fewer more expensive higher performance parts rather than many cheaper slower parts....

    You could bring the RAID analogy into this, but RAID has been revised to mean Independant, not Inexpensive....

    Lastly, of course 8 or 16 crusoes will beat 2 PIII's... but once you've added up the rest of the costs, I'd bet that the pricing would level out once you got past 4 or 8 chips in a box... Or else the Intel machine would take off in comparison to the Transmeta box...

    But since we're talking speed, why not just get a dual alpha and be done with it?
  • That used to be Henry Ford's gimick:

    "My customers can buy a Model T in any color they want, as long as they want BLACK."

    It may be funny in retrospect, when considering Ford cars, and sad in prospect when considering Transmeta processors but...

    Just look around when driving. How many different colors do you see on Ford cars? How many different models does Ford make today?

    Give Transmeta some time.

    The pace of microprocessor evolution is much faster than that of cars. Give them a year or three, and we'll see not only more chip models, but also more supported emulations.
  • With all the news about Rambus' patents and dealings with Intel, along with AMD's problems getting manufacturers to make enough motherboards for the Athlon and Duron, it's good to see that Transmeta is bringing more competition to an industry that has to deal with too many predatory monopolies.
  • I can't seem to find any pricing information. Does anyone know projected prices for these notebooks? Also, forgive my ignorance, but what does Q4 translate into for normal person time?
  • Okay, some people pointed out the obvious down-sides of a low-power SMP box. Point taken. I already knew most of the downsides. I was just dreaming.

    What I'd like is SMP to come to a more sensible level. Yes, there's not much point in SMP on the average to low end right now, but it's a self perpetuating cycle. The common software doesn't do SMP well and therefore SMP is rare, therefore the common software developer doesn't worry about SMP, etc.

    I'd like SMP on the consumer end so that starting Netscape doesn't make my MP3 skip. I'd like SMP on the consumer end so that slocated doesn't kill performance while I'm actively doing something else. Can't a properly threaded OS use one processor for the high-priority process and use the others for background?

    I'd also like my SMP box to have consumer-level power requirements. A 500Mhz PIII is fine for 99.999% of consumer requirements. Therefore, 4 CPUs as powerful as a PIII 500 but having very low power requirements would be good.

    As someone pointed out, SMP would probably add a lot of complexity and power requirements of its own though.
  • The article mentioned IBM, HP, etc. claiming 8+ hour battery life on notebooks. If they can indeed bring such long life and keep the weight down to the ultra-light class, I predict a surge in laptop sales.

    IMHO, Battery life is one of two killer apps that keeps the Palm platform at virtually a monopoly level over other hand helds (ease of use is the other, even non-geeks like to use Palms). The only thing that could stop Transmeta from eating Intel's mobile cake is if some manufacturer could produce an equivalently powered notebook that doubles the battery life over what Transmeta offers. Transmeta has some serious mindshare right about now.

    But as for me and my house, we will remain content with $200 486 based laptops that we pick up on ebay.

  • Actually the "desktop" NetWinders are not intended to be workstations. None of their products are. They are intended to be small to midsized office servers (that's where the "OfficeServer" name comes from). That's also the reason it comes with Apache & Squid & ht/Dig and all the other junk it has preconfigured.

    From that standpoint the NetWinders are very nicely priced.

    ---

  • Remember, you have to look at this from the corporate world. Yes, you and I and 99% of the readers of /. can make super, kick-ass servers with just about any bit of HW lying around. However, if you were a small to min-sized business with no technical people (I've been doing business with some companies where the system/network administrator was just a guy who knew slightly more than where the on/off switch was) and the cost of the NetWinder is paltry for the capibilities it gives.

    Remember, whenever making a statement about technology and the "real world" you should make sure to look at the real worlds point-of-view first.

    ---

  • If you like small and cheap, you might like the EspressoPC [saintsong.com.tw] from Saint Song. Ars Technica did a review [arstechnica.com] on it, and they seemed to really like it. Just think, a middle of the road machine twice the size of a Game Boy. Even better than a hacked I-Opener.
  • Just because Sun hasn't released a drastically new chipset sence 1995, or was it 96, when the Ultra-1's came out doesn't mean that they should be attacked. They have been working on the Ultra 3 architecture for quite some time now and should be release this summer. Theoretically it should be fairly different, at least from what I have been told.
  • Although LCD technology is not exactly new, the introduction of LCD screens into the desktop market in mass quantities is. This introduction, along with the relatively cheap material, will drop the price of LCD monitors rather quickly as companies start to ramp up their production of them.

    The price of materials for an LCD screen will soon become cheaper than that of your regular CRT tube. A good analogy is the difference between vaccume tubes and transistors. At first it was still cheaper to use tubes until transistors had developed a little more, and then the transistors took off (Side note, we can thank the Apollo program for pushing for miniturization of computers, their need for as little weight as possible made SI possible much sooner than through regular inovation).
    Liquid crystal technology is still developing, not only by the companies that make the displays but also by research teams at universities. I know that the physics department at the university that I am attending is doing reseach on LCD's and making them even cheaper.

    Demand increasing
    +Production increasing
    +Competition existing
    +Inovation
    =Lower costs.

    In my opinion LCD prices will continue to drop until they are below the price of CRT tube monitors. Any major breakthroughs will only serve to increase the speed at which it drops.
  • The current Transmeta chip is not supposed to be the most powerful thing out there, and definately not to be a primary server. It is there to be used in devices that don't need a full powered Intel, AMD ... the point is that it has lower power consumption, and produces low heat. It is perfect for a laptop, palm device, and embedded devices, and that is the market that they are after. In my opinion it would be a waste of time, and effert, for Transmeta to implement SMP support into the Crusoe. SMP support takes up space on the chip, space they could use for something more practicle for the market they are aiming at.

    Also, just because you have more than one processor doesn't mean that it is the greatest thing out there. There are a lot of things one needs to consider before dealing with SMP machines, like is it necessary, or will the software that you are using effectively use the SMP capabilities of the machine? I don't need SMP to do what I do at home, in fact on fast expensive chip is better for what I do at home. Frankly, in the lab that I work, the code isn't optimized to take advantage of SMP, and is actually faster to run my multiple symulations on seperate machines than load them up onto one SMP machine.

    The one use I could see for having SMP capabilities for the Transmeta chip is for a backup chip for a server attached to a UPS. The power goes out, you don't want the server to die, no one can connect to the server because their machines are dead, switch to the Crusoe that is emulating the same chip as your server and let it keep the server alive until the power comes back up.
  • It's true that Apple doesn't have a 'BIOS', but they do have undocumented chips which you would need to engineer around.

    Good point. I was thinking from the point of view of porting alternative operating systems to run on Apple's hardware, while the thread was actually about building alternative hardware to run Apple's OS.

    Yes, since Apple abandoned CHRP/PREP, their motherboards use custom PCI controllers, power management units, and other proprietary chips. This effectively prevents other companies from building their own boxes that MacOS will support.

    On the other hand, Darwin [apple.com] (the kernel for OS X Server and the upcoming MacOS X) is open source, which should allow alternative hardware manufacturers to port the OS to new hardware themselves -- assuming anyone has the motivation to do so.

    On top of that, you would need to include a retail copy of MacOS with every machine, even if Apple would sell you a copy and the licence would allow you to install it.

    This, on the other hand, is something that only Apple (read: Steve Jobs) could fix.

  • Thanks for the run down...from someone who wishes he was there for the demo, and Linus- thanks.

    -Sleen
  • One thing everyone forgets about emulating Macs, IE emulating PowerPC's, is that Apple still controls the Mac BIOS. That's why you don't see any Mac clones. (well Apple did experiment with licencing the BIOS but decided against it in the end) The Crusoe isn't going to be any different.

  • You're right, there often is. However, this has not stopped DRAM manufacturers such as Samsung in the past from selling product below what it costs them to make it just to keep marketshare.
  • This is essential to the industry. By having as many non-server architectures as possible means more competition, lower prices, and better value for consumers.
  • I'd rather see an overall price drop in laptops. The bulk of their cost comes from the LCD screen. Someone needs to market a cheap alternative. I can get a whole computer for $500 now, and laptops are still 3 or 4 times that.
  • What I would like to know is, because the Crusoe is in two parts: the VLIW cpu, and the Code Morphing engine, why couldn't code bypass the morphing engine, and execute strait on the VLIW? The code morphing engine is there to essentially emulate an x86, and can get as fast as a 700mhz, meaning that the VLIW must be significantly faster with out having to deal with the emulation. Why couldn't this overhead be simply bypassed to create faster applications that are designed for it, while still allowing x86 code to be run properly?
  • Not to mention that their stooge Rambus is trying to milk us for every dram chip manufactured using their so-called intellectual property.

  • I don't suppose you thought to ask Linus, but I would be very interested to know how well these machines perform running Linux as compared to say, N/T. Was anything mentioned along these lines ?

  • I honestly think there's a glass floor if you will on prices... supply/demand, labor costs, these are things that take revolutions and generations to put dents in... I think computers and networks are as complicated as cars... while they are better now-a-days, I still can't go out and get a sufficient vehicle for $100 brand new...

    ----

  • Is a high power battery. Like the new titanium e2. [usatoday.com]

    I like batteries... a lot.

  • It's true that Apple doesn't have a 'BIOS', but they do have undocumented chips which you would need to engineer around. There's a reason they need to patch the MacOS everytime a new machine ships.

    On top of that, you would need to include a retail copy of MacOS with every machine, even if Apple would sell you a copy and the licence would allow you to install it.

    If a Crusoe-powered Mac laptop is true, there's only one rational scenario -- Apple is making it. (However, with the G3 running at 500+Mhz one wonders if it is worth the engineering effort.)
  • I kind of envy people who think different, they get to ride the short bus and go to special schools.

    Where as people who don't read an entire post before making a knee-jerk reaction are towers of intellect.

    Perhaps you should go here [m-w.com] and look up "sarcasm", if you can manage to read the whole entry before wondering why I sent you to a pronounciation guide.

  • Rebel has been looking for faster chips for about a year now. At first they were going to go for the next incarnation of the StrongARM, but since Intel bought it, development slowed (they didn't want to compete with the Pentium). So, now that Crusoe is out, it's the next best choice. The 275 MHz StrongARM is just a little slow nowadays.
  • I have been watching the Megahertz and 3D graphics cards race of the market with disgust

    but you have to admit that a 1GHz CPU that pumps out enough heat to warm an olympic sized swimming pool, and a graphics cards requiring a souped up AGP bus to run, packing a cooling fan the size of a car radiator are plain straight-out fun

    and that's what computers should be about.
  • The Merced should get them going again
  • With the demand for portable and smaller devices a low power alternative to the standard Intel Pentium processor will excell. It makes perfect sense for notebook vendors and anyone trying to squeeze weight and size and extend battery life that the less power you need and the less cooling required the lighter and longer running your product will be.

    I am anxious to see some good performance benchmarks. I want some good real world software tests of the Crusoe in comparison with the AMD and Intel processors.
  • . However, I am curious as to how differently Crusoe would have been accepted had Torvalds not helped in its development.

    Probably with something like: "Trasmet-what?" ;-) Kidding.

    People seriously interested the industry have been watching this story since Transmeta started, because back then only a few people were realizing the problems with power. (And back then Crusoe wasn't targeting mobile, wasn't it going to be a PC killer?)

    I think Crusoe would have been accepted in probably much the same way if Linus wasn't part of it. It's the low-power operation makes is the meat of the issue, something that would have gotten lots of press anyway. I still don't see it's x86 compatibility as big a benefit as they expect.

    The real bloodbath will start when you can run equivalently priced Transmeta- and Intel-based machines side-by-side. You think Linux vs. Windows is a jihad? Ain't nothin by comparison. It's gonna be ugly.


    ---
  • The underlying instruction set changes too often.

    The idea is to present one set of instructions to the Morphing engine, then have the freedom to change whatever they need to on the underlying chip and it's instruction set(s).

    That seems like more of a bonus to me than being able to stare at a MacOSvapor bootup screen ;)

    Linus has stated (I believe around the time he got his butt kicked in Quake at the initial demo) that they don't even "want" people to port the OS to the underlying chip.
  • Wasn't the whole point of portable unix and C and crap that makes up the foundation of userland is that you don't have to give a shit about the underlying processor? Everything works the same. Just recompile. Well hurrah. We've succeeded! It's a tremendous accomplishment! I want to see dancing in the streets for godsakes.
  • From zdnet's [zdnet.com] link, "Crusoe-powered ThinkPad due by end of year While the notebooks are being described as prototypes, IBM (NYSE: IBM) has said it is committed to introducing a ThinkPad notebook powered by a Crusoe processor by the end of this year. ".

    My question is where are the Transmeta chips produced? If it's somewhere in South East Asia, it would be possible to test drive these babies by the end of the year. If not, I'll just have to be content reading /. reports from you fellow readers!

  • imagine onk posting beowulf cluster comments on his commodore 64 and 300 baud modem
  • Just got back from PC expo and I'd like to start off by saying that LINUS was THERE. Not only was he there but apparently like me and a whole of six other people knew who he was paid him much attention. He wasnt up on a pedestal giving some keynote or anything, he was just out on the floor at the transmeta booth giving interviews to reporters and talking to the few of us that know his face. I ran down to the linux pavillion and bought a tux polo for him to sign and though he didnt seem that happy about signing a cotton shirt with a ballpoint pen he nonetheless obliged. Anyway onto the products. . . They had a few (no more than ten) notebooks at the pavillion and a working webpad that really knocked my socks off. The notebooks were VIAO-like in form factor and even at the end of teh day were very cool to the touch. They were running NT (I THINK - dont remember very well) and were each a different cutesy color. WHile the engineer said named the fore-mentioned big four producing the books this year, I saw nothing but Hitcahi prototypes on the floor. They were running some type of movie on power consumption of crusoe vs. a speed-stepped PIII while decoding DVD. Judge hype-worthiness for yourself, but the running graphs showed that as soon as the DVD decoding starts, the Crusoe adjusts accordingly for power down to an average of circa 1 watt while the PIII kept eating power at the same rate it started at. I was told the tiny translucent pink laptop I was using could run for a full seven hours of normal use but then again they were all plugged in. The webpad on the other hand was NOT a hands on demo but was a real knock out. We were all close enough to know that it was real but not close enough to know what it was running. Anyway - my two cents - they have a good product -looking at the chip prices quoted at the original press conference months ago these things should take a significant part of the market away from chipzilla and as an economics student Im happy to see another player successfully enter the market.
  • I'd like SMP on the consumer end so that starting Netscape doesn't make my MP3 skip.

    Unless you have a years-old processor, you shouldn't need a second CPU to keep MP3s from skipping. Your problem is more likely misconfiguration of the system you have.

    Are you using DMA (or UDMA if you've got it, but switching from PIO to DMA is the big CPU saver) on your hard disks? If not, then every heavy disk access (like starting up a big program) is begging your CPU to drop whatever else it's doing (like playing MP3s) and handle the flood of interrupts from the drive.

    If you're using Linux, try putting "hdparm -d 1 /dev/hda" in your rc.local script. (or "man hdparm" to find more tweaks, turning on DMA is just the most important).

    If you're using Windows (even Win98), there are lots of motherboards with UDMA drivers that don't get installed by default; check the manufacturer's website. And even if the drivers are installed they may not be turned on; there's a checkbox somewhere in one of those Control Panel doohickies.

    If that is insufficient by itself, try making your MP3 player a niced or realtime process. If you make xmms suid root, for example, there's an option to make it a realtime process. (why it doesn't give you a separate option to set a nice level, I have no idea). This can be a stability/security risk, so don't do it on the company web server.
  • This is all well and good, but what I want to know is - can I afford one? Has anyone seen mention of prices for a computing device using a Transmeta processor?

    ---
  • The real questions are whether
    1) Crusoe with linux can use x86 linux binaries


    That was the real question months ago. Where you been, under a rock?

    --
  • > Methinks you have forgotten that
    > Transmeta has a non-existent PR
    > department.

    They have a mailing list for product announcements [transmeta.com]. They should use it, that's all I ask for.

    ------------------
  • I have to correct myself. I haven't tested it personally, but according to my trusty copy of the IEEE 802.11 Handbook, FHSS devices should be interoperable with 11Mbps DSSS devices. Given that, $78 is a pretty good deal for a PCMCIA wireless card. I would still recommend an Airport for an access point, however.
  • This will start happening pretty fast. With Apple AirPort's going for $280.00 at PCConnection which can be configured under Linux with a Java management client, the other folks won't have a choice but to drop prices.
  • Um, exactly how an AirPort, which is 802.11b compatable, can run at 11Mbps, includes a NAT/DHCP server, a modem/PPP client, can be configured from Linux, and priced at $280 is somehow "overpriced crapola" compared to an access point which runs 802.11 at 2Mbps for $360. But whatever.

    Their PCMCIA cards are attractvely priced at $78, however. Too bad most of the 802.11 manufactures have moved to DSSS instead of FHSS so that they can run at 11Mbps, so the Webgear stuff won't work with most new equipment.
  • It's the slowest high end chip. It's the only performance processor that posts lower specfp2000 numbers than the x86.
    The 450 mhz us II is less than half the speed of the 667 mhz ev67 alpha.
    Sun needs the us III core more than intel needs the p7.

    --Shoeboy
    (former microserf)
  • Link for you:
    Apples power failure. [realworldtech.com]

    I believe that the new intel mobile chips with speedstep actually consume fewer watts than the G4.

    What's definite is that x86 slaps the g4 all over the place on both bandwidth and integer performance. And the ghz t-bird beats the 500mhz g4 by a very wide margin in floating point.

    As far as your assertion that x86 instructions require a lot of transistors, that's only if they're implemented in hardware. The crusoe uses software translation. (not the same as software emulation)

    --Shoeboy
    (former microserf)
  • by Pope ( 17780 )
    Q4 = Tomorrow.

    Repeat when necessary :)

    Pope

    Freedom is Slavery! Ignorance is Strength! Monopolies offer Choice!
  • Look at WebGear [webgrear.com] for cheap wireless networking stuff. I have some of their stuff and it works great. Even has Linux drivers.

    The fact it has nothing to do with Apple's overpriced crapola is an added bonus.
  • Are the Netwinders hideously overpriced? I looked around and couldn't find anything for under $1000 US. Is it just that form factor? Am I missing something?

  • I wouldn't be so sure. Some things are always going to be expensive, for various reasons. I know we're jumping industries here, but a car catalyst is never going to be cheap beacuse it's made of platinum or rhodium. Now, I don't know enough about the details of LCDs to know whether this is definitely the case, but with digital watches and mobile phones te small ones have been here in huge volume for some time. Laptops and LCD monitors aren't a small market either, though they're smaller in relative terms.

    Anyway. I'd say, looking from the outside, that LCD prices probably haven't got that much further to fall without a major breakthrough.

    LEPs are looking very interesting and may well produce some nice results but I can't see LCDs usable in decent webpads becoming that cheap.
  • Erm...

    The nature of a catalystic converter - yes, I was just being lazy - means it actually has to be made out of the materials they use. Cheap alloys aren't a possibility.

    I can't see electric cars coming all that soon, though. Battery technology is still a _major_ problem and they've been working hard with no real results for ages. I'd have to say that hydrogen power, in one form or another, looks by far and away the most likely.

    If we want to look at it from another side, electric cars wouldn't actually help the environment as a whole - only the local environment. The electricity has to come from somewhere and, right now, power stations are _worse_.
  • Yes, they are hideously overpriced.

    I love the formfactor, but they clearly want to go for the rack mount ISP market, not the market for people who want small, adequate workstations. I really want a small inexpensive workstation that can also in a pinch be put into enclosures or do light duty workgroup server duty.

    HP is making a small PC called (ugh) "E-PC" for about $549 list for Celeron 500 64MB RAM 8.4 Ultra ATA, 100Base T etc and standard connectors; the case is 8.9x24x27.2 cm. It comes preloaded with (ugh) Windows, but also has a quick swap disk feature so you could bulk reprogram the disks for Linux. I'm trying to get my boss to get a couple of these.

    These aren't quite as small as the netwinders, I think, and probably not quite as quite. It is a lot more flexible and fast though.

  • It's still too expensive considering what you could do with x86 Linux.
  • When you add in the cost of electricity for an always on computer, the Netwinder pays for itself in about two years compared to a Intel-based PC. It uses only about 7 watts average and 15 watts max.


    True. I'd really like reduced power requirements. However, computers often get replaced eery couple of years, so financially its a wash.

  • I have no idea if it's possible with Crusoe, but I would love an SMP box based on Transmeta chips.

    Think about it. TM chips are tiny, cheap, and low-power. You could cram a bunch of them into a case and motherboard that are within reach of common people. So a Crusoe 700Mhz = PIII 500 Mhz. I guarantee 8 or 16 Crusoe 700s will beat any two PIII 8xxs (1GHz PIII is not certified for SMP).

    Of course, for various reasons, the consumer market is anti-smp. Intel would rather sell you one super-expensive chip than two cheap chips. OEMs would rather you buy a whole new system than simply add another processor. etc.
  • From now on, everyone needs to enter many fake or flawed submissions. It could be a completely made up story, it could have fake URLs, or it could be a real story with incorrect facts and misleading titles.

    The general idea is that after they realize that so many submissions are flawed, the editors will be more inclined to check the sources before posting.


    You wouldn't perchance be a fan of Scotsman's tactics [xwrestling.com] to ensure fact-checking among Internet reporters on a different subject, would you?

    His two shots at this experiment got him a lot of vitriol from Internet wrestling "reporters"...and also a lot of accolades from people sick of the pseudo-news that infest a lot of the newsboards that have little better to do than post rumours and rip off Dave Meltzer, Wade Keller and Dave Scherer.

    Despite all the potential for abuse and harassment, it seems such an operation did force people to check their facts more carefully, lest they lose credibility. It's something of a nasty guerilla tactic that can easily go haywire and result in...errr...unintended consequences.

    Then again, I'm sure Taco and the crew deal with at least 100 bullshit submissions a day. Like they need any more:)
  • Ok, this is actually a really interesting point.

    Why are the G4s giving motorola such a hard time getting them to run fast? Complicated instruction dependencies? Not deep enough a pipeline?

    If the Transmeta people are able to claw back some performance by having a simpler core run fast and then dynamically compiling it, that would be cool.

    On a related note, since the TM is a VLIW proc, it should (?) have a decent chance of dealing ok with the AltiVec instructions, no?

    Johan
  • Sorry, Macs don't have anything called "BIOS." On most PowerPC and related architectures (including PowerMacs, IBM workstations and servers, and others), Open Firmware [sun.com] performs most of the functions of the PC BIOS. OF is an open standard that is not controlled by Apple. Of course, Linux and NetBSD developers have had to deal with Apple's various poor OF implementations, but that's a different story...

    What you're thinking of is the Mac ROM, which used to be a megabyte or two of low-level code stored in a ROM chip on the mainboard. These were copyrighted by Apple, and were necessary for MacOS to run. This meant that PowerMac cloners had to license the ROM from Apple, and that MacOS emulators and virtual machines (Mac-on-linux [maconlinux.net], SheepShaver [sheepshaver.com] et al) needed a ROM image in order to run.

    Since MacOS 8.6 however, the "ROM" actually resides in a system file loaded into memory straight from disk. New Macs being manufactured don't even have a traditional style "MacOS ROM" chip. Programs like MOL boot MacOS just fine using nothing but the ROM file from the install CD.

  • Windows NT3.51 came in a version for x86, MIPS, sparc, Alpha and PowerPC. NT4.0 still comes in a version for the Alpha architecture, but is no longer supported. Microsoft ported their most popular apps, and gave complete compiler suites away for the Alpha architecture trying to get companies to support the Alpha-NT platform

    That was a ploy. They had DEC write their own x86 software emulator for Alpha NT (for a long time DEC had more programmers working on NT than Microsoft), and Alpha NT still sucked rocks. Microsoft shelled it for the 2000 series, although we have a beta version running here. One of Microsoft's favorite cards to play was threatening Intel with boosting alpha Windows development. You may recall it was this threat that caused Intel to halt its development of multimedia tools for its newer (at the time) MMX chips. Intel was trying to use software to boost its share over AMD/Cyrix, but Microsoft viewed that as a low level threat.

    Alpha NT never had the most important software available. Like Office, for example. Its strongest value to Microsoft was the leverage it used against Intel. You could run such things in x86 emulators, but if you think that is viable I think you never tried it.

    no one can simply force the PC architecture to switch to a different processor.

    Microsoft can and has kept processor development back. How can a new processor be successful when the OS on 95% of the world's computers only talks x86 ?? Microsoft NEVER made a serious attempt to make its product cross platform. They had a monopoly, and they knew they could force the chipmakers to bend to their will.
  • Umm, so what if a car's catalytic converter (I guess that's what you mean when you say "car catalyst") is currently made of expensive materials like platinum or rhodium? You're assuming no one will innovate to bring prices down. If a company can find a way to make a catalytic converter using cheap alloys, that company could wipe up the competition in sales--so there's pressure to innovate. Companies always invest a lot in R&D, and often invent exotic but cheap alloys that can do the same thing as more expensive materials. TECH DOESN'T SIT STILL, it moves forward. People will eventually invent cheaper ways to do any given task, to create any given object. That includes catalytic converters and LCD screens. Not that we'll need catalytic converters in ten or fifteen years, since the electric car is becoming more of a reality and even more of a necessity, what with OPEC's price-fixing bullshit.
  • What on earth are you talking about? How can you possibly say "never" when talking about price drops on computing hardware? Just like everything else in this industry that starts out expensive, manufacturers will find cheaper ways to make LCDs, and soon enough you will see them on the backs of cereal boxes.
  • If someone from Transmeta's PR department is reading this: Hello? /please/ get your act together! Methinks you have forgotten that Transmeta has a non-existent PR department. Their way of hyping the release date was to hide that message in the html source, remember?

    All the more reason why they are cool!

  • Now we just knock down the cost of 802.11 a notch, and those pads start to look pretty darn nice :-)
  • They still won't get past the higher LCD display costs [cnet.com], and memory costs... Webpads and most electronic networkable devices will never, IMHO, be under $100.

    About the only thing I see promising is up-coming disposable cell phones. Techlust seems to come at some cost, and its more than the ultimate feast at Red Lobster.

    ----

  • by Anonymous Coward on Tuesday June 27, 2000 @11:01AM (#972815)
    I have a solution to the problem that's been plaguing slashdot for the past several months. We can make it so the slashdot editors check the facts on every story before they post it.

    It's really quite simple.

    From now on, everyone needs to enter many fake or flawed submissions. It could be a completely made up story, it could have fake URLs, or it could be a real story with incorrect facts and misleading titles.

    The general idea is that after they realize that so many submissions are flawed, the editors will be more inclined to check the sources before posting.
  • by Hanno ( 11981 ) on Tuesday June 27, 2000 @12:24PM (#972816) Homepage
    I registered with Transmeta's product mailing list some time ago, yet I still haven't received a single message from them about upcoming products. What's going on there?

    If someone from Transmeta's PR department is reading this: Hello? /please/ get your act together!

    That said, I look forward to a Transmeta-powered product, I really do. Not having a car and spending most of my travelling on bicycles, in buses, on trains, I always bought ultra-light computers for my work.

    Raw CPU speed never was a *big* issue for me while on the road, but portability and more than that battery time (!) are (I wrote the Battery Powered Linux Mini-Howto sometime in 1997, you can still find it in your /usr/local/howto :-) but it has now been superseded by Werner Heuser's excellent Laptop Howto).

    I also want a Transmeta-powered desktop mini-workstation, running silent without a fan, with minimal footprint on the desk and little power consumption.

    I want a mini-server that serves as a home network file server, ISDN router, fax, answering machine and MP3 player that I can leave running all day and that doesn't consume a few dozen watts while idle.

    These products are literally what I have been waiting for since a few years. I have been watching the Megahertz and 3D graphics cards race of the market with disgust and hoped for a sign of sanity in all this useless power- and feature-sucking.

    Transmeta, here's a possible client. Please don't disappoint me. :-)

    ------------------
  • by Shoeboy ( 16224 ) on Tuesday June 27, 2000 @11:22AM (#972817) Homepage
    It's amazing to think that they have spent so many years getting away with regurgitating the pentium pro core, without any solid competition.
    And just what, pray tell, is wrong with the p6 core? It's gone from 150 to 1000 mhz. (if you can get one.) It stomped every competitor in integer performance when it came out. It still powers past everything but the athlon and alpha in integer speed. It's proven to be flexible and scalable. Is it a problem that they've been improving a good core rather than making a new one?
    If you want to bash a company for regurgitating an obsolete core, pick Sun.

    --Shoeboy


    (former microserf)
  • by iKev ( 73931 ) on Tuesday June 27, 2000 @12:39PM (#972818)
    ZDNet's take on it [zdnet.com]
    Says that NEC's Crusoe laptop can run for 8 hours running MS office apps..
  • Worse than that, the Duron [slashdot.org] is now beginning to eat up the marketshare of their once-unique Celeron.

    Intel basically was untouched in four areas: High-end server, desktop, low-end desktop, and laptop. No, for our purposes, Cyrix and the K6-2 do not count as competition :-). Now, they compete with the Athlon in the desktop, the Duron in the low-end desktop, and Crusoe in the laptop. And their only remaining area of exclusitivity is only temporary wile AMD gets its act together with the dual Athlon. And this area was already one of their weaker spots, with competition from Solaris and Alpha in the very high-end we-don't-care-how-much-it-costs department.
  • by _xeno_ ( 155264 ) on Tuesday June 27, 2000 @11:40AM (#972820) Homepage Journal
    There is a cheaper alternative to the LCD screens being developed right now. It's called LEP or Light Emitting Plastic. A British company called Cambridge Display Technology [cdtltd.co.uk] recently came up with a method of creating LEP displays through what should be a rather cheep and high-yield process.

    Their press release [cdtltd.co.uk] gives some details about how they're creating displays for use (currently) on small handheld devices, but suggests that they could easily be scaled up for use in larger devices.

    "The pre-production colour light emitting polymer display being shown by CDT and Seiko-Epson has a colour density similar to current LCDs. The techniques being jointly developed by the two companies means that the manufacturing cost of an LEP display will be significantly less than the cost of producing conventional LCD or cathode ray tube displays," said Dr. Shimoda, general manager of basic research, Seiko-Epson.

    Whether or not this technology will be used in the future to really lower the price of laptops is to be seen in the future, but it seems like something cool.

  • by Golias ( 176380 ) on Tuesday June 27, 2000 @11:18AM (#972821)
    It is starting to bug me that none of the Transmeta-related press releases I have seen so far seem to give many clues about the price of these chips. Sure they are low-power, low-heat chips filled with all kinds of crunchy goodness, but how do they stack up against, for example, the fondly remembered Strong-ARM in terms of "!" for "$"?

    For that matter, how to they compete in price against the AIM-cabal's G3, which runs for hours off the static electricity of your body and actually cools the air around its own heatsinks, but costs more than a new family car... or the good old "de"Celeron, which doubles your power bill and is so hot that you can cook an egg on the next desk over, but can be bought in the Wal-mart bargain bin for $3 a pop and overclocked to 2500 MHz?

    (Disclaimer: The above might contain a few slight exaggerations of the strengths and weaknesses of various CPU's.)

  • by Hanno ( 11981 ) on Tuesday June 27, 2000 @12:42PM (#972822) Homepage
    > Do you feel that the Linux community has
    > fallen victim to Transmeta's brilliant
    > public relations? Since Linus Torvalds
    > has joined Transmeta, everyone has
    > become aware of Transmeta's entire
    > product line...

    Yes and no.

    I was just as excited about the PowerPC chips when they were new on the market and hoped for affordable PowerPC hardware that could replace the vanilla Wintel box on my desktop. And the PowerPC chip wasn't tied to "one famous person", at least not for me.

    I remember discussions years ago with my co-students during computer science classes about how great this chip would be and how it would be a *true* alternative to the Intel product line.

    However, Apple killed off the clones and PRPC (or whatever the non-Apple Power PC hardware setup was called) was stomped before it really took off.

    So here I sit now, waiting for the next contestant. The iMac is interesting, but then again, Transmeta looks like the better idea to me now. See my other post in this thread about that...

    ------------------
  • by Frac ( 27516 ) on Tuesday June 27, 2000 @11:07AM (#972823)
    I do. Athlons are eating them up in the high performance desktop market (they better hope that the dual-Athlon chipset from AMD takes forever), and while Transmeta is nipping at their heels at the notebook market.

    Which is good of course. Maybe Intel might stop being so sluggish all these years. It's amazing to think that they have spent so many years getting away with regurgitating the pentium pro core, without any solid competition.

    Go get your free Palm V (25 referrals needed only!)

  • This is a kind twisty story. Buuuut.

    Back in the day, Windows and Mac were real competitors. When Mac introduced a new line, with a new processor, they would allow new instruction sets. The old instruction sets would be emulated in software. They figured the new processor was faster, and that eventually everything would get recoded anyway. Worked pretty well for them.

    Well, not so in Redmond. Microsoft would not recode X86 instruction sets for other processors. If a chip manufacturer wanted to capture the Windows market, they had to do X86. So, we see efforts like those at AMD in the Athlon to do hardware translation, and at Transmeta to do software emulation.

    Meanwhile, Mac already has a G4 that is comparable to the Crusoe for power consumption, and per CPU cycle as fast as X86 processors get. The bottom line is that it is really dumb to stick to old inefficient instruction sets just because your monopolistic OS company refuses to do a little work to earn its tens of billions.

    I hope Crusoe is REALLY successful. AMD and Intel will NEVER create a chip as low on power as a G4 and as fast. Or even a Crusoe. X86 instruction sets simply require a LOT of transistors to work.
    The real questions are whether
    1) Crusoe with linux can use x86 linux binaries
    2) Crusoe linux laptops are faster than G4 linux laptops (once those actually run)

  • Hopefully, the Crusoe Netwinder will be as fully-functional as the current Netwinder family, and not "just" the "gateway" product mentioned in the article. After all, it could be used as more than just a gateway/firewall/router type of box; I'd love to use one as a small file server networked to my current AMD box. All I need is for it to have a PCI slot capable of using a Promise IDE RAID card; it would free up a PCI slot and a couple drive bays from my current computer, and I'd be very happy with it--serving files to a local box isn't so processor-intensive that the Crusoe would be overwhelmed.

    I've been hoping that the Crusoe would make it into some desktop products like the Netwinder, since notebook products are inevitably too bloody expensive (and fairly useless to me, anyway--drive bays! I need more drive bays! not a hobbled travelling PC). After all, not everyone wants/needs the screaming-fast performance of an 800MHz Athlon; some of us want a Crusoe desktop machine for both the "wow factor" and to support the company, but given the nature of the new processor and its chipset DIY commodity mobos and retail processors aren't realistic for any time soon.

    Personally, my old 400MHz K6-2 is still fast enough for everything I do on a daily basis--the occasional video clip rendering notwithstanding (just leave the box alone for a few hours, and...), and I bought the thing over a year ago to support AMD even though the Inetl Celeron 366 was faster for the same price. I can't in good conscience buy five-year-old tech from Intel, I have to support competition in the market--and now with Thunderbird and Duron it looks like AMD is finally tromping Intel on all fronts, hooray for supporting the underdog. Look at all the good we who have supported AMD through the dark ages have now brought about: AMD chips that whomp Intel, but just as significant, lower Intel prices and higher Intel clockspeeds. How can self-respecting geeks buy Intel, knowing that Intel's own roadmap had us all still using processors in the 400-600MHz range right now, with AMD's Athlon being the only thing which picked up the pace? And now, because of AMD, Willamette is going to be released not just for servers, but as a replacement for the ancient-cored P!!! with the P!!! Coppermine becoming their new low-end processor to compete against Duron. Intel had been planning to shove that integrated bullshit Timna (think Cyrix MediaGX) down our throats as a Celeron replacement, with the P!!! for higher-end consumers, and the Willamette for high-end workstations and servers only. That's what their roadmaps said before AMD put the Athlon squeeze on. Folks, unless you need a dual- or quad-processor box *right now*, don't support Intel--they don't deserve it. AMD is now the real innovator, even though they're still the little guy. And Transmeta, support Transmeta as well--they're innovating in different and commendable ways. But there's no excuse for supporting Intel, the company which was going to move us at a snail's pace until AMD stepped it up.

    Anyway, I was planning to turn my old 400MHz box into a fileserver at the end of the year/early next year when I can get my hands on a dual Athlon motherboard, but I'd gladly buy a Crusoe desktop to do that fileserving. Thank you, Rebel.com, if you follow through and make it good.

    So, am I the only one who wants a Crusoe desktop, or are there other technogeeks out there who'll buy a Crusoe desktop system aither in its Netwinder incarnation or otherwise?
  • by panaceaa ( 205396 ) on Tuesday June 27, 2000 @12:09PM (#972826) Homepage Journal
    Do you feel that the Linux community has fallen victim to Transmeta's brilliant public relations? Since Linus Torvalds has joined Transmeta, everyone has become aware of Transmeta's entire product line... and not only that, most of you are crazy about it! Don't get me wrong; Transmeta really does have a great product with Crusoe. However, I am curious as to how differently Crusoe would have been accepted had Torvalds not helped in its development.
  • Kinda sad how the upshot of all the nifty news technology is the rather boring (though admittedly valuable) goal of low power consumption. The original press conference read sort of like the Cheese Shop Sketch.

    "The crusoe can emulate any chip at all."
    "Like a PowerPC?"
    "Theoretically."
    "Or an Alpha?"
    "Technically."
    "Or a Dragonball?"
    "Probably."
    "So what can it actually emulate?"
    "Any chip at all, so long as it's an x86."
  • by Hogarth ( 98887 ) on Tuesday June 27, 2000 @12:17PM (#972828)
    The real appeal of Crusoe are these:
    1. Low Power Consumption
    2. Low Heat Generation
    3. Non-Intel Manufacture
    4. Software-Alterable Emulation
    5. Multiple OS Support

    (1) The fact that Crusoe doesn't require a whole lot of power doesn't mean that it will automatically be thrown into notebooks with standard Li-Ion batteries and get 1000 hours of battery life. The idea is that if you're more efficient, you need less physical battery. That means smaller system. That means ultralight, which is everything with notebooks these days. I know plenty of folks who would give up a very fast and loaded desktop-equivalent notebook for a VAIO that is light, decent, and looks good.

    (2) Less Heat eliminates need for a processor cooling fan. eliminating the need for a fan makes the package smaller. see also #1.

    (3) Companies, especially Big Blue, are tired of forking over notebook profits to Intel. period. Crusoe is very affordable, and that will bring prices for notebooks down all over the place. Intel will still dominate, but at least you won't have to pay a premium price just because you want a PIII-600 in your notebook.

    (4) Being able to alter the emulation processes at the software level means that this is a chip that will grow and improve with time. Intel's coming out with more SSE multimedia extensions? Patch crusoe. Boom. Upgraded Crusoe. No hardware swap required.

    (5) Yes, being able to theoretically emulate anything is pretty lame. Alpha and Dragonball, while potential, are not likely to get emulated with this. What will? PowerPC. I guarantee that Transmeta has a team trying to get down the instruction set for the G4. I know many a web designer who would run out and buy a system if they could run both Windoze and MacOS on the same hardware without something like Virtual PC. It's not as appealing to the market, but it may eventually be part of Transmeta's value proposition. I'd expect to see it within the next 12 months, and expect to see companies like Dell and NEC making notebooks that have full G4 support and Apple jumping on the bandwagon.

    Bottom Line: Crusoe is revolutionary. It will take a while for the waves to be made, but we'll all be using faster, cheaper, lighter, cooler, better computers as a result. Intel needed some additional competition.

  • by cow_licker ( 172474 ) on Tuesday June 27, 2000 @10:59AM (#972829)
    Here is the official press release. http://www.rebel.com/corporate/press/20000627.html

"The chain which can be yanked is not the eternal chain." -- G. Fitch

Working...