Posted
by
Hemos
from the they-will-unveil-it dept.
Magorak writes "ZDNet is running an article on the upcoming PC Expo show in New York which is going to feature new notebooks from NEC and IBM which will have Transmeta's TM5400 processor in it. There's mentions of Transmeta's background, future company partnerships, and other goodies."
This discussion has been archived.
No new comments can be posted.
That would be a very interesting project. Too bad that Transmeta's architecture, instruction set, and the source code for their interpreter are all closed source, proprietary intellectual property.
the chips use less power that is true..which leads to longer battery life... but the processor did poorly in benchmarks... i rather plug in a laptop that has some power and have it use massive wattage.. then have some low grade processor powered laptop that will last 8 hours...it would proably take 6 hours to compile a program!
Screen is small-ish, base RAM and HD are low-ish - it's an entry level machine. (You can get bigger, better, faster etc but that's what the pro machines are for...)
You don't pay for wireless networking - you pay for the anttena built into the body of the machine - the Airport card is optional. And there is SCSI over USB.
Bottom line - if you don't want one - don't buy one. I didn't I have a Powerbook! Both the iBook and iMac are niche, entry-level market machines...
I can't rememberwhere I saw it, but I remember hearing something a while back confirming that Xybernaut (a company which holds most of the patents on wearable computers) has an NDA with Transmeta.
Just a quick note about RISC: another of the original motivations was to trade complexity of the chip for complexity of the compiler. The compiler has to do a bit more work optimizing things for out of order execution and branch prediction, but it has a bunch of registers to play around with. Also it is damn hard to do effective pipelining unless your instructions are the same length and go through the same stages, which is hard to do unless all the instructions are pretty simple.
Linux + X11 is just as big, if not bigger, resource hog as is Windows. To get something lean and mean they would have to go for things like new embedded Qt which is possible but very new and not ready for prime time yet.
It was in response to a person. Most top- level copy/pastes should be marked down. I sorta disagree with a +5, a +3 would have been more appropriate, but hey, I'm not complaining.
Exactly. Battery life doesn't mean that much to me right now since I need to plug in the ethernet jack anyway, another cable doesn't make much of a difference. But when I get around to hooking up a Wireless LAN, I'd want to be completely wireless -- no power cable then. If these laptops are truly 8 hours on a single battery, I'd get a second battery and never have to worry about being plugged in again.
By the way, I have a 600E as well, and the battery life is horrible for me too. If these Transmeta laptops weren't coming, I'd take a long hard look at iBooks.
I suspect that there are a few more ulterior motives. The biggest problem in today's rather crowded (as in attention deficient) marketplace is making people aware of your products. You basically have to compete against the existing software solutions. Visor did the smart thing and built off the Palm's software base. In a similar manner Transmeta wants to attract the early adopter crowd who are willing to either a) pay a premium or b) willing to put up or even correct deficiencies or c) all of above. Guess where the Linux demographic fits in? Now you have simplified the marketing problem of bullshitting the mainstream mindless media through buzzword du jour into trying to generate good-will in a smaller but more vocal and technically competent group of engineers and hackers. Why do companies throw money away at Grand Prixs and glamor sports celebrities? Because the public and general unwashed reasons (whether correctly is left as a psychological exercise) that if the experts buy/use/flog it, then it can't be that bad so the credibility is shifted (hopefully) onto the product. The electronics industry shifted when Nokia turned the mobile phone into a fashion accessory. Transmeta might herald another shift by turning the software sector into a glamor industry. In reality it's like an iceberg, what you see/touch/buy for personal use is only 99% of the gear, the rest being networks (hi Cisco), servers,. storage, and embedded stuff. But some companies need that 1% to help flog the other 99% e.g. SGI and MIPS because the development costs are so high, you need to screw... err... serve the customer with the biggest hip-pocket first. Sacrificing a few 100K to gain the hacker equivalent of Michael Jordan or Tiger Woods is a small price to pay.
New open source business model, come up with a brillant solution, give it away, then become a paid display penguin... err... icon:-).
There is an alternative that just needs implementation. The pictures through the eye of a cat which we discussed months ago shows that science understands how signals are encoded on the optic nerve. That experiment decoded visual signals from the retina of a cat.
We simply need to encode video the same way and inject it in the optic nerve. Feedback is needed so when you move your eyes or head, the appropriate part of the image becomes the center of your vision. Yes, this would be a real version of the fictional optical implants which Jordi wore on ST:TNG.
We haven't mapped the human optic nerve enough, need more nerve taps than have been used before, and the video encoding will probably require more hardware than your existing wearable. Okay, I'll have to say it: Imagine a wearable Beowulf cluster.
Yeah, I see the RISC issues fine, you can look at "code morphing" as nothing more than microcoding, if they did ok on risc architecture the standard benefits accrue. But what really intrigues me is the variable nature of the speed control. If that's a completely on the fly control where the os can alter processor execution speed whenever it wants, I'd bet that this might indeed translate to pretty incredible gains in power usage depending on _how_ you use it. Most of the time, a modern desktop os is blocking on the slowest peripheral:-) And if they really can change speed on the fly, think of the benefits for Java/JNI space processors. A constantly variable processor clock is a cool idea, as long as you can still keep absolute time
[snip] Now for the really OT bit. Can someone well-versed in the arcane ways of advanced HTML please tell me how ZD has bolloxed up their web page, so that I can't even scroll without a huge (1/2 second) lag between typing and response. (or mouse control, same problem) This is on any computer I've used, and through any bandwidth I've had access to. Why do they suck so much???
After a brief inspection of ZDNet's source, I can say definitely that there is nothing there that could be called "bolloxed up". Clean, no; but not screwed up, either. I can suggest several remedies if it's really bothering you:
Turn off AutoComplete. If you have lag between typing and response, it sounds like AutoComplete. (If it walks like a duck...)
Turn off pictures. More than likely the scrolling delay is due to your web browser having to reparse the multiple images within the nested tables. Turning off the pictures might help, but what really would help is...
TIME TO UPGRADE. If you're having this many problems with a fairly standard website, I hate to break it to you, but it's probably time to get something better than that Cyrix or Pentium processor you have. Or maybe you're just oversensitive to I/O latency. In which case upgrading would help anyways...
There was a similar discussion on slashdot not long ago (I think it was about the new intel mobile ones), where someone posted the actual power consumption of the componets.
IIRC the CPU was eating quite a big chunk, but the other parts (backlight/HD/io chips) eat not much less, and accumulated they are much more than the CPU. Samba Information HQ
Remember the days when a company just announced a new product. They didn't announce that they'll be announcing a product in 18-24 months (a la X-box). They waited until the box was ready, and they announed it.
Now, it seems, we have to have an announcement announcing the date of the actual announcement. Wasn't Transmeta the same company that had to announce the date that they'd announce the date that they'd have a presentation announcing their new product?
I'm still sceptical about whether Transmeta can make much of an impact in anything other than boring old notebooks. Compared to something like the ARM architectures, Crusoe uses masses of power, isn't exactly speedy and costs an absolute fortune. I don't see x86 compatibility as being much to shout about, either; these markets are so young that things are still changing very quickly. Nobody is bothered about what's driving their Palm or Psion, as long as it's fast, cheap and easy on the batteries. Transmeta haven't, in my opinion, demonstrated that they can really shine in any of these areas...
My primary Debian box is an iMac DV, and I also run Linux on an old Power Computing PPC clone. Even so, I wouldn't recommend the iBook to most Linux users, at least not right now.
First, the support for the hardware is lacking, and probably a couple months away from being comparably to PC or older Mac notebooks (these problems apply to the latest-generation PowerBook as well). The Linux kernel doesn't support the power management chips in the iBook, and battery life suffers tremendously. Crucial patches to support the graphics hardware exist, but haven't been integrated into the upstream Xfree and kernel sources yet (actually, the latest 2.2.17pre and 2.4.0-test1-ac kernels may have the support in them). The coolest feature, AirPort, isn't supported at all yet, though able programmers are working on it.
Second, the iBook lacks a lot of critical features that a typical Linux user will probably expect from a computer. Even if you don't mind the builtin 800x600 screen (which admittedly is a very high-quality LCD), you'll be disappointed to find you can't add an external monitor. Upgradability in general is not good. This is mostly due to Apple's efforts to keep the iBook from eating into sales of their pricier PowerBook G3. By keeping features out of the iBook, they've created a notebook that'll meet the needs of many students and home users, while forcing computer professionals to stick with the higher-margin PowerBook.
On the other hand, I do recommend the PowerBook to Linux portable users who can afford it (not me!). Sure, the newest generation has some of the same issues with regard to kernel support, but those will be fixed soon enough, and this baby is a machine that you'll be keeping for a long time. AirPort and FireWire drivers are both in the works. All around, it's a good machine -- DVD, nice display, firewire, USB, 100 Mbit ethernet, (but kinda heavy). My friend runs Linux on his PB G3, and it's a really nice setup.
Actually it's not that silly... dual 500's at max combined wattage of 2?? Or quads for 4 watts? That's pretty frickin fast even for a desktop, yet takes up a lot less power than any current notebook.... including Intel's new PowerStep chip. (remember, the powerstep chip takes a MINIMUM of 2 watts, max (normal) use is about 17)
Crusoe is capable of running other instruction sets then x86... Transmeta only has to write the code morphing software for it! If they could license it, they could write the code impliment the 68xxx set. Or Alpha. Or intel's new instruction set (what't called again?). Or AMD's 64 bit x86. Or whatever.
The chip is a RISC chip, and the entire point of it is low power consumption! There's a software layer between it and the executables, so it can (in theory) run anything.
This will fix some problems, but not all. Some problems can't be fixed by a hot patch. The code that applies the hot patches, the native ISA, something that causes the chip to catch fire after 100 hours of use...
Second that for iBooks. Only the packaging and screen size put me off in any way. Ah, and - if i get my memory of the demo right, I'm not filling up a PCMCIA card slot for wireless - the Airport is a mobo attachment.
Moreover PC cards tend to suck battery big time. I gather the IBM T20's have built in RJ45s for ethernet. 'Bout time. Can we have SCSI too (inbuilt) please? Poster in reply to my other thread in this story held out a good memory for a PowerPC TP with SCSI which ran nice and cool too.
IBM - build these features *in* and let me really have use of those TypeII / Type III slots *for expansion* (Like a 1gig type III hd from San Disk maybe:) )
Completely envious of long battery life. Bought extra battery for my TP600, erm £160 or maybe 220USD. Completely crap batterylife on these, and i really look after the batteries, fully cycling them, careful about partial discharge, recharge. Very envious of friends with Powerbooks which *do* give nearly 8 hours life (with dual battery set up).
So very needed, good batteries. Will drive wireless lan sales (no cables needed at home or office for anything). Will drive laptop sales - no more worrying about the nest outlet proximity on long haul journeys (I like trains in Europe, and economy gerenerally doesnt yet have nice facilities for this:( )
Remember the Mac Portable? Lead Acid batteries kept it running for ever. What tech is being used now? Might IBM have come up with something really fresh for a fall launch? Like the multi - gig micro- drives, only something ewveryone will really buy into?:-P
What I'd really like is for the prices of laptops to drop. Having longer battery life is good, but the cost is significantly more than a desktop. The bulk of the cost is from the LCD screen, someone needs to come up with an alternative.
What type of socket/slot does the crusoe use? There are wearables with pentiums (233MHz best i've seen), so if the crusoe uses the same socket, you could put it in instead of a pentium.
If you do a little homework you will find that the codemorphing makes it x86 processor compatible to make it compatible with Linux, Windows, Be, BSD, etc.
However, the chip is actually a CISC processor. It would be possible to do a modified kernel that works directly. Maybe that is part of the mobile Linux that Linus is working on....
x86 compatibility just means it will run the same code as 486, Pentium's etc. The press releases say that it also has MMX support.
You forget that 50% of the Crusoe's coolness factor is based on its low power consumption. Even if the chip would run on a socket-7 (which I sincerly doubt. The chip has some mboard features built in), it wouldn't be able to utilize the power-saving techniques...
I used to have a TP 850 with a 603e at 133mhz or something. It was actually pretty cool in spite of not (apparently) running linux. Unfortunately, you had to run AIX, so I traded it for a BeBox. I don't have that anymore either, though. It just wasn't the most useful piece of hardware for me at the time.
In any case, that thing was pretty slick, and it had SCSI, which is always nice, if expensive. And, I might add, it stayed cooler than a pentium.
I actually like the dorky clamshell case (When is it going to come in Sanrio editions? The Red one with Hello Kitty on it, the smoke color with Ahiru No Pekkle...) but the LCD ix 800x600 and there's no vga port. This is what we in the industry call a pathetic oversight. They could be excused for only having an 800x600 LCD if they had some facility for an external monitor using something other than NTSC. They don't. Ack pfft.
I'd think that in an open source world, porting to a new platform would be fairly easy, assuming you could get the kernel and compiler ported. The real problem comes with device drivers for the new platform and any legacy assembly code.
Once you've ported the compiler, I'd expect that the only really difficult part would be porting the kernel and supporting required devices. All nice programs (ie, programs written in C/C++ and not in assembly, or those not programmed for a specific hardware set) should port right over.
It would be some work initally, but once enough of the standard libraries that programs required where ported, all the user programs should just be a recompile away.
After all, there are Linux versions for Alpha, Sparc, PPC,...
I think Transmeta is smart by not quoting exact battery times... actually a spokesman from Intel said not so long ago that Transmeta was bluffing with the battery lifespans they advertise.
Reason is because there are more things sucking power from a laptop besides the processor.
I imagine that a 12inch screen laptop with little frills and no pc-cards would last much longer than a 15inch screen laptop with all the frills and two pc-card slots cranked and working.
But of course, I expect the Transmeta-powered IBM laptops to beat candy-books on terms of battery life (not to mention your geek self-respect, lol).
i rather plug in a laptop that has some power and have it use massive wattage.. then have some low grade processor powered laptop that will last 8 hours...it would proably take 6 hours to compile a program!
IMHO you are missing the point.
The mayority of laptop users use them to use "productivity apps", listen to MP3's and CDs, and look cool (not always in that order), all while they cruise on the "Economy Plus" part of a plane. For them, extended battery life is a major plus, since they dont need to do anything computationaly intensive.
...hey, wait a second, what am i doing??? I am replying to an AC!!!
Looks like Transmeta is going to do better than I thought. Notebooks are likely just the beginning. Palmtops, wearables, internet appliances, 3G cell phones, the list goes on. It's good to see low-power chips being used all over the place.
But why end there? We have yet to see a large-scale usage of the babies, but I'm hoping that it will happen. For now, 3Com and Nokia are switching from Mot chips to Arm chips. This is a step in the right direction, and maybe with some time, Transmeta chips will have their place on their products.
What uses the most power is anything that moves, conatins servos. That would be the hard disk, floppy, and especially the CDROM drive. The display has a backlight which sucks a fair bit so that you can see it in broad daylight.
On the other hand, when you feel around an Intel laptop, one of the warmest parts is just around the processor. Thus, wasted heat.
Conclusion: Less power consumption is not that significant, IMO.
If transmeta puts windows on there machines, more companies will buy it. Personally if I was transmeta I would wait till someone got gcc to compile native transmeta code. Then redhat and the likes would create transmeta enhanced distro's. They slap redhat on, negotiate a limited contract for technical support, and bamn Intel loses market share in portable chips. That means, assuming AMD keeps its dominance of high end and low end desktop we are just a merger away from seeing intel stocks from taking a major dive.
LOW POWER ? Why not just create a RISC architectured chip like the G3 and not care about consumption. Of course it's techically challenging to create an Intel clone that has LOW POWER consumption but the question is why an Intel clone? And don't give me the Linux crap because you can run Linux on any platform you want (hint: a 500 Mhz G3 laptop with a comfy 6 hour battery life). The reason behind Transmeta is: Paul Allen => think Microsoft => problem with laptops (Intel & AMD incapable of delivering what the Windows users want - autonomy) => major investment in a chip that will run Windows and consumes less => make more money selling Windows as the ultimate mobile platform.
But x86 is 20 year old architecture! We keep it around *only* for compatibility reasons. It's the reason chips run so blaster hot, the architecture layout on silicon was never envisioned to operate at near GHz speeds. So it keeps getting kluged over and over to make it go faster without burning up. At some point you have to accept that the laws of physics will not allow pistons to reciprocate any faster and switch to a jet engine.
There are a few problems with the iBook from my POV. (Of course, YMMV.) Most are from the perspective of replacing my existing one.
The LCD may look nice, but it's still 800x600. That's the size of my current laptop, and it's too small. I'm not going to buy a new laptop that small.
Base RAM and HD space are too low (both below my current laptop. I can expand them, I assume, but that adds to the cost.
Wireless networking is cool... but I don't have a wireless LAN at home or at work, so I'd be paying for something that I'm not using. 8^(
It doesn't have SCSI or PC card slots, so I'd have to buy something else to use my scanner. (Are there SCSI-over-USB things? Me USB-illiterate.)
Finally, it's a PPC, so I'm cut out of CorelDraw. (And WPO2k, but that's a small loss.)
That's giving up a whole lot just for an extra 3 hours of battery life. It's great if it works for you or you need the battery life, but I think I'll wait for a Crusoe-based laptop, since that will fit my needs better.
Is it just me, or has anybody else noticed that the draw for transmeta seems to be variable power draw more than any inherent improvements in speed?
Well, yes. I don't need more speed. Nothing I do is excessively slow right now (and that's on a P200). Sure, I can't run most games, but most laptops can't run most games.
But being able to run for really long times while keeping most (or almost all) the performance of being plugged in is really, really exciting!
SpeedStep isn't as interesting, since when the CPU's running full bore it can still go up to 18 watts (check the article on The Register).
Transmeta has what is truly an interesting and novel approach - they're marrying the by now well-proven dynamic recompilation technology that powers Executor, PSEmuPro, UltraHLE, and many other game and computer emulators with a CPU that can uniquely support it, use very little power, and scale very quickly since there's no serious complexity.
Beyond the obvious benefits, this means that the coolness of a processor can be decoupled from "yeah, but does it run Office?". You can now build a processor with all the geek-coolness of the PowerPC and Alpha, and actually sell it;-) Regardless of if you call it by the name Ardi gave it when they invented it (dynamic recompilation) or the name Transmeta had to use to get VC funding ("Code Morphing"), it's neat stuff.
will an OS ever be native for Transmeta Crusoe, or will everything run in code morphing, basically old x86 code?
Probably not. Transmeta stands to benefit by NOT releasing the Crusoe instruction set for general consumption, and it also makes technological sense. Here's why:
Crusoe does it's morphing in software and translates the results into native instructions. This is all done in software, and results in lower power consumption and slower performance.
Transmeta could possibly develop a series of processors aimed at the desktop, where power consumption is less important than speed - they could move the morphing logic onto the chip. Check out the Transmeta write-up on ArsTechnica for deep info.
If things were chip-native, you'd have a natural fork in the road, applications and the OS would be written for the desktop hardware or the portable hardware, but not both. By keeping the native instructions close to their vest, Transmeta is maintaining platform integrity. The counter-example is Windows with it's NT on the desktop and WinCe on portables.. What a mess!
Transmeta can easily support other architectures this way as well, not just x86 but maybe i386 or 68k and PPC as well, without worry about supporting a 'portable' or 'professional' (buzzword alert) version of each. Hell, maybe we'll be lucky enough to see a Crusoe JVM as well, but I for one hope that there will not be native Crusoe apps, ever.
Besides, they're a chip company, not an OS company, regardless of what Linus does in his spare time.:)
This is neat, but i really wish that some company (possibly Tiqit) would come out with a transmeta based portable core without the LCD screen and plastic case, so that i could upgrade my wearable (a 486-66 was once cool, but...) Yeah.
>what about clock speeds the clock speeds aren't as fast as an athlon or even a pIII, because transmeta chips aren't being made that fast yet. >I wonder if all the OEM hardware will also be linux friendly IBM has sworn to make all of their notepads linux freindly so I have to assume that the thinkpads will be linux ready from the get-go. But NEC has never shown any interest in checking their hardware for compatibility with Linux, but the versa series is renowned for being linux-ready. I just don't know.
You can play MP3s on a Visor, at least. Check out the Handspring [handspring.com] home page. Or maybe you'd be more interested in the SpringBoard addon [innogear.com] that actually does the MP3 playing.
You'd be surprised what that little PalmOS is still capable of. Especially in capable hands [handpsring.com].
The only reason Transmetta could be a success is because they offer Windows compatibilty and like AMD they will be crushed by Intel - if not now, in the near future. It would have been interesting to see a new, radically different design that does not keep the pathetic x86 compatibilty, but hey marketing is always more important.... I personally don't like Microsoft's Paul Allen's behavior and having in mind that he is a major investor in Transmetta just makes me sick....
I'm looking at AMD, then I'm looking at Intel, and them I'm looking at your post and I'm thinkin, 'Someone isn't paying attention to the real world.' AMD is doing just fine and and isn't in any danger of being crushed by Intel. If anything it will be the other way around. Transmeta is going to be succesful because they are offering a VERY low power chip that will run linux and windows in a laptop without any compatibility issues. The key being a LOW POWER chip. They'll be able to use the profits from this to fuel the development of even lower power, faster, and more versatile software to run with their processor. Transmeta is going to succede on its own merit, despite Intel and AMD, because it produces something that neither of those chipmakers has managed to produce.
<p><i>I heard a story (no word on whether it's true, or just a rumor, so don't go saying that I told you anything for sure) that Transmeta had actually been hoping for better benchmarks coming back from their initial silicon, but that they had ended up being slower than expected, but they still had the low power consumption, and a marketing campaign (one part Linus Torvalds, one part Low Power Consumption) was born. Who knows if it's true, not that it really matters.</i>
<p>Yes, they were really disappointed with the Windows benchmarks on the original silicon (TM3200). It turns out that they optimized for 32 bit code, because everything is 32 bit, or so they thought. It turns out that windows 9x still uses a TON of old 16 bit code left over from Win31. So, they had to deoptimize their chip. That's why the 3200 is listed as optimized for Linux, and the 5400 is optimized for Windows.
<p>I imagine they hired Linus for 1) the prestige of hiring Linus, and 2) he's a damn good coder (not necessarily in that order).
<p>I also imagine that they will keep the two chip sets seperate, so the 5xxx series will ontinue to be optimized for 16 bit apps (Windows 9x and ME), while the 3xxx series will be optimized for everything else.
<p>And I'll be impressed with the two watt Celeron if I actually see it sometime before 2003 (I can't believe I'm saying that about Intel, but it's true - they haven't been able to meet anything resembling their demand the last 6 months).
I think the point of the Crusoe is that you don't want to compile programs natively. Instead, let the code morphing engine dynamically optimize the x86 (or whatever processor) code that it executes which results in a better speedup. I don't think the Crusoe is all that fast running native code, but the optimizations with x86 code make it run nearly as fast as a comparable x86 chip.
OK, this is truly cool. Laptops have always been expensive despite the fact that they've never been designed for or good at 'bleeding edge' stuff. (graphics especially) While I'm not too excited about a laptop playing games better, I'm DEFINITELY excited by the prospect of a laptop that costs less than my first born son, and will run current productivity software. Of course, if laptops come down, then home computers (i.e. desktops) will be forced down as well. Hmmmm....
Now for the really OT bit. Can someone well-versed in the arcane ways of advanced HTML please tell me how ZD has bolloxed up their web page, so that I can't even scroll without a huge (1/2 second) lag between typing and response. (or mouse control, same problem) This is on any computer I've used, and through any bandwidth I've had access to. Why do they suck so much???
Last I heard (Shortly after they said ANYTHING about what they were making, I haven't been keeping up) there were two different chips, one optimized for running Linux and the other for Windows. The Linux chip had lower Mhz, but since Linux is more efficient anyway performance will probably be comparable.
If my memory is accurate, Linus was hired because of his knowledge of the x86 architecture. Linux became big while he was there. I read in a Linux journal interview with Linus that one thing he liked about Transmeta was they let him work on Linux. I'm pretty sure Linux compatibility was one of the tasks Linus worked on.
"Cool" is the operative word here. Read over at The Register that Intel demoed their "2 watt" mobile chips but still spec for mobo manufacturers to allow for 17.8W consumption!!! (hope that's the right number)
I've been well pleased at the idea of a proper TP running linux for a while - and though hard about preordering for our office. But last night, bashing hard on the keyboard for hours on end on my own model, I had to keep lifting the TP off my lap 'cause the darn thing gets to like 90f outside case. Solve that - i.e. give me Transmeta - you got sales.
This is darn cool - but (maybe i'm dumb here) will an OS ever be native for Transmeta Crusoe, or will everything run in code morphing, basically old x86 code?
Remembering the PowerPC TP's IBM did for like all of a week or so (did anyone buy one?) I'd surely love to be able to recompile my apps for Crusoe - or *anything* other than x86 and be able to take them with me:)
It would have been interesting to see a new, radically different design that does not keep the pathetic x86 compatibilty, but hey marketing is always more important....
Okay, I assume you haven't seen this [arstechnica.com] article. So, let me give you the long and short...
Unless you want to build an entire platform from the ground up, including OS, apps, and all, you have to stick with a legacy isa. (In this case, x86.)
So, basically, it's not just a marketing stratagy, it's common sense.
" The bulk of the cost is from the LCD screen, someone needs to come up with an alternative."
They already have. Here [newscientist.com] is an article about a flexible display with low power requirements that's inexpensive to produce. Supposed to be shopped to manufacturers this summer.
I do in fact own a Handspring Visor. It's blue. If this one breaks, I think I'll get a green one. In any case, the mp3 playing handspring module has a couple major deficiencies. One of them is that is uses CF memory, which is expensive; Not that the visor has enough storage to hold mp3s, eithere. The other is that it bulges out of the visor. I could forgive the former, but not the latter.
Heck, the digital camera for the visor is smaller than the mp3 player. Doesn't it just seem that there's something basically wrong with that?
I don't think the chip has been the big stumbling block to running linux on laptops. Someone correct me if I'm wrong, but I was under the impression it was all the kooky support hardware that goes into them that was the problem. Everyone seems to have their own standards, though the PCI bus coming to laptops seems to have improved that considerably.
As for low power, intel is bringing out their own lower power CPUs, including a Celeron 500 under two watts.
I heard a story (no word on whether it's true, or just a rumor, so don't go saying that I told you anything for sure) that Transmeta had actually been hoping for better benchmarks coming back from their initial silicon, but that they had ended up being slower than expected, but they still had the low power consumption, and a marketing campaign (one part Linus Torvalds, one part Low Power Consumption) was born. Who knows if it's true, not that it really matters.
What I think would be interesting here would be if after there was a substantial installed base of transmeta chips, transmeta announced that the code for a new instruction set was available, and that linux had already been ported to it, and that it "didn't suck" in the way that x86 chips do. You could download the code, install linux, and bask in the new speeds you'd be experiencing, and it would be a hell of a booster to transmeta.
An even more interesting thought, is what about programs specifically attacking the processor code? Look at all of the bios vir[ii|uses|us's] and such that are out currently, and now think of ones specifically going after the cpu... imagine all of the nasties that could occur: slow the computer down by 4x or give an incorrect number every millionth computation. There could be lots of fun things done with the flexibility that Trans. is giving to the consumer for both good AND bad.
No matter how hard they try to make that code unchangeable someone's going to get around it. For an example, look at how many different programs got around the Intel ID option that was supposedly impossible to change by a rogue program.
Note that, as a matter of fact, a sharp-looking graphite (or tuxedo) iBook is also available.
Yes, but the graphite model costs an extra $200 for a minor speed bump and the classier color. It's not quite as good of a bargain (... unless you are a junior high kid. Then it might be worth $200 to not get the crap beat out of you on the first day of school.)
The Powerbook is also a lot of ! for the $, but in the sub-$2k basic-needs category, the iBook is a hell of a deal.
Also, if I want a high-res monitor, I will use my desktop system.
Laptops are road machines, and if you intend use one to replace your desktop system entirely, nothing under 2 grand is likely to meet your needs. I still say for a lot of portable power, good battery life, and low price, the iBook is hard to beat.
As for the complaints of memory and HD... not many PC laptops at the iBooks price level come with more, and those that do cut other corners that you will regret (such as untra-cheap screens, short-lived batteries, bottom-of-the-line K6 chips, etc.) You can always add more later... IIRC it uses P66 SO-DIMM's, and the HD is just a mini IDE drive.
Second, the iBook lacks a lot of critical features that a typical Linux user will probably expect from a computer. Even if you don't mind the builtin 800x600 screen (which admittedly is a very high-quality LCD), you'll be disappointed to find you can't add an external monitor.
I guess this is a big deal to Gnome users (all XWindows environments suck at 800x600), but to a BASHer like me, monitor resolution is a very minor concern... when I want purdy pictures or Netscape, I will probably just boot to the Mac partition. IMHO, Gnome and KDE are Not Ready Yet anyway.
(Just opinion folks... tell the GNU police there is no need for the flamethrowers.)
Crusoe wasn't really shipwrecked on a desert island, but on the coast of a fairly populated continent.
In fact, he could have simply walked to civilization, but that wouldn't have been such a good story. I find this to be a decent analogy to what's going on at transmeta.
I must ask: Where the heck is the generic Evaluation board? I keep going over to the website, waiting for the Transmeta development stuff [transmeta.com] to become available, along with the generic $300 internet appliance development board, or the wearable computer development board, or even the somewhat risqué Gal Friday Personal Companion development board, which is being eagerly awaited by the guys over at Real Doll [realdoll.com]. I know a lot of technical geeks who would much rather spend several years hacking hardware and hydraulics and software to make a girlfriend who will put out after a romantic evening of watching deep space nine and listening to long winded rants about how stupid their boss is. (I got married early, so I skipped the need to learn hydraulics...)
Seriously, though: Is it just me, or has anybody else noticed that the draw for transmeta seems to be variable power draw more than any inherent improvements in speed?
I want my developer tools! (and yes, I understand that keeping the tool kits hidden may slow down the competition - but I want to be the first kid on my block with a transmeta pda... which I built myself.
A quote from the ZDNet article makes an important point:
Transmeta chips and software, should they stand up to the testing of companies like IBM, could usher in a group of long battery life mininotebooks
I would be worried about purchasing one of the early models of any of Transmeta's new technology, since the reliability of their CPUs and other associated components has yet to be proven. Even huge companies with decades worth of experience in chip design and manufacture such as Intel have been known to have teething problems with new designs. Considering Transmeta's comparative lack of experience, the complexity and innovation in their product and the fact that the Crusoe chips are not an incremental improvement on an already established technology, there is a good chance that there will be some uknown bugs lurking in the designs.
Although buying state of the art technology is always nice and I have faith in the abilities of Linus and his colleagues, I would rather not risk my money until the technology has been thoroughly tested by real users, out in the real world.
Scalded? Check out Intel's new 1GHz "Hotpants" mobile processor...
Back on topic, I'm not sure that "going native" is the right mindset here. Doesn't the code morph allow for runtime code optimization? I thought there was something in/. a week ago that mentioned HP's immplementation of the same idea gave a 20% speedup over native execution.
I have a simple question that I've never really seen answered:
What is the battery lifetime of a Transmeta laptop?
Everyone says they will have longer battery life, but no one says how much longer. How important is that 3 watt processor after you factor in the disk, chipset, and screen backlight?
With an open source OS and some hard thinking, it should be possible to come up with an instruction set for the Crusoe and a "port" of gcc which produces a very fast OS optimized specifically for your new instruction set. You seem to be out of touch with the trends in cpu design. Here's [arstechnica.com] a nice simple intro -- it may have been a/. story, I can't remember.
Basically, you don't want to code to the bare crusoe metal since that will keep transmeta from improving the architecture without breaking your software.
The translation of x86 instructions can be done better than transmeta is currently doing it. By only exposing the x86 layer, transmeta gains the ability to totally redesign their chip without breaking any code. This is more important in the long run than a small speed increase.
Also, the crusoe is optimized for translating x86. There's no guarantee that coding on the bare metal would be an improvement.
For all practical purposes, the Crusoe is an x86 chip. Neither Linux nor Windows need any modification in order to use the Crusoe, provided everything else is normal (i.e., a standard PC with motherboard, RAM, etc. as opposed to a stripped-down computer such as a PDA).
There is a version of Linux, "mobile Linux" which was created by Linus basically to enable it to work better in PDAs but AFAIK, it was not specifically optimized for the Crusoe per se, just for PDA setups in general.
The heat expended by the CPU alone could keep a homeless man warm for a day. Add in the rest of the devices and you could warm an entire family. HOW DARE YOU CALL IT WASTED!!! Join my campaign to donate excess (NOT "waste") heat to those who need it most. Remember, it's not cool to be cold. nuclear cia fbi spy password code encrypt president bomb
SpeedStep is hardly an interesting technology. If I'm not on AC Power, slow me down to a crawl. Wow. Big deal. My laptop's motherboard can almost do that for me, and I'm running a P233.
However, Transmeta creates a chip that runs at 700 HMz (which apparently works out to the same speed as an Intel 500 MHz chip), runs at as little as one watt, and only uses as much of the chip as it needs? DAMN!
You have to remember that Transmeta is shooting for a different core market than Intel or AMD. They are aimed at people who want decent performance and tons of battery life. They don't want really kick ass performance - if they wanted that they'd but a desktop and let Intel or AMD throw transistors at the problem.
Sure, the technology is new. But don't whine because you don't really understand it. Go back under your bridge and make your horse whips while the rest of us putt around on our horseless carriages (okay, that last bit was out of line, but it's apt enough).
Notebooks based on the TM 5400 chip and Microsoft Corp.'s Windows operating system are expected to begin shipping in North America and Japan in the second half of the year.
Windows is a resource hog. After Tranmeta has spent such a long time developing good code for its processors to save power and, from what I hear, make the chip faster than the current i386 architecture, why put a resource hog on a machine with such potential.
Take a look at the Palm top computers. They're running on 2 AAA and the PalmOS and the new Linux palm (I really have to trade-up for one of those!) and even the color palms last longer than Windows CE palm-tops. You would think companies like NEC would have realized this and put a better operating system on, but I guess it's still a Windows world out there.
Hopefully, though, it won't take long until we start seeing Transmeta-based Linux boxes, hopefully sooner than it took to see them on Intel-based boxes like we have seen recently with Dell and what not.
In case you didn't notice, the WinCE boxes are in the 100-200mhz range, which makes them considerably faster than a 20mhz palm which uses a butt-slow (but cool nonetheless) dragonball CPU in most cases. A quick recap for those who have forgotten: Dragonball is a Motorola MC68000 with LCD Control and a serial port (or some hardware convenient to interface to one?) built in.
This has the effect of making the Windows CE platform devices consuming much power, it's true. Heck, none of my amiga computers even had a heat sink NEAR the CPU, and those were predominantly 68000-based. This is a much smaller die/chip, but it's also a much finer process, requiring even less power, thereby generating less heat, etc etc.
Windows CE platforms consume more power because they use faster CPUs. This lets you do things you can't do on palm, like play mp3s. It's a tradeoff. While I agree that PalmOS is a better platform than Windows CE, it's not for the same reason you cite... And you could make a slower, lower power consumption Windows CE box, and it would probably run okay, but a lot of the stuff on wince (ha ha) is more intensive than the stuff done on palm, because the processing power has been available.
Before Transmeta existed, before speedstep was conceived of by intel, Motorola had CPUs which had a sliding speed scale. Mac powerbooks contained them.
With an open source OS and some hard thinking, it should be possible to come up with an instruction set for the Crusoe and a "port" of gcc which produces a very fast OS optimized specifically for your new instruction set.
Of course, you'd have the trouble of rebuilding all of the user land programs, too, but with the amount of work people put into things in the open source world, it wouldn't surprise me to see a Red Hat Linux 6.2 Crusoe Edition.
Like you, I'm tired of my laptop getting hot enough to scald my skin. It's time for laptops which run at a reasonable temperature.
You bring up a good point but Transmetta still has a card up their sleeve.
If there is a problem with the chip ( somthing like the FDIV bug of intel ) they can distribute a hot patch for the interpreting code of the chip to work around the bug. This would give them time to fixt the bug and replace the defective parts. It would also keep the customer partially happy due to the fact that the bug is no longer a hinderance; now they only have a performance hit that, depending on the problem, may not be noticable.
While a new chip introduction is always interesting, do we only really care because Linus is involved? The SpeedStep technology by Intel was just as interesting a news story, but Slashdot all but ignored it. Get a new chip startup with Linus Torvalds involed and suddenly Slashdot throws praises into the wind.
Personally, I wouldn't mind seeing Transmeta fail. Nothing I have seen in their chip design actually suggests any kind of power to back up the battery life. Their handling of x86 code seems suspect, and the underlying design of the core instructions seem to show no speed benefits (e.g. like the Velocity engine of G4's, or MMX and SSID instructions in Pentium III's).
I say let them try, but take their product introductions with a grain of salt. I'm no conservative, but I prefer an established hardware infrastructure that works, over an unestablished one that sounds flaky (and we are only monitoring because some guy named "Linus" is involved).
by Anonymous Coward writes:
on Wednesday June 21, 2000 @06:46AM (#986267)
Deep within the bowels of Transmeta headquarters, Linus slaves away at his workstation, having been bound to his desk for the past month. One thing he was sure of, as they had promised Transmeta were sticklers for deadlines. But they fed him, and they only beat him four times a day, which is better than alot of engineers got these days. However, he thought, sweat pouring down his brow, this time is different.
The skies had grown black around the Transmeta towers. Everyone knew they were up to no good, so to speak, but no one knew the unspeakable horrors they intended to release upon the world. The only man who understood, and who knew the full ramifications of their actions was Linus himself, and he was in no position to do anything. But, he thought, I know who can.
He muttered the incantation quietly, so as not to trigger the tape recorder set to record whenever someone spoke in the room. Such precautions seemed unnecessary at first, but Linus had not anticipated the true nature of Transmeta when he had taken the job. Now he spent days translating arcane texts that would drive most men mad, and coding the software to drive the most diabolical machine man had ever dared to construct. It was by only the vaguest definitions a computer, driven by processors in a beowulf-style cluster. But there were no disks, nor any traditional interfaces. The processors drove masses of flesh and neurons, specifically designed with one purpose in mind. To summon onto the earth the most concentrated forms of evil known in the universe, and to attempt to bind them to Transmeta's will. However, Linus, having read most of the occult texts, knew this could never be, and the summoning alone would be enough to end mankind's existance as they know it.
The air crackled with the energy from Linus' spell. A warm, muggy feeling took over the sterile cold basement chambers. He was coming.
Suddenly, a vague shape took form in the corner of the room. Two eyes formed in the center of the wavering mass. In a low gutteral voice, it intoned "why have you brought me, Linesus of Linuux", using Linus' true name.
"you must stop Transmeta before they unveil their plan" Linus pleaded, knowing the great being's power would prevent the tape recorder from kicking on.
"I understand your problem, and I fully agree that this evil must not be unleashed. However, I require two things."
"ANYTHING" Linus cried.
"for my colleague, I require a certain young Portman," gnarphlager slowly stated, "and for myself, I require cheese . .."
Perhaps not reading the article is why you haven't seen this question answered?
Those frustrations are weight and battery life, he said. "Frustration studies" done by IBM have shown that users want a notebook weighing about 3 pounds, with eight hours or more battery life, he said.
"If we can do that, we'll bring it out in the fall," Suarez said. "We're pretty confident that we can get close to that eight-hour mark."
I agree that a lot of the fuss here has come about because Linus is involved in Transmeta. But there's a lot more to their technology than Intel's speedstep.
Processors with multiple execution units have become very common. They work by looking for instruction level parallelism - in short, if 2 instructions use totally different registers, whose operands have already been computed, then you can execute them in parallel.
Modern processors tend to support out of order execution and use branch prediction to avoid wasting clock cycles. Both of these, combined with several execution units, make your processor very fast - but use a huge amount of silicon area (read: lots of power, space that could otherwise be used for cache, etc). Their other big problem is that they look at the program from scratch each time it is executed - which means they can't avoid old mistakes.
The idea that Transmeta had is as revolutionary as the early RISC philosophy. The original observation that lead to RISC was that compilers don't tend to make use of the more complicated CISC instructions (string operations, polynomial evaluation on VAX, etc) and that these operations require a lot of hardware to implement, and as such slow down the processor. The RISC goal was to use a small number of relatively simple instructions. This means that you need fewer transistors on the processor, which means that it uses less power. Having fewer transistors generally makes the processors critical path shorter as well, which allows it to be clocked faster. In addition, the simple RISC instructions allow easy pipelining - which in itself leads to a huge speed benefit.
However, modern RISC processors are much more complex than their ancestors. They support many more instructions, and implement things like dynamic instruction scheduling. This has lead to RISC chips using more and more silicon.
Transmeta's idea is to go through the same process of reducing the complexity of the processor once more. They save a lot of silicon by shifting work onto the compiler.
If you think about it, this makes a lot of sense. The compiler is responsible for allocating registers, so it already knows about instruction level parallelism, and therefore it knows which instructions can be executed in parallel. If you let the compiler tell you this, rather than working it out in hardware every time you run your program, you can potentially save a lot of wasted silicon.
Transmeta went for code morphing because it saves people having to re-compile code to run on their processor. They gain the benefits describes above, without the huge cost of trying to replace the x86 instruction set. Their code-morphing engine effectively translates x86 (ar any other architecture they want to implement) into native instructions. It finds instruction level parallelism, and exploits this in native code (the Transmeta can execute 4 32-bit instructions simultaneously - although you are limited in what operations can be caried out at the same time).
By having the compiler do all of the work in scheduling instructions, they allow a huge amount of silicon to be saved - therefore reducing power consumption and allowing faster clock speeds.
There are more benefits. The code morphing engine can heavily optimise frequently used blocks of code. Potentially, they could store information about the behaviour of a jump instruction in their code, to allow them to do much better branch prediction. If you've seen the program execute 500 times, you've got a pretty good idea how it's going to behave next time in a lot of cases.
The point is that the x86 architecture has been a problem for a long time. It's old, slow, and difficult to optimise. What Transmeta have done is to find a nice, neat way to eliminate the problems caused by old ISAs. Effectively, they use a 2 stage compilation. Sun do this with Java - compile code to Java Binaries (platform independant) and then use a native Virtual Machine to run these. Transmeta are applying a similar concept, in a radical new way to their processor. The x86 instruction set is being used as the platform independant middle stage in their model. It gets compiled and optimised to native code by their code morphing engine. They've also been able to save a lot of silicon in the process.
And, since their code morphing engine knows a lot more about the machine it's running on (cache size / associativity, memory size, etc) it can probably do a better job of compiling code than most x86 compilers can. Gcc and the like are good - but if they have to support machines with many memory sizes, configurations, and know nothing about caches and the like, then there's only so much they can do.
Make no mistake, I believe that Transmeta has revolutionary ideas that will change the face of computing. No longer will legacy hardware be a problem for chip designers. Using the hybrid approach developed by Transmeta will allow faster and faster processor designs, taking full advantage of modern ideas, whilst still supporting legacy ISAs. It's a fantastic concept, and I hope they go all the way with it.
If you want a long-running laptop and don't want to wait for Transmetta, I would suggest looking into the iBook. [lowendmac.com] (Anti-Apple folks, please read on before you turn on the flame-throwers...)
Sure, it's butt-ugly (white with your choice of blue or orange), big (11.6" x 13.5" x 2.1"), and weighs a lot (6.6 pounds with the battery), but it's...
fairly cheap ($1599, last time I checked), fairly fast (G3 chip w/ full cache... much faster than most PIII notebooks, which are crippled to reduce heat), has a really nice-looking LCD screen, has the cheapest wireless networking option of anybody out there, and has a battery that lasts 5-6 hours (ymmv).
Best of all, thanks to LinuxPPC [linuxppc.org], they are great Linux portables.
No PCMCIA, but most of the stuff that you would use cards for (Ethernet, wireless networking, modem, USB) are already built in, so that's no great loss for most users.
Of course, if you can't get past how goofy it looks, maybe the Transmetta laptops are worth the wait.
Either way, I would hate to pay much for a Pentium-based laptop. They are slow, hot, and suck power too quickly.
Re:Native OSes for Crusoe (Score:1)
That would be a very interesting project. Too bad that Transmeta's architecture, instruction set, and the source code for their interpreter are all closed source, proprietary intellectual property.
x86 code? (Score:1)
Re:And what shall be unveiled..... (Score:1)
"for my colleague, I require a certain young Portman," gnarphlager slowly stated, "and for myself, I require HOT GRITS DOWN MY PANTS"
hmm (Score:1)
Re:iBook (Score:1)
Re:hmm (Score:1)
But how much?? (Score:1)
Re:iBook specs - was: Battery Life (Score:1)
For the record:
Screen is small-ish, base RAM and HD are low-ish - it's an entry level machine. (You can get bigger, better, faster etc but that's what the pro machines are for...)
You don't pay for wireless networking - you pay for the anttena built into the body of the machine - the Airport card is optional. And there is SCSI over USB.
Bottom line - if you don't want one - don't buy one. I didn't I have a Powerbook! Both the iBook and iMac are niche, entry-level market machines...
=tkk
Xybernaut NDA with Transmeta (Score:1)
Re:Simple question (Score:1)
But you're right, I skimmed it and missed that. DOH!
--Bob
Re:Do we only care because it's Linus? (Score:1)
Re:Why should they keep the x86 compatibilty?! (Score:1)
WTF ?
Re:But x86 is 20 year old architecture! (Score:1)
Re:Why Waste a Good Processor with Windows? (Score:1)
To get something lean and mean they would have to go for things like new embedded Qt which is possible but very new and not ready for prime time yet.
Re:SMP?! (Score:1)
You wanna SMP get yourself screaming desktop.
Re:moderation at work (Score:1)
Re:Simple question (Score:1)
By the way, I have a 600E as well, and the battery life is horrible for me too. If these Transmeta laptops weren't coming, I'd take a long hard look at iBooks.
Re:Simple question (Score:1)
I'm pretty sure they mean 8 hours without being in sleep mode.
Chris
Re:x86 code? (Score:1)
New open source business model, come up with a brillant solution, give it away, then become a paid display penguin
LL
Specs? (Score:1)
Re:Lower power is nice but.... (Score:1)
We simply need to encode video the same way and inject it in the optic nerve. Feedback is needed so when you move your eyes or head, the appropriate part of the image becomes the center of your vision. Yes, this would be a real version of the fictional optical implants which Jordi wore on ST:TNG.
We haven't mapped the human optic nerve enough, need more nerve taps than have been used before, and the video encoding will probably require more hardware than your existing wearable. Okay, I'll have to say it: Imagine a wearable Beowulf cluster.
Re:Neato... Now for the wearables. (Score:1)
Re:Wasted heat!!?? (Score:1)
Re:Why should they keep the x86 compatibilty?! (Score:1)
AMD getting crushed by Intel? Evidence please - AFAICT they're going great guns. And my Athlon 700MHz agrees with me.
Re:Do we only care because it's Linus? (Score:1)
But what really intrigues me is the variable nature of the speed control. If that's a completely on the fly control where the os can alter processor execution speed whenever it wants, I'd bet that this might indeed translate to pretty incredible gains in power usage depending on _how_ you use it. Most of the time, a modern desktop os is blocking on the slowest peripheral
And if they really can change speed on the fly, think of the benefits for Java/JNI space processors. A constantly variable processor clock is a cool idea, as long as you can still keep absolute time
Re:Why should they keep the x86 compatibilty?! (Score:1)
Calm Down (Score:1)
[snip]
Now for the really OT bit. Can someone well-versed in the arcane ways of advanced HTML please tell me how ZD has bolloxed up their web page, so that I can't even scroll without a huge (1/2 second) lag between typing and response. (or mouse control, same problem) This is on any computer I've used, and through any bandwidth I've had access to. Why do they suck so much???
After a brief inspection of ZDNet's source, I can say definitely that there is nothing there that could be called "bolloxed up". Clean, no; but not screwed up, either. I can suggest several remedies if it's really bothering you:
HTH
Re:Simple question (Score:1)
IIRC the CPU was eating quite a big chunk, but the other parts (backlight/HD/io chips) eat not much less, and accumulated they are much more than the CPU.
Samba Information HQ
Whatever happened to just making an announcement (Score:1)
Now, it seems, we have to have an announcement announcing the date of the actual announcement. Wasn't Transmeta the same company that had to announce the date that they'd announce the date that they'd have a presentation announcing their new product?
Can't we just go back to acutal ANNOUNEMENTS?
Aye Carumba!
Why use Transmeta? (Score:1)
iBook pro/con (Score:1)
First, the support for the hardware is lacking, and probably a couple months away from being comparably to PC or older Mac notebooks (these problems apply to the latest-generation PowerBook as well). The Linux kernel doesn't support the power management chips in the iBook, and battery life suffers tremendously. Crucial patches to support the graphics hardware exist, but haven't been integrated into the upstream Xfree and kernel sources yet (actually, the latest 2.2.17pre and 2.4.0-test1-ac kernels may have the support in them). The coolest feature, AirPort, isn't supported at all yet, though able programmers are working on it.
Second, the iBook lacks a lot of critical features that a typical Linux user will probably expect from a computer. Even if you don't mind the builtin 800x600 screen (which admittedly is a very high-quality LCD), you'll be disappointed to find you can't add an external monitor. Upgradability in general is not good. This is mostly due to Apple's efforts to keep the iBook from eating into sales of their pricier PowerBook G3. By keeping features out of the iBook, they've created a notebook that'll meet the needs of many students and home users, while forcing computer professionals to stick with the higher-margin PowerBook.
On the other hand, I do recommend the PowerBook to Linux portable users who can afford it (not me!). Sure, the newest generation has some of the same issues with regard to kernel support, but those will be fixed soon enough, and this baby is a machine that you'll be keeping for a long time. AirPort and FireWire drivers are both in the works. All around, it's a good machine -- DVD, nice display, firewire, USB, 100 Mbit ethernet, (but kinda heavy). My friend runs Linux on his PB G3, and it's a really nice setup.
Re:Battery life (Score:1)
Baka not working extrans.... (NT) (Score:1)
Dorky look... (Score:1)
-----
Re:SMP?! (Score:1)
-----
your missing the point (Score:1)
The chip is a RISC chip, and the entire point of it is low power consumption! There's a software layer between it and the executables, so it can (in theory) run anything.
-----
Re:I won't buy one just yet. (Score:1)
Re:Do we only care because it's Linus? (Score:1)
Great. That's all I need: an optimized death in Quake...
Re:Simple question (Score:1)
Second that for iBooks. Only the packaging and screen size put me off in any way. Ah, and - if i get my memory of the demo right, I'm not filling up a PCMCIA card slot for wireless - the Airport is a mobo attachment.
Moreover PC cards tend to suck battery big time. I gather the IBM T20's have built in RJ45s for ethernet. 'Bout time. Can we have SCSI too (inbuilt) please? Poster in reply to my other thread in this story held out a good memory for a PowerPC TP with SCSI which ran nice and cool too.
IBM - build these features *in* and let me really have use of those TypeII / Type III slots *for expansion* (Like a 1gig type III hd from San Disk maybe :) )
==Idle Random Thoughts. Usual Disclaimers Apply==
Re:Simple question (Score:1)
Completely envious of long battery life. Bought extra battery for my TP600, erm £160 or maybe 220USD. Completely crap batterylife on these, and i really look after the batteries, fully cycling them, careful about partial discharge, recharge. Very envious of friends with Powerbooks which *do* give nearly 8 hours life (with dual battery set up).
So very needed, good batteries. Will drive wireless lan sales (no cables needed at home or office for anything). Will drive laptop sales - no more worrying about the nest outlet proximity on long haul journeys (I like trains in Europe, and economy gerenerally doesnt yet have nice facilities for this :( )
Remember the Mac Portable? Lead Acid batteries kept it running for ever. What tech is being used now? Might IBM have come up with something really fresh for a fall launch? Like the multi - gig micro- drives, only something ewveryone will really buy into? :-P
Lower power is nice but.... (Score:1)
moderation at work (Score:1)
Re:Neato... Now for the wearables. (Score:1)
"It's not a niche..." (Score:1)
It took me a while to sort through all those negatives.
But seriously, why is it that micronotebooks are popular in Japan, but not North America?
Re:Simple question (Score:1)
Re:Do we only care because it's Linus? (Score:1)
CodeMorphing and x86 Compatibility (Score:1)
except... (Score:1)
Re:not so hot thinkpad? yeah! (Score:1)
I used to have a TP 850 with a 603e at 133mhz or something. It was actually pretty cool in spite of not (apparently) running linux. Unfortunately, you had to run AIX, so I traded it for a BeBox. I don't have that anymore either, though. It just wasn't the most useful piece of hardware for me at the time.
In any case, that thing was pretty slick, and it had SCSI, which is always nice, if expensive. And, I might add, it stayed cooler than a pentium.
Re:Battery life (Score:1)
I actually like the dorky clamshell case (When is it going to come in Sanrio editions? The Red one with Hello Kitty on it, the smoke color with Ahiru No Pekkle...) but the LCD ix 800x600 and there's no vga port. This is what we in the industry call a pathetic oversight. They could be excused for only having an 800x600 LCD if they had some facility for an external monitor using something other than NTSC. They don't. Ack pfft.
Re:Native OSes for Crusoe (Score:1)
Once you've ported the compiler, I'd expect that the only really difficult part would be porting the kernel and supporting required devices. All nice programs (ie, programs written in C/C++ and not in assembly, or those not programmed for a specific hardware set) should port right over.
It would be some work initally, but once enough of the standard libraries that programs required where ported, all the user programs should just be a recompile away.
After all, there are Linux versions for Alpha, Sparc, PPC, ...
SMP?! (Score:1)
Re:Simple question (on battery life times) (Score:1)
Reason is because there are more things sucking power from a laptop besides the processor.
I imagine that a 12inch screen laptop with little frills and no pc-cards would last much longer than a 15inch screen laptop with all the frills and two pc-card slots cranked and working.
But of course, I expect the Transmeta-powered IBM laptops to beat candy-books on terms of battery life (not to mention your geek self-respect, lol).
Re:hmm (Score:1)
IMHO you are missing the point.
The mayority of laptop users use them to use "productivity apps", listen to MP3's and CDs, and look cool (not always in that order), all while they cruise on the "Economy Plus" part of a plane. For them, extended battery life is a major plus, since they dont need to do anything computationaly intensive.
Yey! (Score:1)
But why end there? We have yet to see a large-scale usage of the babies, but I'm hoping that it will happen. For now, 3Com and Nokia are switching from Mot chips to Arm chips. This is a step in the right direction, and maybe with some time, Transmeta chips will have their place on their products.
Re:Simple question (Score:1)
What uses the most power is anything that moves, conatins servos. That would be the hard disk, floppy, and especially the CDROM drive. The display has a backlight which sucks a fair bit so that you can see it in broad daylight.
On the other hand, when you feel around an Intel laptop, one of the warmest parts is just around the processor. Thus, wasted heat.
Conclusion: Less power consumption is not that significant, IMO.
Re:Why Waste a Good Processor with Windows? (Score:1)
Re:Why should they keep the x86 compatibilty?! (Score:1)
The reason behind Transmeta is:
Paul Allen
=> think Microsoft
=> problem with laptops (Intel & AMD incapable of delivering what the Windows users want - autonomy)
=> major investment in a chip that will run Windows and consumes less
=> make more money selling Windows as the ultimate mobile platform.
But x86 is 20 year old architecture! (Score:1)
Re:Battery life (Score:2)
The LCD may look nice, but it's still 800x600. That's the size of my current laptop, and it's too small. I'm not going to buy a new laptop that small.
Base RAM and HD space are too low (both below my current laptop. I can expand them, I assume, but that adds to the cost.
Wireless networking is cool... but I don't have a wireless LAN at home or at work, so I'd be paying for something that I'm not using. 8^(
It doesn't have SCSI or PC card slots, so I'd have to buy something else to use my scanner. (Are there SCSI-over-USB things? Me USB-illiterate.)
Finally, it's a PPC, so I'm cut out of CorelDraw. (And WPO2k, but that's a small loss.)
That's giving up a whole lot just for an extra 3 hours of battery life. It's great if it works for you or you need the battery life, but I think I'll wait for a Crusoe-based laptop, since that will fit my needs better.
Re:Yeah, But Where's Friday? (Score:2)
Well, yes. I don't need more speed. Nothing I do is excessively slow right now (and that's on a P200). Sure, I can't run most games, but most laptops can't run most games.
But being able to run for really long times while keeping most (or almost all) the performance of being plugged in is really, really exciting!
Dynamic recompilation is cool technology (Score:2)
Transmeta has what is truly an interesting and novel approach - they're marrying the by now well-proven dynamic recompilation technology that powers Executor, PSEmuPro, UltraHLE, and many other game and computer emulators with a CPU that can uniquely support it, use very little power, and scale very quickly since there's no serious complexity.
Beyond the obvious benefits, this means that the coolness of a processor can be decoupled from "yeah, but does it run Office?". You can now build a processor with all the geek-coolness of the PowerPC and Alpha, and actually sell it
Crusoe native OS - doubtful. (Score:2)
Probably not. Transmeta stands to benefit by NOT releasing the Crusoe instruction set for general consumption, and it also makes technological sense. Here's why:
Crusoe does it's morphing in software and translates the results into native instructions. This is all done in software, and results in lower power consumption and slower performance.
Transmeta could possibly develop a series of processors aimed at the desktop, where power consumption is less important than speed - they could move the morphing logic onto the chip. Check out the Transmeta write-up on ArsTechnica for deep info.
If things were chip-native, you'd have a natural fork in the road, applications and the OS would be written for the desktop hardware or the portable hardware, but not both. By keeping the native instructions close to their vest, Transmeta is maintaining platform integrity. The counter-example is Windows with it's NT on the desktop and WinCe on portables.. What a mess!
Transmeta can easily support other architectures this way as well, not just x86 but maybe i386 or 68k and PPC as well, without worry about supporting a 'portable' or 'professional' (buzzword alert) version of each. Hell, maybe we'll be lucky enough to see a Crusoe JVM as well, but I for one hope that there will not be native Crusoe apps, ever.
Besides, they're a chip company, not an OS company, regardless of what Linus does in his spare time.
Neato... Now for the wearables. (Score:2)
Re:Simple question (Score:2)
engineers never lie; we just approximate the truth.
Re:Specs? (Score:2)
the clock speeds aren't as fast as an athlon or even a pIII, because transmeta chips aren't being made that fast yet.
>I wonder if all the OEM hardware will also be linux friendly
IBM has sworn to make all of their notepads linux freindly so I have to assume that the thinkpads will be linux ready from the get-go. But NEC has never shown any interest in checking their hardware for compatibility with Linux, but the versa series is renowned for being linux-ready. I just don't know.
Devil Ducky
Re:Why Waste a Good Processor with Windows? (Score:2)
You'd be surprised what that little PalmOS is still capable of. Especially in capable hands [handpsring.com].
Re:Why should they keep the x86 compatibilty?! (Score:2)
I'm looking at AMD, then I'm looking at Intel, and them I'm looking at your post and I'm thinkin, 'Someone isn't paying attention to the real world.' AMD is doing just fine and and isn't in any danger of being crushed by Intel. If anything it will be the other way around. Transmeta is going to be succesful because they are offering a VERY low power chip that will run linux and windows in a laptop without any compatibility issues. The key being a LOW POWER chip. They'll be able to use the profits from this to fuel the development of even lower power, faster, and more versatile software to run with their processor. Transmeta is going to succede on its own merit, despite Intel and AMD, because it produces something that neither of those chipmakers has managed to produce.
Kintanon
Transmeta not on PC Expo 2000 List of Exhibitors (Score:2)
That's half true (Score:2)
<p>Yes, they were really disappointed with the Windows benchmarks on the original silicon (TM3200). It turns out that they optimized for 32 bit code, because everything is 32 bit, or so they thought. It turns out that windows 9x still uses a TON of old 16 bit code left over from Win31. So, they had to deoptimize their chip. That's why the 3200 is listed as optimized for Linux, and the 5400 is optimized for Windows.
<p>I imagine they hired Linus for 1) the prestige of hiring Linus, and 2) he's a damn good coder (not necessarily in that order).
<p>I also imagine that they will keep the two chip sets seperate, so the 5xxx series will ontinue to be optimized for 16 bit apps (Windows 9x and ME), while the 3xxx series will be optimized for everything else.
<p>And I'll be impressed with the two watt Celeron if I actually see it sometime before 2003 (I can't believe I'm saying that about Intel, but it's true - they haven't been able to meet anything resembling their demand the last 6 months).
Re:Native OSes for Crusoe (Score:2)
OT: Next generation? (Score:2)
Hmmmm....
Now for the really OT bit. Can someone well-versed in the arcane ways of advanced HTML please tell me how ZD has bolloxed up their web page, so that I can't even scroll without a huge (1/2 second) lag between typing and response. (or mouse control, same problem) This is on any computer I've used, and through any bandwidth I've had access to. Why do they suck so much???
Re:x86 code? (Score:2)
If my memory is accurate, Linus was hired because of his knowledge of the x86 architecture. Linux became big while he was there. I read in a Linux journal interview with Linus that one thing he liked about Transmeta was they let him work on Linux. I'm pretty sure Linux compatibility was one of the tasks Linus worked on.
Matthew Miller, [50megs.com]
not so hot thinkpad? yeah! (Score:2)
"Cool" is the operative word here. Read over at The Register that Intel demoed their "2 watt" mobile chips but still spec for mobo manufacturers to allow for 17.8W consumption!!! (hope that's the right number)
I've been well pleased at the idea of a proper TP running linux for a while - and though hard about preordering for our office. But last night, bashing hard on the keyboard for hours on end on my own model, I had to keep lifting the TP off my lap 'cause the darn thing gets to like 90f outside case. Solve that - i.e. give me Transmeta - you got sales.
This is darn cool - but (maybe i'm dumb here) will an OS ever be native for Transmeta Crusoe, or will everything run in code morphing, basically old x86 code?
Remembering the PowerPC TP's IBM did for like all of a week or so (did anyone buy one?) I'd surely love to be able to recompile my apps for Crusoe - or *anything* other than x86 and be able to take them with me :)
==Random idle thoughts, usual dislaimers attributable thereto.==
Re:Why should they keep the x86 compatibilty?! (Score:2)
Okay, I assume you haven't seen this [arstechnica.com] article. So, let me give you the long and short...
Unless you want to build an entire platform from the ground up, including OS, apps, and all, you have to stick with a legacy isa. (In this case, x86.)
So, basically, it's not just a marketing stratagy, it's common sense.
Re:Lower power is nice but.... (Score:2)
They already have. Here [newscientist.com] is an article about a flexible display with low power requirements that's inexpensive to produce. Supposed to be shopped to manufacturers this summer.
carlos
Re:will it run any x86 OS (Score:2)
--Josh
Re:Why Waste a Good Processor with Windows? (Score:2)
I do in fact own a Handspring Visor. It's blue. If this one breaks, I think I'll get a green one. In any case, the mp3 playing handspring module has a couple major deficiencies. One of them is that is uses CF memory, which is expensive; Not that the visor has enough storage to hold mp3s, eithere. The other is that it bulges out of the visor. I could forgive the former, but not the latter.
Heck, the digital camera for the visor is smaller than the mp3 player. Doesn't it just seem that there's something basically wrong with that?
Re:Why should they keep the x86 compatibilty?! (Score:2)
I don't think the chip has been the big stumbling block to running linux on laptops. Someone correct me if I'm wrong, but I was under the impression it was all the kooky support hardware that goes into them that was the problem. Everyone seems to have their own standards, though the PCI bus coming to laptops seems to have improved that considerably.
As for low power, intel is bringing out their own lower power CPUs, including a Celeron 500 under two watts.
I heard a story (no word on whether it's true, or just a rumor, so don't go saying that I told you anything for sure) that Transmeta had actually been hoping for better benchmarks coming back from their initial silicon, but that they had ended up being slower than expected, but they still had the low power consumption, and a marketing campaign (one part Linus Torvalds, one part Low Power Consumption) was born. Who knows if it's true, not that it really matters.
What I think would be interesting here would be if after there was a substantial installed base of transmeta chips, transmeta announced that the code for a new instruction set was available, and that linux had already been ported to it, and that it "didn't suck" in the way that x86 chips do. You could download the code, install linux, and bask in the new speeds you'd be experiencing, and it would be a hell of a booster to transmeta.
Re:I won't buy one just yet. (Score:2)
No matter how hard they try to make that code unchangeable someone's going to get around it. For an example, look at how many different programs got around the Intel ID option that was supposedly impossible to change by a rogue program.
Anyway that was my thought and mine alone
Re:iBook (Score:2)
Yes, but the graphite model costs an extra $200 for a minor speed bump and the classier color. It's not quite as good of a bargain (... unless you are a junior high kid. Then it might be worth $200 to not get the crap beat out of you on the first day of school.)
The Powerbook is also a lot of ! for the $, but in the sub-$2k basic-needs category, the iBook is a hell of a deal.
Also, if I want a high-res monitor, I will use my desktop system.
Laptops are road machines, and if you intend use one to replace your desktop system entirely, nothing under 2 grand is likely to meet your needs. I still say for a lot of portable power, good battery life, and low price, the iBook is hard to beat.
As for the complaints of memory and HD... not many PC laptops at the iBooks price level come with more, and those that do cut other corners that you will regret (such as untra-cheap screens, short-lived batteries, bottom-of-the-line K6 chips, etc.) You can always add more later... IIRC it uses P66 SO-DIMM's, and the HD is just a mini IDE drive.
Re:iBook pro/con (Score:2)
I guess this is a big deal to Gnome users (all XWindows environments suck at 800x600), but to a BASHer like me, monitor resolution is a very minor concern... when I want purdy pictures or Netscape, I will probably just boot to the Mac partition. IMHO, Gnome and KDE are Not Ready Yet anyway.
(Just opinion folks... tell the GNU police there is no need for the flamethrowers.)
Yeah, But Where's Friday? (Score:2)
In fact, he could have simply walked to civilization, but that wouldn't have been such a good story. I find this to be a decent analogy to what's going on at transmeta.
I must ask:
Where the heck is the generic Evaluation board? I keep going over to the website, waiting for the Transmeta development stuff [transmeta.com] to become available, along with the generic $300 internet appliance development board, or the wearable computer development board, or even the somewhat risqué Gal Friday Personal Companion development board, which is being eagerly awaited by the guys over at Real Doll [realdoll.com]. I know a lot of technical geeks who would much rather spend several years hacking hardware and hydraulics and software to make a girlfriend who will put out after a romantic evening of watching deep space nine and listening to long winded rants about how stupid their boss is.
(I got married early, so I skipped the need to learn hydraulics...)
Seriously, though:
Is it just me, or has anybody else noticed that the draw for transmeta seems to be variable power draw more than any inherent improvements in speed?
I want my developer tools! (and yes, I understand that keeping the tool kits hidden may slow down the competition - but I want to be the first kid on my block with a transmeta pda... which I built myself.
I won't buy one just yet. (Score:2)
I would be worried about purchasing one of the early models of any of Transmeta's new technology, since the reliability of their CPUs and other associated components has yet to be proven. Even huge companies with decades worth of experience in chip design and manufacture such as Intel have been known to have teething problems with new designs. Considering Transmeta's comparative lack of experience, the complexity and innovation in their product and the fact that the Crusoe chips are not an incremental improvement on an already established technology, there is a good chance that there will be some uknown bugs lurking in the designs.
Although buying state of the art technology is always nice and I have faith in the abilities of Linus and his colleagues, I would rather not risk my money until the technology has been thoroughly tested by real users, out in the real world.
Re:Native OSes for Crusoe (Score:2)
Scalded? Check out Intel's new 1GHz "Hotpants" mobile processor...
Back on topic, I'm not sure that "going native" is the right mindset here. Doesn't the code morph allow for runtime code optimization? I thought there was something in /. a week ago that mentioned HP's immplementation of the same idea gave a 20% speedup over native execution.
Simple question (Score:3)
What is the battery lifetime of a Transmeta laptop?
Everyone says they will have longer battery life, but no one says how much longer. How important is that 3 watt processor after you factor in the disk, chipset, and screen backlight?
--Bob
Re:Native OSes for Crusoe (Score:3)
You seem to be out of touch with the trends in cpu design.
Here's [arstechnica.com] a nice simple intro -- it may have been a
Basically, you don't want to code to the bare crusoe metal since that will keep transmeta from improving the architecture without breaking your software.
The translation of x86 instructions can be done better than transmeta is currently doing it. By only exposing the x86 layer, transmeta gains the ability to totally redesign their chip without breaking any code. This is more important in the long run than a small speed increase.
Also, the crusoe is optimized for translating x86. There's no guarantee that coding on the bare metal would be an improvement.
--Shoeboy
(former microserf)
Re:x86 code? (Score:3)
There is a version of Linux, "mobile Linux" which was created by Linus basically to enable it to work better in PDAs but AFAIK, it was not specifically optimized for the Crusoe per se, just for PDA setups in general.
-JD
Wasted heat!!?? (Score:3)
nuclear cia fbi spy password code encrypt president bomb
RE: SpeedStep (Score:3)
However, Transmeta creates a chip that runs at 700 HMz (which apparently works out to the same speed as an Intel 500 MHz chip), runs at as little as one watt, and only uses as much of the chip as it needs? DAMN!
You have to remember that Transmeta is shooting for a different core market than Intel or AMD. They are aimed at people who want decent performance and tons of battery life. They don't want really kick ass performance - if they wanted that they'd but a desktop and let Intel or AMD throw transistors at the problem.
Sure, the technology is new. But don't whine because you don't really understand it. Go back under your bridge and make your horse whips while the rest of us putt around on our horseless carriages (okay, that last bit was out of line, but it's apt enough).
Why Waste a Good Processor with Windows? (Score:3)
Notebooks based on the TM 5400 chip and Microsoft Corp.'s Windows operating system are expected to begin shipping in North America and Japan in the second half of the year.
Windows is a resource hog. After Tranmeta has spent such a long time developing good code for its processors to save power and, from what I hear, make the chip faster than the current i386 architecture, why put a resource hog on a machine with such potential.
Take a look at the Palm top computers. They're running on 2 AAA and the PalmOS and the new Linux palm (I really have to trade-up for one of those!) and even the color palms last longer than Windows CE palm-tops. You would think companies like NEC would have realized this and put a better operating system on, but I guess it's still a Windows world out there.
Hopefully, though, it won't take long until we start seeing Transmeta-based Linux boxes, hopefully sooner than it took to see them on Intel-based boxes like we have seen recently with Dell and what not.
Re:Why Waste a Good Processor with Windows? (Score:3)
In case you didn't notice, the WinCE boxes are in the 100-200mhz range, which makes them considerably faster than a 20mhz palm which uses a butt-slow (but cool nonetheless) dragonball CPU in most cases. A quick recap for those who have forgotten: Dragonball is a Motorola MC68000 with LCD Control and a serial port (or some hardware convenient to interface to one?) built in.
This has the effect of making the Windows CE platform devices consuming much power, it's true. Heck, none of my amiga computers even had a heat sink NEAR the CPU, and those were predominantly 68000-based. This is a much smaller die/chip, but it's also a much finer process, requiring even less power, thereby generating less heat, etc etc.
Windows CE platforms consume more power because they use faster CPUs. This lets you do things you can't do on palm, like play mp3s. It's a tradeoff. While I agree that PalmOS is a better platform than Windows CE, it's not for the same reason you cite... And you could make a slower, lower power consumption Windows CE box, and it would probably run okay, but a lot of the stuff on wince (ha ha) is more intensive than the stuff done on palm, because the processing power has been available.
Re:Do we only care because it's Linus? (Score:3)
Before Transmeta existed, before speedstep was conceived of by intel, Motorola had CPUs which had a sliding speed scale. Mac powerbooks contained them.
Native OSes for Crusoe (Score:4)
Of course, you'd have the trouble of rebuilding all of the user land programs, too, but with the amount of work people put into things in the open source world, it wouldn't surprise me to see a Red Hat Linux 6.2 Crusoe Edition.
Like you, I'm tired of my laptop getting hot enough to scald my skin. It's time for laptops which run at a reasonable temperature.
Re:I won't buy one just yet. (Score:4)
If there is a problem with the chip ( somthing like the FDIV bug of intel ) they can distribute a hot patch for the interpreting code of the chip to work around the bug. This would give them time to fixt the bug and replace the defective parts. It would also keep the customer partially happy due to the fact that the bug is no longer a hinderance; now they only have a performance hit that, depending on the problem, may not be noticable.
Do we only care because it's Linus? (Score:4)
Personally, I wouldn't mind seeing Transmeta fail. Nothing I have seen in their chip design actually suggests any kind of power to back up the battery life. Their handling of x86 code seems suspect, and the underlying design of the core instructions seem to show no speed benefits (e.g. like the Velocity engine of G4's, or MMX and SSID instructions in Pentium III's).
I say let them try, but take their product introductions with a grain of salt. I'm no conservative, but I prefer an established hardware infrastructure that works, over an unestablished one that sounds flaky (and we are only monitoring because some guy named "Linus" is involved).
And what shall be unveiled..... (Score:5)
The skies had grown black around the Transmeta towers. Everyone knew they were up to no good, so to speak, but no one knew the unspeakable horrors they intended to release upon the world. The only man who understood, and who knew the full ramifications of their actions was Linus himself, and he was in no position to do anything. But, he thought, I know who can.
He muttered the incantation quietly, so as not to trigger the tape recorder set to record whenever someone spoke in the room. Such precautions seemed unnecessary at first, but Linus had not anticipated the true nature of Transmeta when he had taken the job. Now he spent days translating arcane texts that would drive most men mad, and coding the software to drive the most diabolical machine man had ever dared to construct. It was by only the vaguest definitions a computer, driven by processors in a beowulf-style cluster. But there were no disks, nor any traditional interfaces. The processors drove masses of flesh and neurons, specifically designed with one purpose in mind. To summon onto the earth the most concentrated forms of evil known in the universe, and to attempt to bind them to Transmeta's will. However, Linus, having read most of the occult texts, knew this could never be, and the summoning alone would be enough to end mankind's existance as they know it.
The air crackled with the energy from Linus' spell. A warm, muggy feeling took over the sterile cold basement chambers. He was coming.
Suddenly, a vague shape took form in the corner of the room. Two eyes formed in the center of the wavering mass. In a low gutteral voice, it intoned "why have you brought me, Linesus of Linuux", using Linus' true name.
"you must stop Transmeta before they unveil their plan" Linus pleaded, knowing the great being's power would prevent the tape recorder from kicking on.
"I understand your problem, and I fully agree that this evil must not be unleashed. However, I require two things."
"ANYTHING" Linus cried.
"for my colleague, I require a certain young Portman," gnarphlager slowly stated, "and for myself, I require cheese . .
Re:Simple question (Score:5)
Those frustrations are weight and battery life, he said. "Frustration studies" done by IBM have shown that users want a notebook weighing about 3 pounds, with eight hours or more battery life, he said.
"If we can do that, we'll bring it out in the fall," Suarez said. "We're pretty confident that we can get close to that eight-hour mark."
Re:Do we only care because it's Linus? (Score:5)
I agree that a lot of the fuss here has come about because Linus is involved in Transmeta. But there's a lot more to their technology than Intel's speedstep.
Processors with multiple execution units have become very common. They work by looking for instruction level parallelism - in short, if 2 instructions use totally different registers, whose operands have already been computed, then you can execute them in parallel.
Modern processors tend to support out of order execution and use branch prediction to avoid wasting clock cycles. Both of these, combined with several execution units, make your processor very fast - but use a huge amount of silicon area (read: lots of power, space that could otherwise be used for cache, etc). Their other big problem is that they look at the program from scratch each time it is executed - which means they can't avoid old mistakes.
The idea that Transmeta had is as revolutionary as the early RISC philosophy. The original observation that lead to RISC was that compilers don't tend to make use of the more complicated CISC instructions (string operations, polynomial evaluation on VAX, etc) and that these operations require a lot of hardware to implement, and as such slow down the processor. The RISC goal was to use a small number of relatively simple instructions. This means that you need fewer transistors on the processor, which means that it uses less power. Having fewer transistors generally makes the processors critical path shorter as well, which allows it to be clocked faster. In addition, the simple RISC instructions allow easy pipelining - which in itself leads to a huge speed benefit.
However, modern RISC processors are much more complex than their ancestors. They support many more instructions, and implement things like dynamic instruction scheduling. This has lead to RISC chips using more and more silicon.
Transmeta's idea is to go through the same process of reducing the complexity of the processor once more. They save a lot of silicon by shifting work onto the compiler.
If you think about it, this makes a lot of sense. The compiler is responsible for allocating registers, so it already knows about instruction level parallelism, and therefore it knows which instructions can be executed in parallel. If you let the compiler tell you this, rather than working it out in hardware every time you run your program, you can potentially save a lot of wasted silicon.
Transmeta went for code morphing because it saves people having to re-compile code to run on their processor. They gain the benefits describes above, without the huge cost of trying to replace the x86 instruction set. Their code-morphing engine effectively translates x86 (ar any other architecture they want to implement) into native instructions. It finds instruction level parallelism, and exploits this in native code (the Transmeta can execute 4 32-bit instructions simultaneously - although you are limited in what operations can be caried out at the same time).
By having the compiler do all of the work in scheduling instructions, they allow a huge amount of silicon to be saved - therefore reducing power consumption and allowing faster clock speeds.
There are more benefits. The code morphing engine can heavily optimise frequently used blocks of code. Potentially, they could store information about the behaviour of a jump instruction in their code, to allow them to do much better branch prediction. If you've seen the program execute 500 times, you've got a pretty good idea how it's going to behave next time in a lot of cases.
The point is that the x86 architecture has been a problem for a long time. It's old, slow, and difficult to optimise. What Transmeta have done is to find a nice, neat way to eliminate the problems caused by old ISAs. Effectively, they use a 2 stage compilation. Sun do this with Java - compile code to Java Binaries (platform independant) and then use a native Virtual Machine to run these. Transmeta are applying a similar concept, in a radical new way to their processor. The x86 instruction set is being used as the platform independant middle stage in their model. It gets compiled and optimised to native code by their code morphing engine. They've also been able to save a lot of silicon in the process.
And, since their code morphing engine knows a lot more about the machine it's running on (cache size / associativity, memory size, etc) it can probably do a better job of compiling code than most x86 compilers can. Gcc and the like are good - but if they have to support machines with many memory sizes, configurations, and know nothing about caches and the like, then there's only so much they can do.
Make no mistake, I believe that Transmeta has revolutionary ideas that will change the face of computing. No longer will legacy hardware be a problem for chip designers. Using the hybrid approach developed by Transmeta will allow faster and faster processor designs, taking full advantage of modern ideas, whilst still supporting legacy ISAs. It's a fantastic concept, and I hope they go all the way with it.
Battery life (Score:5)
Sure, it's butt-ugly (white with your choice of blue or orange), big (11.6" x 13.5" x 2.1"), and weighs a lot (6.6 pounds with the battery), but it's...
fairly cheap ($1599, last time I checked),
fairly fast (G3 chip w/ full cache... much faster than most PIII notebooks, which are crippled to reduce heat),
has a really nice-looking LCD screen,
has the cheapest wireless networking option of anybody out there,
and has a battery that lasts 5-6 hours (ymmv).
Best of all, thanks to LinuxPPC [linuxppc.org], they are great Linux portables.
No PCMCIA, but most of the stuff that you would use cards for (Ethernet, wireless networking, modem, USB) are already built in, so that's no great loss for most users.
Of course, if you can't get past how goofy it looks, maybe the Transmetta laptops are worth the wait.
Either way, I would hate to pay much for a Pentium-based laptop. They are slow, hot, and suck power too quickly.