Slashdot is powered by your submissions, so send in your scoop


Forgot your password?

C'T visits Transmeta 56

The german technical magazine C'T recently visited Transmeta's office in Santa Clara. Transmeta's roster is impressive, including not only Linus but also Robert Collins and Christian Ludloff both well known for their work on finding undocumented instructions and registers. Transmeta's LongRun technology (reducing CPU power by varying frequency and voltage) only works with APM, and without it the TM5400's net consumption is 5W at 43 degres Celsius and 600Mhz. At 700Mhz the TM5400's performance is slightly under that of a 500Mhz PIII. The TM5400 will be the first processor to use IBM's new CMOS8S copper process. In the interview David Ditzel denies having used Elbrus technology in Crusoe. For non german speakers, there's always Babelfish.
This discussion has been archived. No new comments can be posted.

C'T visits Transmeta

Comments Filter:
  • Intel and AMD haven't exactly been "racing towards having the first chip out on copper". Intel bowed out of the race well over a year ago when it said that it had no plans to use copper interconnect on it's 0.18um process and may consider it for 0.13um. AMD, on the other hand, has said that they will be shipping their 0.18um process with copper interconnect before the end of the Q2 (IIRC).

    The megahertz race has been going on since Q4 last year and is still heating up, but the copper race never really happened since Intel doesn't feel that it needs Cu in 0.18um.

    * Speaking only for myself *
  • The actual performance difference between a 500 Mhz pentium and a 700 Mhz pentium is much less than you think, in most situations. Having something that gets you 85% of the performance, at half the power consumption and heat dissipation is a BIG win for a laptop.

  • CT visits Transmeta

    You mean, CmdrTaco visited Transmeta? Cool! But why in the world did he write about it in German???

    Oh, never mind...

    Bravery, Kindness, Clarity, Honesty, Compassion, Generosity

  • Hmm. It mentions experimentation with a native Crusoe Linux, but the emulated x86 version was better. One case where a native port may not be worth the trouble (particularly with the code differences between the chips).
  • And Crusoe has four times the battery life of any of them.

    You do not understand. The point of this chip is NOT HIGH PERFORMANCE. Nor is it low cost. It is ACCEPTABLE performance, with ACCEPTABLE battery life--which in my and most people's opinions, no mobile devices have reached yet.

    Even at their most "efficient" setting, the SpeedStep PIII chips use as much power as a 400MHz PIII; only when they're plugged into a wall are they any faster. By contrast, Crusoe should be able to switch between a much much wider range of power-usage modes (perhaps even continuous range), and will be considerably more efficient at all of them. Furthermore, its response-time will be faster than most notebooks, which quickly power-down and back up in order to save battery life already.

    So. Once again. It's not for desktops. It's not for Quake3 and it's *certainly* not for Ultima 9. It's for non-3d-game Win32 compatability--mainly web browsing--on ultralight laptops with several times the battery life of today's x86 ultralightweights, and for wireless x86-compatible internet webpads with 20 hour battery life. Your comments are completely irrelevent to the market Crusoe is targeted at.
  • Crusoes island

    On one Friday with Transmeta

    Reports on a journey in c't are rather rare. In the latter, which I can remember, a certain Rob. S. Pierre described a press journey from Apple to Ireland [1] - and whirled up much dust thereby. It struck me following a Intel Pressereise (to the IDF) after Santa Clara, not to Intels Headquarters, but to Transmeta.

    Hardly a stone throw away of Intels headquarters the start UP company Transmeta in a beautiful park logiert distributed on three in to two-story houses. From the lying close road with the defining name ' Freedom Circle ' one looks directly on the relatively ugly large concrete/glass blocks of the ' large brother '.

    Instead of in open-plan offices with many ' Cubicles ' those sit altogether more than 200 Transmeta coworkers mostly in small offices, although there are also some larger laboratories. There I am a first European journalist in its holy halls, insured one me, and above all all all this, who was allowed to spy a whole day in the offices and laboratories.

    In Linus Torvalds ' office I felt directly as at home. Torvalds - style genuine in the SuSE t Shirt - acknowledged that he had begun with Transmeta, because he finally times which other one to make wanted than in each case Linux, Linux. It did not express it, it suggested however nevertheless that it Linux had out-hung in the meantime quite to the neck. And there the function was to participate as a chief architect in the code Morphing software (CMS) a welcome alternation and a large challenge.

    But the spirit, which one called... to time sits Torvalds nevertheless again mainly at Linux, to ' mobile Linux ', which is meant as accessories to the Crusoe processor for small harddiskless systems (in 32 MByte the Flash).

    Mobile Linux runs by the way as x86-Software and is dynamically gemorpht. With a native Linux version one - in such a way I experienced besides - had experimented also, but proved as fewer effective than the emulated x86-Version. Transmeta does not also even want to publish the complex native instruction set of the Crusoe processors, because the company can keep itself so any modifications open. There are already differences between the CMS versions of the two Crusoe processors to TM3120 and TM5400.


    Of the effect of the current savings technique LongRun also for the virtual Northbridge the responsible person development conductor Marc meat man, by the way a German, could convince me in the Transmeta lab. Between 266 and 600 mc/s in five steps (the minimum gradation is 33 mc/s) and between 1,1 and 1,6 V of regulating TM5400-Crusoe also times 6 W used, remained on average at usual software however with approximately 1 W and became - without radiator boxes - even times lukewarm within the point area. LongRun needs Idle statuses for its rule work. Without APM under DOS LongRun does not operate and the 600-MHz-TM5400 swallows then, approximately at the Doom play, constantly 5 W, which heated the processor in addition, only on 43 C. Into these values consumed is the compatible Northbridge with included, partial to Intels BX, which is integrated already with the Crusoe in the processor or is virtualisiert of the CMS. To the comparison: the original BX Northbridge alone used up well 2 W - and knows neither PC133 nor a second DRAM INTERFACE for DDR-266 like the TM5400.

    Since CMS is also gentleman over the Northbridge, it can drive not only the processor, but also the memory if necessary more slowly, which supplies additional Einsparpotenzial. Smaller, discovered bug in the Northbridge software just was quite useful, in order to demonstrate directly times the advantages of a processor implemented to a large extent in software: Before my eyes software Engineer Peter Anvin reloaded during operation a corrected CMS version, wars.

    For the Crusoe BIOS Robert Collins is responsible, who came unexpectedly together with many other coworkers from before four years the adjusted, secret Pentium project of Texas Instruments to Transmeta and with its well-known Website and Intthe el Secrets had created itself already times with Chipzilla. Collins described me that Crusoe loads those with boats first into few 100 milliseconds about 2 MByte (decompressed) large CMS, which translation Buffer (8 to 14 MByte) creates and then, as each tidy x86, with which execution of the code at the address F000:fff0 begins. The BIOS is coded complete in x86-Code and transfers system information like the SDRAM parameters picked out from the SPD EEPROM over a common storage area to the CMS software. For chip-record and Transmeta the Phoenix BIOS has processor-independent section licensed.


    Likewise via TI to Transmeta the ' x86-Validator ' Christian Ludloff came, on which TI had at that time become attentive by its c't article overPentium secrets[ 2, 3 ]. Ludloff typed itself with millions of assembler lines in the wahrsten sense of the word the fingers wund, in order to program ugly, malicious, hinterfotzigen code, only in such a way over-accumulated the poor processors up with Stolperfallen. Thus he discovered zuhauf bug and anomalies not only with earlier Crusoe/CMS versions, but naturally also with the competitors. So some entry for instance in Intels Specification update would have to actually carry the thank saying ' thanks for to of CL '. Ludloff plans to make in the summer a part of its assembler often commodity, for example all processor structures, exceptions and so on Website very popular on its with Insidern accessible the public - that becomes bitebite bites!

    I with the occasionally expressed assumption, he confronted Transmeta boss Dave Ditzel had designt himself during his time as a Sun coworker in Moscow ' enriched ' at the ideas of Boris Babaian, the E2K names with the Russian Sun Distributor Elbrus already for a long time a processor. Babaian was even one week before with it, answered Ditzel. One conversed nicely. The Designs of the E2K (High end VLIW without dynamic optimization, very high floating decimal point performance) and Crusoe are nevertheless very different. A certain inspiration already gave it, but to no more than from HP -, Sun or other Designs.


    Ditzel contradicted also publications, according to which in April products with Crusoe processors are to come out. At the earliest end of the second quarter, thus probably in June, should be so far it for the small Crusoe TM3120. The more efficient brother TM5400 with the ingenious LongRun Stromspartechnik is in the third quarter to expect. It will be the first processor on the market, which is manufactured in IBMs new CMOS8S-Kupfer-Verfahren.

    When possible oem comes beside S3/Diamond and FIC also Quanta computer in question, one the largest Notebook manufacturer, that manufactures for many different in the job, which draufdrucken then their Dell, HP or other Logo. Sybase had as large software house already admits given that its SQL Anywhere studio (mobile data bases and data synchronisation) was to be extended particularly for Crusoe systems. And the Augsburger company Infomatec AG announced that they were received together with the Crosstainment AG and Transmeta a partnership, in order to develop on base of its Java network Technology (JNT) together reference models of Internet-capable terminals with Crusoe processors.

    With absolute performance specification Transmeta is further reserved, the values do not look so badly. A 700-MHz-TM5400 comes for instance on Pentium-III-500-Niveau. But even with the so important Ziff-Davis bench mark Winstone lies under it it a piece, which results according to Transmeta from the fact that these bench marks only once ' touches many routines ', so that the dynamic optimization cannot access. The ZD lab saw and wants the lack however in the discussion with Ditzel in the next version for a realistic frequency distribution to provide.

    And finally there was the most frequently placed question there whether now Transmeta goes still in this year to the stock exchange? Ditzel answered much saying: ' it would not surprise me... ' (as)


    [1] Rob S. Pierre, Pferde, Apple and the Irish Macs, c't 10/87, P. 82

    [2] Christian Ludloff, wondrous transformation, 2/95, P. 242 c't

    [3] Christian Ludloff, between the lines, c't 11/94, P. 266

  • Intel and AMD have been racing towards having the first chip out on last count, it looks like AMD is going to beat them to it. Is that what you're thinking of?
  • Sigh. (I know it's bad to complain about something free, but) It's tough trying to figure out a sentence like:

    Of the effect of the current savings technique LongRun also for the virtual Northbridge the responsible person development conductor Marc meat man, by the way a German, could convince me in the Transmeta lab.

    Wish I knew German, if only to see how babelfish ended up with "Marc meat man."

  • Thanks! Both to you and the guy below. The Fleischman - "meat man" thing is cracking me up.
  • by frederik ( 86671 )
    I've posted this story over a week ago :(.
  • IBM developed the copper interconnect technology...
    however there is currently no
    commercially available chip that uses it...
    I believe they're saying the Crusoe will be
    the first commercially available chip using
    the technology...
  • When I first saw this article, my first thought was that Rob had been invited out for a visit with Linus, et al. As I read further, it became apparent I had the wrong CT. Reading a little more thoroughly, all I can say is ... Doh!
  • Quick question: What is APM? Thanks!
  • Von der Wirkung der Stromspartechnik LongRun konnte mich der auch für die virtuelle Northbridge verantwortliche Entwicklungsleiter Marc Fleischmann, übrigens ein Deutscher, im Transmeta-Lab überzeugen.

    You mean this? Why use bablefish when there is the instant /.-translationcrew at your disposal? Al you need to do is ask...

    I just hope I translate this correctly (I'm Dutch). So, here it goes:

    Developmentleader Marc Fleischman, a German btw, who's also responsible for the virtual Northbridge, could convince me of the effect of the powersavingstechnology 'LongRun' at the Transmeta-Lab.
  • 1. The price of a Celeron 500 and motherboard is about $160 (see For $350, you can get an Athlon 700 and motherboard, which for most applications will run nearly twice as fast. This difference isn't much considering the cost of the rest of the PC. If you are buying a $160 chip and board, you aren't buying entry-level; you are buying junk. Sure the most of the current generation of software doesn't need this kind of performance, but you can bet the next generation of apps running on Windows Millenium will hog it right up. Already you need more than a Celeron to get really good performance out of games like Ulitma 9 and Quake 3.

    2. AMD is phasing out the K6-2 for the desktop market, and the mobile market will soon follow. They will soon introduce a mobile socket version of the Athlon. Intel is also moving ahead quickly with its SpeedStep PIII chips. Consider that in the beginning of 2001, the standard for computing will be chips well in excess of 1GHZ, and it would be silly to buy something in a few months that lags behind today's PIII 500 (and likely fares much worse against the Athlon in floating point operations.)

  • Well, There may always be a bablefish, but its usefulness will always be what we've learned to expect from babel. I use Babel to translate SAP commentary, and it's useful. But weird.

    An example from the quoted site is

    In Linus Torvalds ' office I felt directly as at home. Torvalds - style genuine in the SuSE t Shirt - acknowledged that he had begun with Transmeta, because he finally times which other one to make wanted than in each case Linux, Linux. It did not express it, it suggested however nevertheless that it Linux had out-hung in the meantime quite to the neck. And there the function was to participate as a chief architect in the code Morphing software (CMS) a welcome alternation and a large challenge.

  • Just say no to Babelfish.

    I keep waiting for someone to do it better. We now have the Systran engine in both altavista and, both of them willing to mangle anything you throw at them for free.

    I notice that The Register keeps quoting L&H translations, but I'm too cheap to pay for one. Is there any non-systrans site on the web?

  • For all of you who have some dollars to spend: The last sentence is very interesting:
    When asked if there might be an IPO this year, Ditzel answered:
    `Es würde mich nicht überraschen ...'
    in english:
    `I wouldn't be surprised ...'
    In my opinion: hype + real technology == excellent deal
    I hope I am lucky and get some of those :)


  • I've been running a copper-interconnect G3 processor in my PowerBook 2400c since last summer.

    I believe that every G3 and G4 chip shipped in a Macintosh since October has been a "copper" PowerPC.
  • >> Like the OGG troll running around today.. Okay yes it detracts from a good conversation and its a waste of bandwidth butt it is comical even if immature.

    Its "OOG", not "OGG", *I* am an OGG not to be confused with said troller "OOG". Hey man, if you had my name you'd be self consious too.
  • >> I found it: the required SDRAM uses 1W, when not in sleep mode. This should be used constantly, so figure a constant extra 1W, on top of the 5W for the processor, when the processor is doing work.

    Correct me if I'm wrong, but doesnt RAM consume a static amount of power, regardless of "usage"? After all, the RAM still has to be refreshed every cycle to retain data, therefore "unused" RAM is actually as active as "heavily accessed" RAM. I actually meant to reply to the parent of this reply, sorry...

  • Just say no to Babelfish.
  • Battery life with some grunt is important for some applications, such as for a public neural LAN. (an R&D project I am currently working on.)
    High-bandwidth often requires high speed CPU grunt to make routing decisions and other network monitoring tasks. Especially with the promise of UltraWideBand technology becoming available in the future, I look forward to the day of being able to power an external node purely by a low cost low power solar panel.
    A transceiver which is barely indistinguishable from noise, transmits data at 5Mbps and can run on a pen light battery for a couple of years.
    Wack a couple of these together with a Transmeta CPU, solar panel (enough to power two 50uW 5Mbps transmitters) to create a node on a publically accessable neural network and 'Voila'!
    A self-configuring, self-healing, low-cost, high-bandwidth, high-density network.
    I'm interested in doing it! If anyone else is, please let me know.
    Check it out at [].
    C.Burgess - email:
  • although i'm not that fit in processor-design, i think it's not possible to do cool code optimizations executed in hardware. if you compare it with the possibilities java-jit has. it's not just code-morphing:
    • parallelization,
    • experimental-code-morphing (trying different possibilities)
    • total destruction/inlining of functions/the OO-design and reconstruction in native code (similar extreme to an white-paper of MIT: "burning c code into silicon")
    somebody knows what kind of optimizations existing jit's do? i've no time to study kaffe src-code...

    i could imagine, that something like java-bytecode will become the bin of the future and half-intelligent vm's would transform it with hardwaresupport to the current cpu's native code/"codes", because future cpu's will be dpu's (distributed pu).

    one could also imagine, to let a intelligent vm compile the code for special cases (functions for special parameters, that often appear) and in worst case, (if the assumpted parameters do not fit) just has to interpret the original code...

    too much imagination to stay alive... but i'm also nothing.
  • here is the correct link to c't []

    Althought you could have guessed :-)
    Check it out! It is the best German computer magazine.
  • I'm not so concerned about Elbrusian technology as I am about Elbonian [] technology.
  • The code translation software optimizes on the fly. In particular, it perform unsafe optimizations and back off if a problem occurs. No static compiler can do that,; they always have to account for the worst case.

    A static compiler can also do a "fast case" and "worst case" version of code. The advance load and check instructions on the IA64 are built for that sort of thing, and the speculatave load as well. The DEC/Compaq GEM compiler can do an alias and noalias version of a function and do an alias check on entry.

    Existing CPUs don't have many features that encurage this, but it can definitly be done. It will bloat the code some, but if the common cases is common enough the extra i-cache traffic will be minimal.

  • Finally an article in which the Europeans have a little more insight views :-)

    Anyway, I can't wait untill this article gets out. I have a subscription to the Dutch version of C'T (note the ', if we don't use it some lame Dutch "magazine" will think we mean them) and I buy the German one from time to time. The C'T itself is one of the more clued magazines around. It's one of the very few who are doing Linux articles (and more offcourse) which date even back to before the major hype. Personally I think you can best compare the C'T with a Byte but with less advertisement & more clued stories.

    Therefore I think its only fair that they get to do such a big (guess it will be big) story about Transmeta. And be assured that if C'T does a story it will be a fair story. Stupid things will be called stupid and vica versa.

    If you don't have a subscription I'd really advice you to try an issue. You won't be dissapointed. Unless you can't read German (or Dutch) offcourse :-).

  • by leiz ( 35205 ) <leiz.juno@com> on Monday March 20, 2000 @04:15PM (#1188996)
    1. the entry level pc is a celeron 500 / k6-2 500 right now, not a P3/K7 700, that's more of a midrance pc.

    2. the transmeta cpu is competing on the MOBILE market, not the desktop market, so it's not competing with the latest athlon or sledgehammer.

    There is no statute of limitation on stupidity.
  • by bjparker ( 95974 ) on Monday March 20, 2000 @02:17PM (#1188997)
    >At 700Mhz the TM5400's performance is slightly >under that of a 500Mhz PIII This isn't going to be anywhere near good enough. The entry level for a new PC *right now* is a 700 megahurtz processor. By the time Crusoe is being sold, you will be able to get a socket version of the Athlon with on-die level 2 cache which will *totally* smoke this thing, and likely won't be much more expensive. Plus Willamette and SledgeHammer will be just around the corner. The power consumption had better be *sweet*, or else Crusoe will never get off the ground because of its woeful lack of performance.
  • by SuperJ ( 125753 ) on Monday March 20, 2000 @12:01PM (#1188998) Homepage
    by Justin Osborn
    To the tune of: Big Bad John by Jimmy Dean

    The mystic Transmeta announced it that day
    It was faster than Intel, as some may say
    Conserved batteries, gave heat the slip
    And everybody knew it was a superior chip - Crusoe

    The folks at Transmeta had slaved day and night
    To make a chip that ran just right
    It ran on software, instructions optimized.
    It wasn't at all advertised - the Crusoe

    A guy named Linus, from Finland they employed
    He made a living doing what he enjoyed
    Hacking the kernel, making it mobile
    Spewing out code and letting it compile - for Crusoe

    The OEMs quickly made alliances
    For laptops, handhelds, and web appliances
    It seemed that they would rake in the dough
    The folks at Transmeta would get the biggest take though - from Crusoe

    I walked to the store to get a laptop
    Crusoe was cheap, they were $2G a pop
    I imagined all the time from rechargin I'd save
    As I read the review the magazine gave - the Crusoe

    I pulled out my laptop flying overseas
    The project wasn't finished, couldn't be at ease
    The hard drive clicked as I typed and typed away
    But the battery just didn't seem to drain away - from Crusoe

    My presentation was going well but it was going long
    My boss began to talk, blather on and on
    The Crusoe made it through, and here's my one quip
    In the middle of my laptop is one heck of a chip - the Crusoe

  • by marcf ( 59515 ) on Monday March 20, 2000 @05:24PM (#1188999)
    It is incorrect that LongRun only works with APM.

    In fact, LongRun samples all traditional power management activity, which is:

    • AutoHALT (ACPI: "C1")
      • x86 HLT instruction;
    • Quick Start (ACPI: "C2")
      • Internal CPU clocks off (responding to a STPCLK# signal from the south bridge);
      • CPU is responsible for maintaining cache coherency;
    • Deep Sleep (ACPI: "C3")
      • External bus clock removed from the CPU (by the south bridge);
      • System is resonsible for maintaining cache coherency.
    LongRun effectively samples the time spent in these power management states, and converts it into a fraction of the overall sampling interval. This fraction directly determines the target CPU performance percentage.

    APM is the older, BIOS-based implementation to exercise the above CPU power management states. ACPI is the new OS-based implementation to exercise the CPU power management states.

    Neither APM nor ACPI are necessary for LongRun to function efficiently. In fact, LongRun works perfectly well with all OS's that are smart enough to invoke HLT whenever their scheduler (or GUI event loop) idles. This is true at least for all modern operating systems (all recent Windows versions, Linux and OS/2).

    As a result, LongRun transparently works without any BIOS or OS modifications on pretty much all OS's you care about.



  • by loik ( 95237 ) on Monday March 20, 2000 @03:45PM (#1189000)

    I can't sleep, so I instead of running around th house I decided to translate this article for a change.

    Crusoe's island

    Journey reports are rather rare in c't. In the last one that I can remember a certain Rob.S.Pierre described a press journey from Apple to Ireland [1] - and whirled up a lot of dust with it. Following an Intel press journey (to IDF) I found myself in Santa Clara, not at Intel's headquarters, but at Transmeta.

    Hardly a stone's throw away from Intel's headquarters the startup company Transmeta resides in a beautiful park, in three one- to two-story houses. From the street next to the park with the characteristic name "Freedom Circle" the relatively ugly concrete/glass-towers of the "big brother" are directly visible.

    Instead of open-plan offices with lots of cubicles the more than 200 Transmeta employees sit mostly in small offices, although there are also a bunch of big labs. I was assured that I was the first European journalist in their holy halls, and above all the very first to be allowed to spy for a whole day in the offices and labs.

    In Linus Torvald's office I felt at home from the start. Torvalds - in proper style in a SuSE-T-shirt - confirmed that he started at Transmeta because he wanted to do something else than always only Linux, Linux. He didn't say it explicitly but indicated that he had been sick of Linux at some point. And so the task of cooperating as a chief architect in the Code Morphing software (CMS) was a welcome change and a big challenge.

    But the ghosts one called... at the moment Torvalds is nevertheless working mainly on Linux again, specifically on "mobile Linux" that is intended as an accessory for the Crusoe processor for small diskless systems (in the 32 MB of flash)

    By the way, Mobile Linux runs as x86 sofware and is morphed dynamically. There had also been experiments with a native Linux version, but it proved less effective than the emulated x86-version. Transmeta also does not want to publish the the complex native instruction set of the Crusoe processors because the company can keep the option of making arbitrary changes. The CMS versions of the Crusoe processors TM3120 and TM5400 are already different.


    Marc Fleischmann - a German, by the way, who is also responsible as a development leader for the virtual Northbridge - could convince me of the effects of the power saving technology LongRun in the Transmeta-lab. A TM4500-Crusoe running between 266 and 600 MHz in five steps (the minimal step is 33 MHz) and between 1,1 and 1,6 V consumed 6 W as a peak value, but on average and with usual software it rested at roughly 1 W and got - without radiator - only tepid. LongRun requires idle states for its regulating. Without APM under DOS LongRun does not work and the 600 MHz TM5400 consumes - when running doom for instance - constantly 5 W but still only warms up to 43 degrees. Included in these power consumption values is the Northbridge - compatible in part with Intel's BX - which is integrated into the Crusoe processor or respectively virtualised by the CMS. As a comparison: the original BX Northbridge consumes above 2 W alone - and doesn't know neither PC133 nor even a second DRAM interface for DDR-266 as the TM5400 does.

    Since CMS is also controlling the Northbridge it cannot only slow down the processor but also the memory if necessary, which makes the potential for savings even bigger. A small, recently discovered bug in the Northbridge software was quite useful actually to show the advantages of a processor realised mainly in software: In front of my eyes software engineer Peter Anvin loaded a fixed version of CMS to the processor, and that was it.

    Responsible for the Crusoe BIOS is Robert Collins. He came together with many other employees from the secret Pentium project that was suddenly abandoned four years ago from Texas Instruments to Transmeta and with his well-known website and Intel Secrets he already fought once with Chipzilla. Collins explained to me that Crusoe first loads the 2 MB (decompressed) CMS in some 100 milliseconds when booting, then installs the translation buffer (8 to 14 MB) and after that, as any proper x86, starts execution of the code at address F000:FFF0. The BIOS itself is coded completely in x86 code and passes system information, e.g. the SDRAM parameters read out of the SPD-EEPROM, through a shared memory range to the CMS software. For the chipset and processor independent part Transmeta licensed the Phoenix BIOS.


    The "x86-validator" Christian Ludloff also came to Transmeta vi TI; he was discovered by TI back then by his article about Pentium secrets in c't [2,3]. Ludloff has literally hacked his fingers sore with millions of assembler lines to produce ugly, malicious, !#@#$# code which traps the poor processors over and over. Like that he discovered loads of bugs and anomalies not only in former versions of Crusoe/CMS, but of course also in the concurrents. Many entries e.g. in Intel's Specification Updates should actually say "thanks to CL". Ludloff is planning to make a part of his assembler software, e.g. all the processor structures, exceptions and so on publicly accessible on his website popular among insiders - that's gonna be a tidbit!

    I confronted Transmeta boss Dave Ditzel with the sometimes uttered suspicion that he took some of the ideas of Boris Babaian during his time as a Sun employee in Moscow. Boris Babaian has been designing for quite a time a processor named E2K at the Russian Sun distributor Elbrus. Babaian was at his's place only a week ago, Ditzel answered. They had a good conversation. He pointed out that the designs of E2K (high end VLIW without dynamic optimisation, very high floating point performance) and Crusoe were very different. A certain inspiration came from there, but not more than from HP, Sun or other designs.


    Ditzel also contradicted publications announcing that products with Crusoe processors shall come out in April. At the earliest at the end of the second quarter, probably in June, the small Crusoe TM3120 will be ready. The TM5400 with more performace and the ingenious LongRun power saving technology is to be expected in the third quarter. It's gonna be the first processor on the market made in IBMs new CMOS8S copper process.

    A possible OEM, besides S3/Diamond and FIC, could be Quanta Computer, one of the biggest notebook manufacturers who produces for many others who only print their Dell, HP or other Logo on the notebooks. Sybase as a big software company has already announced that their SQL Anywhere Studio (mobile databases and data synchronisation) is to be extended especially for Crusoe systems. And the company Infomatec from Augsburg announced that they have built a partnership with Crosstainment AG and Transmeta to develop reference models of internet capable consumer devices with crusoe processors together, based on their Java Network Technology (JNT).

    Transmeta is stil reluctant to publish absolute performance data, yet the values don't look too bad. A 700 MHz TM5400 reaches about Pentium III 500 level. But when it comes to the so important Ziff-Davis benchmark Winstone it is a bit under that level; the reason is according to Transmeta that this benchmark "touches" a lot of routines only once so that dynamic optimisation can't work. The ZD lab has recognized the deficiency in a discussion with Ditzel and will ensure a more realistic use of the routines in the next version.

    And finally there is the most frequently asked question if Transmeta enters the stock market this year. Ditzel's answer: "It wouldn't surprise me..."

Experience varies directly with equipment ruined.