Slashdot Mirror
AMD Invests $7.5M in Transmeta
trouserless writes with the news that AMD has invested heavily in Transmeta. The power-conscious chip company has been financially ailing of late. AMD is taking payment in stock, binding the two companies (both with suits pending against Intel) together. PC World reports: "Transmeta did secure a few licensing deals, notably in Japan, but it also wracked up heavy losses. In January 2005 the vendor announced job cuts and said it would switch its focus to licensing its power management technology to other companies. Later that year Transmeta agreed to sell its Crusoe chips to Hong Kong company Culturecom Technology Ltd. for $15 million in cash. Last year's deal with AMD, to resell Transmeta chips in Microsoft Corp.'s pay-by-installment PC initiative, raised the vendor's prospects again. But in March Transmeta said it faced delisting from the Nasdaq because its stock price fell below $1 for more than 30 consecutive days."
7.5 mil is _nothing_
And the Transmeta website that had nothing on it sans a "No hidden message here" text buried with in the html.
*Sigh*
Rapid Nirvana
$7.5 million is nothing-- but Transmeta's stock is also worth close to nothing... this can only help their stock price. Damn! Should have bought.
Does transmeta actually make anything any longer? or have they degraded to the license(old-stuff)-n-sue model.
Now make me a good embedded project board, including supplemental power from the vibration generators! I guess I should just buy the new board from make (only due to the 4 1-amp motor controllers). Well my 64 bit AMD processor runs my 32 bit windows great!
Intel's overwhelming mobile computing dominance probably left AMD with no alternative but to buy their way back into competition. It would be interesting if they expanded their GPU/CPU thing to mobile processors sooner because of this. Anyway, this spices things up for the near future given that Transmeta processors branded as AMD will gain better acceptance in the market in general.
Cheers!
Atheist: Buddhist in a Prius
What? How? The fact that they both have law suits against Intel? I don't get it.
Technology tips and tricks.
http://www.thefreedictionary.com/wrack
My dictionary only lists "wrack" as a noun...
When Linus joined Transmeta I was really pleased to hear about what they had to offer wrt. technologies that did powersaving (540MHz at 1W) and flexible architecture emulating with the core architecture based on VLIW.
:-)
Since Intel haven't really gotten off with their own VLIW architecture in Merced, which is really disappointing since it is the natural next step towards "HW is something that can be emulated in SW" where the only thing HW provides basically is calculation at higher speeds. I'm not downplaying the importance of that, but let's get the task division straight
So, here's to hoping that this investment can further Transmeta's ideas, and I'll again look into buying AMD CPUs
The real news is that Transmeta is still in business. 5 years ago I joined a company that was using Transmeta chips for their primary product. The chips were really slow, even slower than their MHz rating would suggest. We were the only company that I was aware of that were using Transmeta chips at all.
(Maybe the semiconductor industry doesn't carry the same amount of chair-throwing passion among its leadership that software does? I'm actually curious now).
Quo usque tandem abutere, Nimbus, patientia nostra?
You do intend to say that Intel reminds you of SCO, right? Because in the late 1990's Intel had a problem, their CPUs doubled as toaster ovens. Back then it wasn't a huge problem in desktops, but laptops based on Intel CPUs were roasting a lot of wieners. The solution, read Transmeta's patents to find out how to make chips run cool and save energy. Now after investing time and money to develop something new and non-obvious Transmeta has the nerve to sue Intel! Intel should be able to take (steal) what they want from lousy patent trolls (valid patent holders), make a ridiculous profit, and maintain their monopoly. There's just no justice anymore. Not for Intel, Paris Hilton, or Scooter Libby!
Microsoft's investment in SCO. They only kept SCO afloat to keep the lawsuits going... the same reason AMD seems to be investing in Transmeta -- to keep the Intel lawsuit going.
What's up with merging and investing in all these second tier companies? It's like AMD is trying to form some sort of crappy corporate Voltron.
"Things are more moderner than before- bigger, and yet smaller- it's computers-- San Dimas High School football RULES!"
The suggestion is most likely that AMD is doing a random payout simply to keep Transmetas lawsuit against Intel going. Just like it has been argued that Microsoft did for SCO to support the lawsuit against various Linux companies (though I am personally not all that convinced about this theory either).
At a guess, that's probably a month of twos worth of payroll. Most second-round VC financing for an internet startup would produce more than $7.5m. AMD will probably write off more than $7.5m in a year (O.K, maybe not quite that much. Intel would, though)
Don't get your business news from Slashdot, kids.
TFA says Transmeta shot to prominence due to Crusoe. This is wrong; Transmeta shot to prominence because it hired Linus Torvalds and refused to talk about what it was doing.
Classical Liberalism: All your base are belong to you.
Yes, this should be interesting to watch...I wonder if AMD is spreading itself too thin, but as a low-power and small-form-factor enthusiast I would be very interested in seeing what comes out of this, if anything. I would love a CPU that can power down to single-digits-of-watts in a low-power state.
~Eien no Inori wo Sasagete~ Searching for my Hatsumi...
Transmeta, the company with some quite amazing chip technology (do you know how it translates microcode on the software level to simplify hardware etc? pretty exciting stuff) was left in the position of a patent troll.
Investing 7.5 million in Transmetta is called "investing heavily".
YouTube, a company built on nothing (it's just a damn site for low res flash videos), that didn't make a dollar profit before google bought it, costed 1.8 billion.
A typical startup investment from a VC is around 3-10 million dollars and that's not "heavily" at all..
So with numbers that distorted, I know now: we're in a very fragile bubble right now, and when it burst, it'll be ugly. Uglier than before.
I am a VC and would like to invest $50 million (FIFTY MILLION US DOLLARS) in your blog, please.
In other news, I invested $20 in the Canadian Cancer Society.
Microsoft made SCO their tool/third-wheel/etc., and now AMD is doing the same with Transmeta however Transmeta and to a lesser degree AMD are the underdogs. In the case of SCO and Microsoft, they were the proverbial bullies fighting against Linux. Albeit IBM was no underdog, but it's almost an unfair comparison.
Good. Cheap. Fast. Pick Two.
The success of the Opteron came out of the DEC Alpha teams AMD hired away from Compaq. Now AMD is going to get the Transmeta innovations. Intel spends gobs of money on internal research to come up with new innovations. AMD being smaller cant spend the same so it is constantly on the prowl for talented researchers working at companies going down the drain and buys up the innovations at bargain basement prices and in this way manages to match Intel in the innovation game. Expect something as big as the Opteron was to come out in 2 years time.
**Life is too short to be serious**
which seemed to sell the crown jewels off to some obscure company, which then disappears from public view and eventually went into receivership.
(And it happened to ParkPlace/Digitalk, [They were the originators of SmallTalk] which WAS a company I cared very deeply about. Some shareholders were left holding $17M worth of used toilet paper instead of the valued stock they originally bought. and ObjectShare went onto something else.)
Fucked up my OOPL(Smalltalk) consulting career but royally.
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
Probably, probably not. Transmeta did show one thing, softcores might viable, and if thats true, chances are we might be buying from either Xilinx or Altera for our processor, and then deciding if we should use a AMD, Intel, or a custom implementation from opencores.org.
Boy I wish that would happen.
...am glad that AMD is investing money in Transmeta. Transmeta had some interesting concepts that if applied correctly (without the x86 emulation) would probably revolutionize processors where mobility and power savings count a lot more than all out performance like pocket PCs and really small laptops. I'd bet that it'll work even better on mobile graphics though. Dedicated video cards suck up battery life like a sponge, if the power consumption on those can be greatly reduced when not running visually intense programs... ah the possibilities.
And you thought I'd make one of those overlord comments.
The post might be scathing, but it's not trollish. If I had points it would be "informative" because "sarcastic yet accurate" isn't available.
They seem to be nostalgic for the good old days because my Pentium D runs awful hot, even with an oversized Zalman CPU cooler running full speed.
Comment removed based on user account deletion
By the way: THere's a hell of a big difference between a VLIW emulator and a soft core.
Engineering is the art of compromise.
I was actually trying to be funny...
Sarcastic IS funny.
Not for the target of the sarcasm.
"I'm never quite so stupid as when I'm being smart" (Linus van Pelt)
Otherwise, a brilliant move, I have to say. I should write a bot that posts that line once a day and see how much karma it racks up.
Wracked up heavy losses? Illiteracy has even struck PC World?
Neopets - the best free game on the Int
Thanks. Unfortunately the Intel fanboys have the moderator points.
AFAIK, no modern "x86" CPU is actually internally compatible with i80x86 instructions.
It will serve the discussion well to have this comment moderated up a bit, such that someone who -does- know differently will be more likely to see it and elaborate on the subject.
Even though it is late in the game.
I have two Transmeta Crusoe Flybooks and they are very slow compared to PM processors of equal GHz. The company and the technology is interesting, but I am afraid the Crusoe processor has failed to deliver any advantages to the customer. Unless Transmeta can build a faster CPU, people won't buy their products. The CPU is not the only component using up power. Why should we invest so much research in energy efficient CPUs while we still use storage with moving parts (HDDs)? Why not invest all of our resources in researching and implementing better batteries?
The P/E of the S&P 500 is 21 right now. Bond yields are 5% (a P/E of 20), so the market is a bit over-valued, but I wouldn't call it a "very fragile bubble".
Patrick Doyle
I mod down every jackass who puts his moderation policy in his sig. Oh, wait a sec....
speaking of AMD CPUs and Transmeta back in the early AMD64 days, Transmeta's software microcode and on the fly instruction set-to-VLIW recompile (codemorphing) had been put into use to emulate AMD64 so it could be thoroughly tested before the actual chip could be available.
This helped testing that the 64bit work done for other architecture behaved well for AMD64 and contributed to the fact that AMD64 was supported from the very first day the chip went out of the factory.
In addition to the ultra-low power consumption offered by the RISC design, codemorphing is also a very interesting piece of technology to beta-test new instruction sets before the actual hardware is produced by the fab (even if the emulation on VLIW is slower, it's still a lot less slow than 100% software emulation, and it's cheaper than FPGA - given the high number of transistors in modern CPUs).
AMD can even be doing this on purpose. While Intel is just multiplying the number of cores and basically substituting s/gigahertz/number of cores/ in all their marketing material, AMD are planning a lot of interesting stuff for their AM2 socket, with Torenza initiative and similar : the GPU-in-CPU units and similar on-CPU stream processing vector units (à la Cell processor) offers a lot of new possibility, but will also require a lot of development from the software programmer in order to fully harness the possibility.
Having Transmeta still around is handy, because it opens the possibility of having not so expensive and not too much slow kits to test and develop for newer instruction sets, while the actual hardware is still being developed on.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]