Slashdot Mirror

I think a system like this would be useful for a Trusted Computing (TC - http://en.wikipedia.org/wiki/Trusted_computing ) system on Linux. TC does have some good uses, and having the OS cooperate with Intel's hardware (La Grande - http://www.intel.com/technology/security/ )would be great.

FORTRAN is still an evolving language, look for FORTRAN 90, 95, and 2003.

Many graduate students in applied mathematics and statistics write their new code in these languages, especially since one can find a excellent F95 compiler, free for non-commercial uses on Linux.

I've heard that it uses an Intel technology called "Extensible Firmware Interface" (EFI). http://www.intel.com/technology/efi/

From a functional standpoint I don't know if the Open Firmware tricks we all know and love will carry over, but I share your concern. Frankly, if I can't boot to Firewire Target Disk mode then I don't want one, no matter how fast they may be or what other bells & whistles they may have.

Even someone at Apple when they announced the switch said nothing would prevent you from dual booting Windows.

No, what he said was that it would not be actively discouraged by Apple. That doesn't mean that Windows currently supports all the technology they use--most notably EFI.

From the FAQ:

1.4. Does Xen support Microsoft Windows?

Unfortunately we do not currently support Windows; the paravirtualized approach we use to get such high performance has not been usable directly for Windows to date. However recently announced hardware support from Intel and AMD will allow us to transparently support Windows XP & 2003 Server in the near future. We are working on this and intend to have support available by the time the new processors are available.

Duo is that new processor.

VT and Xen 3.0 will
  virtualize Windows XP...
From PR:
The demonstration features a pre-release version of Xen 3.0 virtualizing both Linux and Microsoft Windows XP SP2. Support for unmodified guests, such as Windows, is made possible by Intel's VT technology, which provides a privileged mode of execution that allows Xen to share platform resources between both modified and unmodified guest operating systems, providing CPU, memory and I/O resource guarantees.

New 3.0Ghz Pen4 has VT and Yonah being part of the 65nm tech... also has virtualization.

From the FAQ:

1.4. Does Xen support Microsoft Windows?

Unfortunately we do not currently support Windows; the paravirtualized approach we use to get such high performance has not been usable directly for Windows to date. However recently announced hardware support from Intel and AMD will allow us to transparently support Windows XP & 2003 Server in the near future. We are working on this and intend to have support available by the time the new processors are available.

Duo is that new processor.

VT and Xen 3.0 will
  virtualize Windows XP...
From PR:
The demonstration features a pre-release version of Xen 3.0 virtualizing both Linux and Microsoft Windows XP SP2. Support for unmodified guests, such as Windows, is made possible by Intel's VT technology, which provides a privileged mode of execution that allows Xen to share platform resources between both modified and unmodified guest operating systems, providing CPU, memory and I/O resource guarantees.

New 3.0Ghz Pen4 has VT and Yonah being part of the 65nm tech... also has virtualization.

This relationship will not last long. I mean look at Apple's main site for the Macbook. It doesn't use Intel's new logo, just some stupid Apple gif. Hah, some partnership....

I'm getting kind of bummed about the meager gains we're seeing in recent years with CPU speed. Sure, we're going to multi-core, but there's only so much a programmer can do to wring performance gains out of threading without making things too complex and unmanageable. I want to know when the next real quantum leap is coming, like optical processors or quantum computing... something that gives us an order of magntitude increase instead of another 8%. Looking at Intel's timeline for the next 10 years isn't encouraging. A lot of fuzzy ideas about the Internet, "user experience", and parallelism. I wonder where we'll be performance wise in 30 years, in comparison to the difference between now and 1976.

Intel doesnt offer a Centrino Solo at 1.5GHz! The only single core chip they offer is at 1.66GHz. Read the Intel price list and educate yourself.

I would expect that the Radeon 9550, being a discrete chip with its own memory (albeit only 32MB) would outperform pretty much any current integrated graphics setup, including ATI's own.

From ATI's 9550 page we can see that the GPU has 4 pixel pipelines and 2 vertex shader pipelines.

The 945GM features GMA950 graphics. From Intel's GMA950 page we can see that this core runs at up to 400MHz, and can render up to 4 pixels per clock. In terms of performance however, from ExtremeTech we can see that the GMA950 performs much worse than a 6200TC, which is probably on par with a Radeon 9550. "To put it more bluntly, it's a complete and total rout for the GMA950". Some games wouldn't even run, possibly due to driver bugs (and people claim that Intel has great stable chipsets with no issues).

However to cut costs I can't see Apple continuing to use a dedicated graphics chip on their low-end products. They have to compete with PC laptops, and people don't know that integrated graphics suck (but are still usually good enough for desktop use).

In case you didn't know, Apple's Developer Transition Kit PC (which supposedly runs OS X quite well) uses Intel Graphics Media Accelerator 900 (GMA 900).

Intel's newer integrated graphics (GMA 950) performs a little faster, but is not much different architecturally. Since the chipset shares system and graphics memory, the use of dual-channel DDR2 667 should improve performance a bit due to the increase in graphics memory bandwidth.

While I understand these aren't meant to be gaming "powerhouses", I'd still like to do some light gaming on it. (FWIW, I've play World of Warcraft, at around 20fps, on my old iBook G4 800 and my desktop P3 750 (ti4200), which was acceptable. I'd like to see at least this level of performance).

Anandtech has a small gaming performance review of the desktop version of GMA 950. The benchmarks were run with high quality settings at 800x600. The games (and fps) were Doom 3 (14 fps), Far Cry (10.2), Half-life 2 (20.2), Unreal Tournament 2004 (26.5), and Wolfenstein: Enemy Territory (30.5).

Intel also has a "Game Compatability List" for GMA 950 which "green lights" World of Warcraft as "No known issues with default settings, or all issues resolved." However, this was the same rating given to Far Cry, which only ran at 10.2 fps (high quality 800x600) in Anandtech's review. Also, Half-life 2 is "yellow lighted" as "Game operates with minor issues or workarounds" because of low frame rates, but Anandtech's review had it running at 20.2 fps.

A quick follow up on this: I was perhaps a little too cynical. The Intel 945GM chipset (the "Napa" chipset that includes an integreated GPU)uses the Intel GMA 950, which actually sounds fairly reasonable for a GPU optimized for low power consumption. Whether or not the PowerBook uses the GMA 950 is a different story through.

... don't expect much. The blurb for their "Vista-ready" chipset, they're not promising anything:

"Intel's next generation mobile platform, code-named "Napa", is scheduled to launch in early 2006. Napa-based laptops using either Intel's integrated or certain 3rd party discrete graphics are targeted to support the new Windows Vista Display Driver Model."

Translation: you'll still need a third-party GPU if you want decent 3D graphics performance.

My bad.

http://support.intel.com/support/notebook/sb/cs-00 6408.htm

Intel® PRO/Wireless 3945ABG Network Connection
Driver is expected to be available Q1 2006.

BTW, the link titled view specification chart on http://www.intel.com/products/centrino/duo/index.h tm contains no specifications at all.

BTW, the link titled view specification chart on http://www.intel.com/products/centrino/duo/index.h tm contains no specifications at all.

No, me too. I think we are the only two.

I mean they want to have differentiation between the consumer and pro models. Why not release the ibook with "similar" specs to the current powerbooks with celeron or lowend pentium m processors? This will prevent major cannabalization of the current powerbook lines and allow them to release the yonah based powerbooks as the pro model they really are.

There will be a single single-core version of Yonah at launch and this might be a good candidate for the iBook (with PowerBooks being dual-core). However, at $209, the single-core Yonah seems a bit expensive for a "low end" notebook. In a few months, Intel will likely release Celeron M CPUs based on the Yonah core (single core), but with a lower FSB (533MHz vs 667Mhz) and less L2 cache (1MB vs 2MB). These would be way too late for a MacWorld announcement.

I think there's a very good chance that the current Dothan-based Celeron M will be used in the new iBooks (and maybe Mac minis) that are expected at MacWorld. Based on the current Pentium M core (but with lower FSB and L2 cache), the Celeron M is cheap, fast, and runs cool.

The current Celeron M does lack SSE3 support and many people think SSE3 will be required for the new Intel Macs. However, I think this assumption might be false and is based on the fact that the Dev Kit version of OS X for Intel requires SSE3 (the Dev Kit CPU has SSE3). Since the Dev Kit version was only intended to be used on one CPU, and SSE3 is off by default in XCode, I suspect that SSE2 will be enough for the official release of OS X on Intel.

If 'almost too hot to touch' is below the specs for the processor's operating temperature range...it doesn't matter how hot it feels to the user.

First, I have to mention the obvious. The heatsink is going to be COOLER than the processor itself. (That's the whole reason it works as a heat sink.) If the heatsink is below the maximum temp for the processor it doesn't actually mean you're safe. Without knowing the thermal resistances for everything, you simply cannot say whether you're safe or not. Especially thetaCS, the case to heatsink thermal resistance.

Second, even if you're within specs IT DOES MATTER for reliability reasons.

Even notice how many motherboards die because those big electrolytic caps sitting next to the CPU fail? The failure rate of components is quite signifcantly affected by temperature. For example, an electrolytic capacitor might be rated for 8,000 hrs at 75 C but that increases to 32,000 hours at 55 C. The same is true (but generally to a lesser degree) for virtually all components.

There are also other little niceities like a lower operating temperature meaning a lower leakage current.

It never ceases to amaze me how people with no training will second-guess the basic competency of others with degrees in their field.

It never ceases to amaze me how pompus us people with degrees can be.

I also love the egotistical "we drive 'em hard" implied in the "marathon gaming" bits- as if they're HARDCORE users who STRESS the xbox beyond its limits.

I find this obnoxious too.

I guarantee Microsoft had units running benchmarks/game demos for WEEKS at a time doing burn-in...

Me too, but I certainly DON'T guarantee what the results of that test were. Maybe the results projected that many xboxes would barely outlast their warranty coverage. We don't know. The only thing we can guess is that they were reasonably sure it wouldnt' fail in a manner where it had to be replaced within the warranty period.

On one hand, you would think people could do their job.

On the other hand look at the nice power cords microsoft sent out to 1st gen Xbox owners so their houses don't get burnt down. Could these be those same guys with degrees you're talking about?

People with degrees in their field make mistakes all the time. I'm not saying these guys found one, but acting as if it's crazy to question someone because they may have a degree is out of line.

Speculation surrounds Apple's rumoured x86 release (some rumours claim this month) whether they will adopt Intel's EFI, an OpenFirmware substitute.

Intel has been working on EFI for 7 years. Traditional PCs typically ship with BIOSes because of legacy operating system and hardware support. Apple would have no such problems due to building a new platform. Plus, particularly in the notebook space, designs will be heavily influenced by Intel.

My point in mentioning Apple? That other manufacturers may follow suit once any quirks are worked out and Vista ships.

There's already a linux bootloader, elilo for EFI, developed originally for Itanium.

Then there was EFI. Open standard, some open source, compatible with the BIOS, scriptable, able to be updated without a DOS floppy disk.

This is the 21st century.

I for one welcome our communist, indium antimonide clad overlords!

Content from http://www.skoool.com (flash required) is being used on Linux computers in schools in India.

That said, not everyone believes that branding really provides much value to companies. I'd say that in general whether you believe in the power of branding or not depends in large part on how much money you've spent on branding and how successful your company has been in the marketplace. If you've spent a lot of money on branding and have risen to dominate your market, you likely won't want to stop your branding efforts, if for no reason other than to avoid rocking the boat. I'm sure many big branding spenders don't have much emperical data to back up their belief that branding is effective. On the other hand, bean counters aren't always very good at incorporating intangibles into their calculations. When is the last time you heard a CFO say, "Yeah, we need to spend more money on tech support, so people will get a warm fuzzy feeling after we respond to their problem."

The Intel brand was ranked #5 for Business Week's 2005 Global Brands Scoreboard just below Coke, Microsoft, IBM & GE, with a brand value of $35.6 billion.

They must have done something right... some combination of image, cost, and core-values. Seems Intel stives to have all employees aligning with these values. That 1994 Pentium bug could have really damaged the brand, but they managed it by apologizing, recalling and replacing at a cost of $300 million.

I suggest you read this paper.

Computing is going all low-power and parallel. Check out Intel's Platform 2015.

Read about the benefits Intel ascribes to 64-bit software here. "Processors with Intel EM64T support 64-bit capable operating systems from Microsoft, Red Hat and SuSE." And you won't be able to run them.

There are some applications where a 64-bit CPU can perform FOUR TIMES more work in 64-bit mode than 32-bit mode. One of these is big integer multiplication. Check out Is 32 bits really better than 64?": "If we instead would compare an Athlon XP and an Athlon 64, the latter would be almost 4 times faster. Why 4 times and not just 2 times? Because a 64x64=>128 bit integer multiplication actually performs 4 times more work than a 32x32=>64 bit integer multiplication!"

If you want a low power 64-bit CPU consider an AMD Turion based notebook. Check out this article and its conclusions. In particular, "A lot of people see Dothan's 27W TDP & Turion ML's 35W TDP and assume that Dothan is automatically lower power. Intel computes thermal design power as 75% of the maximum load on the chip, while AMD's TDP rating is derived from the absolute worst case power dissipation of the chip. Part of the total system power is also incorporated into AMD's TDP, as the memory controller is located on-chip. Intel's memory controller is built into the chipset and thus draws power not calculated as part of Dothan's TDP. Also while Turion 64 is at idle (800MHz clock speed), it's performance is likely to be higher due to the higher bandwidth data bus. All of these factors contribute to Turion 64 being more power efficient under low load circumstances."

And the -MT Turions have even lower power consumption: AMD Turion 64 specifications.

My next notebook will not be constrainted to only running x86-32 software.

"... how much more do you think your new P4 processor would cost if it was fabricated in the USA?"

Probably about the same as it costs being produced in Israel. Just because the current P4 is fabricated in Israel doesn't mean that the new ones will be. If it was strictly about cost, they'd be producing the chips in China, or Malaysia where regulations are lower and labor costs are far lower - but it isn't that simple. The Israel fab was probably just the cheapest to convert for producing the new chips.

In fact, the newest technology will be produced right here in the good ole USA in Arizona. In the USA there are more than a handful of Intel fabs:

from http://www.intel.com/pressroom/kits/manufacturing/ manufacturing_qa.htm#1

How many factories do you have worldwide, where are they located and what percentage of your workforce do they employ?

Intel has 11 fabs worldwide today. The company also has six assembly and test sites worldwide. Intel has 15 manufacturing sites worldwide. Sites within the United States are located in Chandler, Ariz.; Santa Clara, Calif.; Colorado Springs, Colo.; Hudson, Mass.; Rio Rancho, N.M.; Hillsboro, Ore.; and Dupont, Wash. Sites outside the United States are located in Shanghai, China; San Jose, Costa Rica; Leixlip, Ireland; Jerusalem, Israel; Qiryat Gat, Israel; Kulim, Malaysia; Penang, Malaysia; and Cavite, Philippines. Approximately half of Intel's 78,700 employees work within Intel's Technology Manufacturing Group.

From http://www.anandtech.com/cpuchipsets/showdoc.aspx? i=2592:

"Intel has announced that it has reopened its Fab 12 facility in Arizona and is preparing to produce the latest in processor technology at this location."

It may come as a surprize to find that there are many other fabs in the USA that are not Intel and are in fact making inexpensive chips. Some of them are overdemand facilities that get contracts to produce chips when the manufacturer can't meet demand by themselves. Not every, inexpensive chip is fabbed overseas, just as not everything from the USA is expensive.

It costs many millions to build a new fab so chips are not always produced in the place where most people would think that it costs the least. In many cases, the production is in the place where it costs the least to convert the equipment since the cost of building a new fab from the ground up can actually be greater than the labor costs of operating the fab over the life of the fab (3-5 years in some cases).

Its function pretty much what this TPM chip seems to want to do. Do not worry, your P4 does not have it nor does your AMD chip. It caused a bit of a stink and Intel backed down. For the time being. At least they claimed. Who knows what is really inside your pc.

Seems this time it has managed to get a little bit further. Now anonimity on the web is a bit of a hassle as /. is probably well aware of. Just count the number of people who feel they can troll websites without fear of being found and getting their teeth kicked in. If you behaved like the GNAA in say a real life setting like a soccer club you would easily find yourselve gently reminded about proper behaviour. On the net there are ways to hide yourselve and it is used.

Any freedom will be used and by some people in ways that other people would rather not want. For instance while you do indeed need legal identification to drive a car and for that indentification to show you are entitled to drive a car there is no real check. Wich means that some people who are not allowed to drive a car do. It is called joyriding and it results in a fair number of deaths. Yet we more or less realize that this is the price to pay for not having to go past a checkpoint when we leave the driveway.

A hardware based identification sounds "nice" if you think about the need to identify yourselve to your bank. Well no actually it doesn't. I don't know how other banks work but the dutch "postbank" sends an SMS to your phone when you have (after giving a login/password pair) entered a transaction to confirm your identity. It works because a wrongdoer can't easily get their hands on both. If you keylog my machine you will not have my hardware phone and if you rob me you will not have my password.

Oh btw this TPM chip does NOT seem to guard against keyloggers. If I get your PIN number I can simply login on your machine and be identified as yours. Not good.

What is worse I think this TPM chip can only work with closed source. A linux machine could of course easily spoof the number unless it was part of the network card (and even Mac adresses can be spoofed) so I think this little chip is far more about MS and (perhaps) Apple wanting to ban opensource software then anything else. Oh they do not need to do it openly. Just that more and more websites and content will require a tpm chip. How about enabling tpm identification by default in every word document so that even in the new "open" format it would only work with MS software, opensource wordproccesor would simply not have the TPM chip to decode it. So you would have to ask each time your send a piece of content please could you send me a non-tpm version?

Nah, nice as it would be to ban the trolls easily I fear that they are the price we pay for the freedom to run the software we want. Time perhaps to mothball one of my current machines just in case in the not to distant future there will be no mobo's left to buy that run linux.

Not that I think that is going to happen. Why? Well our glorious defenders of freedom the chinese. The article suggests that this TPM chips is Bill Gates way of getting the chinese to pay for MS software. Lets hope the chinese are smarter and that the TPM chip is the way to get Chinese even more serious about creating an independent IT system. I rather trust the chinese goverment, who has no control over me a dutch citizen, then our own bought and sold goverments.

If gadgets are so worthless, then things are going pretty bad for Intel. :)
http://www.intel.com/personal/resources/five_must_ haves.htm

the MT-40 turion is 2.2ghz at less than 25W
http://www.amd.com/us-en/Processors/ProductInforma tion/0,,30_118_12651_12658,00.html

compare that with intel's P-Ms which go up to 27W
http://processorfinder.intel.com/

then factor in that the turion wattage includes the very hot memory controller and the PM doesn't, since its on-die for amd chips but on the chipset for intel chips.

therefore amd has the power/heat lead in laptops. they've had it in everything else for a while now.

Actually, in the last couple of models, AMD has been making huge strides in reducing the heat dissipation of their models, beating intel pretty often. Wikipedia has an article showing some of the heat dissipation details..: Wikipedia.
But then again, given that the specs of the notebook show it to be a 500 MHz machine, it is true that the Intel chips of that time were more thermally efficient as show by these websites: Intel CPU thermal ratings and some others including many AMD cores. But then again, we have the AMD Geode that consumes max of 4 watts at 500 Mhz. So... it's really something that AMD is just as competent at I presume.

Some Anonymous Coward wrote:"why the US is in the forefront of technological innovation and Argentina is not." The US used to be in the forefront of technological innovation. That is long past. Now the US is in the forefront of litigation. M$ comes to mind. Bloat and obsolescence are honoured in the US. Check out Asia. M$ is in rapid decline there. The top reason for businesses to switch to Linux is that their competitors have switched and they are at a disadvantage without using Linux. M$ is selling 20th C products in the 21st C. That doesn't fly well. The rest of the world will eat them for lunch. Look at automobiles. The US is still using gas guzzlers. Look at steel. etc. Look at Intel. They are selling their chips at a very high price compared to AMD for similar performance and using more power, too. Intel is going down because the world can see they are second rate. Both of these US corporations do most of their business outside the US and are essentially global e.g. http://www.intel.com/pressroom/kits/manufacturing/ manufacturing_at_a_glance.pdf http://www.amd.com/us-en/Corporate/AboutAMD/0,,51_ 52_502,00.html The cyclotron going to Alaska is sixties technology. It produces only a little gamma radiation if C12 blocks are used to prevent neutron emission. It should have some concrete shielding or be placed underground. Very little chance of environmental contamination. Internal parts may be readioactive.

Parent needs to be modded up more it is the most coherent comment on the topic posted so far. One minor nit pick - a 65nm\45nm fab costs about $3.5billion see here for the investment required for Intel's Fab 28 in Israel. That's an increase of $1.5 billion on the cost of the existing 90nm\65nm Intel Fab 24 in Ireland .

OK, I officially mediawiki too much. Please allow me to rephrase:

Then I swapped it for an 2915ABG, which is natively supported in Linux just fine.

Depends on what wifi card you're talking about. My Acer Aspire used to have a Broadcom chip controlling the wifi, for which my only solution was Ndiswrapper (which incidentally worked pretty well, so long as routers weren't encrypted using anything other than WEP). Then I swapped it for an [http://www.newegg.com/Product/Product.asp?Item=N8 2E16833106219&ATT=Network+Wireless+Ada&CMP=OTC-yah 00TT Intel 2915ABG], which is [http://support.intel.com/support/notebook/sb/CS-0 06408.htm natively supported in Linux just fine].

http://support.intel.com/support/wireless/wlan/sb/ cs-006205.htm This is for my card - 2915 A/B/G. Also, of course, always use the latest drivers.

Are there webcams out for Linux that actually support face-tracking?

The library is still alive. Apparently a group of the DEC developers are now at intel, and the the library is called the Math Kernel Library (MKL). http://www.intel.com/cd/software/products/asmo-na/ eng/perflib/219769.htm

Interesting source of the information: I met someone on the IRC the other day who told me the following...

The biggest thing I wonder about in "The key is stored inside the CPU". This adds cost, but it is possible. It means that to execute your own code, the serial number must be determined so that a replacement flash chip can be properly encrypted. I'm betting it's pretty hard to find this number out without taking apart the processor.

t seems Intel's gone into hibernation mode kind of like they did right before releasing the Pentium 4 in the first place. (Allowing the P3 [and now P4?] market(s) to stagnate and die off?) Come on Intel, what are you up to???

Their processor roadmaps are hardly secret. The Yonah chip comes out in January. It's the first of their 'cool/cheap/multicore' strategy. IMHO it marks the death of the Marketing Department's control of Engineering (NetBurst/GHz,GHz,GHz) and the Pentium-M strategy starts to become the desktop strategy for Intel. Fortunately we all run multitasking OS's these days and care more about cores and threads than clock cycles.

I thought Intel made the first laser on silicon.

link

But what do I know? Maybe it's a different kind of laser.

Another first silicon laser? So who was really first?

----
http://oemagazine.com/newscast/2004/102604_newscas t01.html

Los Angeles, CA | 26 October 2004 -- Researchers at UCLA have demonstrated the first silicon laser, which could lead to more effective biochemical detection, secure communications, and defense against heat-seeking missiles.

----
http://www.intel.com/technology/silicon/sp/

First Continuous Silicon Laser

In a paper published February 17, 2005 by the prestigious scientific journal Nature, Intel researchers disclosed the development of the first continuous wave all-silicon laser using a physical property called the Raman Effect. They built the experimental device using Intel's existing standard CMOS high-volume manufacturing processes. This is the third silicon photonics paper Intel has published in Nature since 2004, beginning with the modulator breakthrough (see the Learn More section).

----
http://www.photonics.com/readart.asp?url=readartic le&artid=325&bhsh=1050&bhsw=1680&bhqs=1

PROVIDENCE, R.I., Nov. 21 -- Silicon has made its way into everything from computers to cameras. But a silicon laser? Physically impossible -- until now. A Brown University research team led by Jimmy Xu has engineered the first directly pumped silicon laser by changing the structure of the silicon crystal through a novel nanoscale technique.

No, the new Celeron D processors are LGA775. http://www.intel.com/products/processor/celeron_d/ index.htm

The actual cache usage is the performance hit here. The big issue with hyperthreading is that it was designed a bit on the cheap, so certain control structures dealing with controlling cache filling aren't in place - the two logical threads fight over who fills the cache. As a result, there's a greater chance of cache misses, causing the CPU to have to fetch data from the significantly slower main memory. Intel has an excellent basic article describing the issues regarding various code/data block sizes and cache sizes, explaining the difference between how to code for hyperthreading processors and how to code for non-hyperthreading processors. What it boils down to is you've got to design your code to use less than half the cache size if you want hyperthreading to be effective.

I would assume Intel's dual core's are the same, though I haven't looked into them much. Indeed, it looks like they use 1mb L2 per processor (Found here.

I was very happy to note Intel released drivers for Linux. I use Linux desktop on Intel 865. I did not want to buy dual-core (Intel 945-based) because no drivers.

BUT, I was very surprised and very much disappointed to see notes on Intel website.
(http://developer.intel.com/design/motherbd/cz/cz_ drive.htm)

For Windows:
Note: The Windows* OS device drivers listed have been evaluated on this Intel® Desktop board. ....

For Linux:
1. All BETA and Linux* information and software contained herein is provided "AS IS" to Intel customers. Intel Corporation disclaims all express or implied warranties and liabilities for the use of this document, the software...

Why such a discrimination? For Windows, fully tested, for Linux, eat as it is!!!!

Isn't it a shame for manufacturer like Intel to say we released drivers untested??? This is unbelievable and very very iresponsible.

We can forgive a poor open source programmer who develop a driver and say I don't have resources to test for all chipsets but it runs on my computer.

What's the execuse multi-billion dollar Intel has to say? No resources? Cannot afford to allocate at least one programmer to test them?

Thanks, I forgot about that. However, the current Dothan-based Celeron M does support SSE2 ("Streaming SIMD Extensions 2"). Also, according to the OSx86 Project site, SSE2 processors will work with a patch for Rosetta. Rosetta works best with SSE3, though.

1. OSx86 Project Forum > Technical Work > Hardware > Sticky: SSE2 and SSE3 information PLEASE READ

   *No one with less then sse2 has any chance

   *SSE3 is perferred and will run rosetta the fastest.

   *SSE2 can run os x intel but with patches such as in coregraphics, there now are new patches that will allow rosetta to run.

2. OSx86 Wiki > HCLPart > 1.14 CPU

I'm not so sure about the current Celeron M anymore. It will work, but I don't know if the lack of SSE3 will have a big performance hit.

I think he means the current Celeron M. I don't know why everybody is assuming Apple will use the upcoming 65nm Yonah core so soon.

A current Dothan-based Celeron M (Pentium M's core, 400MHz bus, 1MB L2 cache) should be a significant upgrade over the G4 processor in the iBook and Mac mini. It's also cool enough for the small form factors and cheap enough for this price range.

From my previous post, the specifications from Freescape indicate 512Kb of L2 cache. Intel is just slightly ahead on this front, and that means a lot (updated this past October).