Slashdot Mirror
Intel Responds to Crusoe
spaceorb writes "According to Zdnet, Intel is preparing to do away with the mobile Pentium III in favor of a new chip, codenamed Northwood, in 2001. Northwood is based on Intel's next generation 32 bit chip Williamette, which will mark the beginning of their transition to 0.13 micron. Also briefly noted in the article is the mobile Athlon. "
No mention in this article of specially reduced power consumption (beyond normal mobile chips) so I find it hard to believe that this chip will be able to outcompete the Crusoe. I don't know if the Crusoe has SMP support but if it did you should easily be able to beat the intel chips on speed as well by putting four of them in a laptop with still lower power consumption.
That raises an interesting question: Can the kind of technology which lets Transmeta optimize instruction translation on the fly be used to optimize the execution of single processer code on a dual procesor system? Just like in the translation case we would keep a software table but instead of most used instructions/jumps it would be of independent blocks of code.
Marriage is the "pseudo-ethics" that cloaks the messy truth of sexuality in the raiment of propriety -- it's "Don't Ask,
You missed the point. Intel is timing this ANNOUNCEMENT to chip away at some of the glow around Transmeta. Yes, they likely have had this in the pipeline forever, but they also almost certainly rushed the announcement out the door after the Transmeta pess conference.
Intel needs to be scared by Transmeta. If they're not, they are doing their shareholders a disservice.
We've known about their 130nm process, and we know that all their x86 chips get put into a mobile format. Also, we know that Willamette's core will be release a la the Pentium IV product very late this year.
FWIW, the Willamette is Intel's first new core since the P6 back in the mid 90s, which found itself inside the Pentium Pro, II, III, and Celeron. Rumours and admissions declare the below to be its likely improvements:
The 130nm process basically will decrease the size of the features on the processor. Basically, imagine drawing stuff with a big fat marker, then getting a nice, fine pen. You can make much more detailed drawings, right? Basically same thing here. Benefits of going from 180nm to 130nm process:
-JC
PC News'n'Links
http://www.jc-news.com/pc
There is no mention of either Transmeta or Crusoe in the story, other than in the "related links" section. I don't think any of this was prompted by Transmeta's actions - Intel has this stuff planned out months in advance. Slashdot is just using any chance it gets to hype Transmeta b/c Linus works there.
--
http://gammatron.weblogger.com
Not to beat the computers vs automobiles down, but this does in some way feel like back when the Japanese automobile manufacturers started taking on Detroit, with its low powered but cheap and fuel efficient cars those high powered gas guzzling models. And Detroit learned in the end people didn't need to be able to do 200 mph so much as they wanted 30 mpg.
The laptop market is going to go in the same direction. People don't use laptops, excuse me, most people don't use laptops to run Quake and other games at ridiculously high fps. In a laptop, having good enough performance and a long battery life are more than sufficient. They may be slower but the question is will most people really miss the performance. If one does not run graphics-heavy games (and Microsoft operating systems and applications), how much CPU power does one need?
The Transmeta people had some good ideas. Pity that it's the x86 instruction set they're stuck with, though it will be interesting to see how good a Java virtual machine they might get out of it. For some reason I like the idea of a standardized instruction set and then forcing the chips to run that everywhere very appealing.
As for Intel, I think they're going to have to eventually learn that they can't be all things to all people. Microsoft is starting to slowly come to that conclusion as well. Sooner or later you have to pick a market segment and hope that it stabilizes. Trying to go against AMD and Transmeta just seems a bit much for one company to do.
> Every company that wants to put a cool mobile > device on the market should do it now and -thus- > should use a transmeta chip. There simply isn't > anything else in the high-end mobile processor > area right now.
.13 micron process that the article mentions. They will have to sacrifice functionality and computational speed in order to do so, but Intel is in a position with some of the best hardware engineers and best fabrication plants in the world. They can design a simpler chip which uses less power if they want to. They have "Evil Empire" status with Microsoft, so many people criticize them here, but they, like Transmete employ smart people. (No, I don't work for them and I don't own Intel stock either).
First, I'd like to point out that Crusoe is not being used in devices right now. Their web page tells us to check back in about 6 months or so. Granted, this is before Intel is going to release their new mobile processor, but I think many people would consider waiting several months for a tried and true technology which will have a larger support infrastructure in place. Where does one buy a motherboard that accepts Crusoe processors? Are they available in sufficient numbers? What about other hardware to go with it? There will be loads of support for Intel processors because companies know that there will be a large demand and they've actually heard of Intel before.
Second, you say "high end". My understanding was that Crusoe is not a high end processor. It trades raw speed for lower energy consumption due to a simpler architecture. Certainly the clock speeds are high, but a 700 MHZ Crusoe will not deliver the same performance from a computational standpoint as a 700 MHZ Athalon or Pentium III. I'm not saying that this is a bad thing, as many people here have pointed out that the number one issue with mobile computers is battery life, but the Pentium III and Crusoe were designed with different goals in mind.
There is nothing that prevents Intel from designing a low power chip which uses fewer transitors than a PIII using the
Why must this new processor be a response to Crusoe or the mobile Athelon? Certainly Intel is watching what its competitors are doing, but the Pentium III has been around for quite a while. I think it's perfectly reasonable that this is just the next generation chip from Intel. Just because they release a new chip, it doesn't mean they're playing a frantic game of catch up. I know we all like Linus here, but Transmeta hasn't even proven that it is a threat to Intel. They have a product, but who here owns one? Who makes CPUs for Palm Pilots? Are they also a threat to Intel?
That's just foolish. Intel constantly has chips in the pipeline. Does anyone seriously think that the Intel engineers were sitting with their feet up on their desks and then suddenly there was the Transmeta announcement? "Good God! We better start designing a new chip!"
Not to mention that Intel constantly makes new roadmap announcements, so the FUD argument doesn't fly either.
I still might point out that Transmeta has not even shipped a product yet, so I would argue against throwing FUD in glass houses.
--
Does anyone here realize how long it takes to design a new architecture? How about 3-4 years? (How long did Transmeta take? Almost 5 years.) My point is that Willamette wasn't rushed out to compete with Athlon, since Willamette designs must have started before anyone had even heard of the K6-2. So now you think that Intel has mentioned a future laptop chip just because of Transmeta??? Did you think that the were going to use the mobile PIII forever? Willamette comes out in Q2 or Q3, so you would expect a mobile version a year later, just like with the P6 architecture. I haven't found except Slashdotters who think that Transmeta is a competitor to Intel's main processor business. It is a competitor to StrongArm and Motorola's embedded chips, period. By the way: anyone see the news about Timna? It is a very low power (no numbers were mentioned)completely integerated chip (sound, video, chipset, all integrated with the CPU) which Intel is supposed to announce next week. *That* might be a competitor to Transmeta, but its design must have started several years ago, too. Remember, everyone can see trends develop. Transmeta isn't some god-like company just because the Finn works there.
Haven't you ever heard of Merced/Itanium or McKinley? Intel is (with the help of HP) very hard trying to make a 64bit chip, but is not doing very well. Itanium is behind chedule, a lot. First Itanium silicons have been produced, and the rumour is that they suck badly. McKinley which is mainly HP effort *should* be coming along just fine, infact there is small chance thatit will be done before Itanium (originally it was cheduled 1.5 - 2 years after Itanium). Making a good 64bit CPU is not easy.
And yes, incompatibility with x86 is a VERY large obstacle to overcome when moving to the IA64 architecture. If it wasn't, we'd all be using 64bit alphas/whatever right now. 32 bits will be enough for consumer market for many more years, and the Intels current x86 CPU is already starting to max out in terms of clockspeed (can't find a 800MHz PIII?). Remember that PII, Celeron and PIII are all based on the PPro, which came out way back in 1996. If Intel doesn't soon come upwith a new x86 core, they will totally lose the high-end x86 market to AMD. As for the low end market...64bit CPUs are friggin huge, and thus way too expensive to be fitted into a $1000 computer.