Slashdot Mirror

First off, you made a small error, the Xeon 5080 has as much L2 cache (2 x 2 MB) as the Xeon 5160 (4 MB, shared). As seen here for the 5080 and here for the 5160 (Intel hasn't updated its pricelist, yet).

Now, given the description of your needs, and given the choice between 2 x 5080 (3.73 GHz) and 2 x 5160 (3.0 GHz), the fastest system would be, without hesitation, the 2 x 5160. Despite its raw clock speed advantage, the 50xx generation is so poorly designed that even a "slow" 2 x Opteron 285 (2.6 GHz) would probably outperform it (or at least match it) and at half the power consumption & heat levels. Even Intel admit it implicitely in the way they talk about Woodcrest ("now we are back", etc). This is the whole reason why Intel developed the Core microarchitecture in the first place after all...

Between a 2 x Xeon 5160, and a quad dual-core Opteron (8 cores), I don't know. It really depends whether your workload scales well with the number of cores, and on how much it depends on memory latency & throughput. Technically speaking a quad dual-core Opteron, such as 4 x Opteron 865 (1.8 GHz), offers a similar or a slightly more important amount of raw computing power than a 2 x Xeon 5160, offers easily up to twice the memory throughput and offers up to half the memory latency (but only for local physical memory, not memory accessed over HT links). But a quad dual-core Opteron would also cost about $1000 more in CPUs alone. But then again the Opteron DDR400 memory would be cheaper than the Xeon FB-DIMM sticks, etc. Overall I estimate a 4 x Opteron 865 would be about $400-$800 more expensive than a 2 x 5160. And we really can't tell which one is better for you because there are too many variables, one of the system could literally be 40% faster than the other one, and it is hard to tell which one without doing real tests on real hardware. If possible, ask the vendor to lend you a system (this is what we do at work).

If I were you I would maybe choose the 2 x 5160, if only because it will be the first Woodcrest system in your hands, and you will be able to test it yourself in your environment with your apps. So you will have a pretty good idea of what it's capable of.

Oh and your $5K quote looks reasonable given the Windows license, the SAS disk, etc.

While I didn't RTFA, I would like to inject my two cents:

Intel's latest VT technology in Intel Macs assists in running an OS in a virtual space and allows that OS to directly (or transparently) access the hardware. AMD is working on a very similar technology that would allow the exact same hardware-accelerated VM. From Intel's Press Release: "Provides headroom for more robust hardware-assisted virtualization solutions." (source)

The summary's mention of "AMD's Pacifica virtualization implementation" seems to suggest that this proof-of-concept "virus/worm/whatever" is very similar to a virtual machine but with the major exception that it utilizes a technology in hardware that is not widely adopted -- yet. Being that the software itself does not have to virtualize the environment and instead taps hardware means that it could potentially be very small compared with a 100+MB download of VMware, and its size and ease-of-distribution (potentially within a worm/virus/trojan/rootkit/etc. could make it a huge, undetectable threat; however, if the user beats the virus to the lowest-level VM (i.e. direct access to hardware), it would be impossible for an infection to be completely undetectable.

And why isn't this innsight in more of those motherboard factory tours? As every
other consumer I have responsibility for what I buy, but it seems I am part of very few that is actually clearly aware of that every time I choose a product.

If people would spend less time on religion and more time interested in planet earth and things like these (when I say planet earth I mean everything also humans), the world might get better. Sure, 99.9% of such caring may all be rediculously futile and a waste of time, but with power comes responsibility, and in time we might get extremely effective of stopping evil companies.

Tommy Hilfiger busted again for use of slave workers in Mae Sot
http://nrk.no/programmer/tv/fbi/3296787.html
http://jooh.no/root/text/Tommy_Hilfiger/email_to_h ilfiger.txt

Maybe I should start a religion and hammer these few words into the cheep so they would think about it in their daily acts. For now I'll post the parents simple and insightful sentences here and there. Someday a company might start up and stamp their products "built in fair work conditions" and enough brainwashed people will actually buy them and the rest is history.

For now buy Intels lead-free NICs.
http://www.intel.com/network/connectivity/products /pro1000gt_desktop_adapter.htm

Some anti-idealists may say don't stress about it but don't listen to them. You are either part of the problem or part of the solution.

You might want to consider writing a back-end for SWIG to support your language, which will make it easy to plug in new primitives and integrate libraries with your language. That will sure save you a lot of work, and enable developers to easily plug their own stuff in! SWIG lets you write interface definition wrappers for libraries (and can even use C++ header files directly, or you can tailor the interface by simplifying and modifying them with annotations), and SWIG automatically generates all the glue code for a wide range of languages, or pure xml that describes the interfaces for other tools to grok. There's an API for extending SWIG and writing new back-ends.

SWIG is brilliantly and practically designed, and deals well with harsh realities and nuances of C++. It totally rocks!

For a great example of a complex library exposed to scripting languages via SWIG, check out the Intel Open Source Computer Vision Library, which plugs into many scripting languages by describing its APIs to SWIG. If you wrote your own language back-end, you'd be able to SWIGify an OpenCV interface to your own language, as well as any other library that has a SWIG interface.

-Don

• My "older" PC network runs fluently over my CAT-5 network; fully seperated from the company network using some LEVEL-1 switches; full speed, my network connectivity never goes down. Although still; it goes fast; but when transferring bigger files like my recorded dj-sets it's sometimes a hog.
• I guess I can also outrule transferring DVD's with that, but with my cheap mediaplayer that ain't a problem either. A backup is made once a month overnight of this thing; my dvd's are ever-after underused and playing through the network (using OPEN netbios (DUH Freecom!) as extension)
• I tried to compare 10/100/1000 differences with a newer PC installed with the intel gigabit networking options, just like 3 of the servers; connected to a gigabit supporting switch to the company backbone. The difference is very minimal; and this talking about ONE pc using an entire switch for itself with 2 servers connected to it (currently as backup network)
• Now I got myself an Apple after 15yrs of PC's; it's a relief; only; it does transfer at speeds of +/- 1mbit/s instead of 10 or 100mbit/s so what's going to give me advantage to that (and what's up with that Apple?) and (off-topic) it crashes out of the blue (literally - no screen ; just hanging ; powerkey still works).

The Intel® Core(TM) Duo Processor also includes support for Intel® Virtualization Technology which is a set of hardware enhancements to Intel server and client systems that combined with the appropriate software, will enable enhanced virtualization robustness and performance for both enterprise and consumer uses.

Lookup the SATA setting in BIOS. The A8V model has SATA settings that can be RAID or AHCI setting which will let OS to detect SATA drive as IDE drive.

ftp://dlsvr03.asus.com/pub/ASUS/mb/socket939/A8V-M X/e2337_a8v-mx.pdf

Look at the 2-26 or page 68 where it shows you how to set the SATA mode as AHCI. This is the only work-around I've found so far.

http://www.intel.com/technology/serialata/ahci.htm
http://linuxmafia.com/faq/Hardware/sata.html

This should be suffice to gather enough info on the workaround. I had a bitching time setting up the SATA on mobo (although it wasn't A8V-MX but with same chipset and bios) until I found out my mobo accepts AHCI standard for SATA controller.

Intel distributes its C++ compiler for free under two conditions: compilation must be for non-commercial use

In other words, you're likely to be forbidden to distribute compiled binaries at all. A lot of programs that hobbyists would compile are under GNU GPL or another license that forbids adding restrictions against commercial fields of endeavor.

See http://www.intel.com/cd/software/products/asmo-na/ eng/compilers/clin/219856.htm

Linux only. So I can either spend $400 on the Windows compiler or spend $400 on purchasing Linux-compatible hardware to replace my Linux-incompatible hardware. And I still can't compile the Linux kernel and distribute the resulting binary because the non-commercial license is incompatible with the GPL. Hmm...

Misconception #553: The Intel C++ compiler costs money for hobbyists. Fact: Intel distributes its C++ compiler for free under two conditions: compilation must be for non-commercial use and you don't get committed support. Given that these are 'hobbyists', then these are perfectly reasonable requirements. See http://www.intel.com/cd/software/products/asmo-na/ eng/compilers/clin/219856.htm.

A fast compiler might be something like Intel's own compiler: http://www.linuxjournal.com/article/4885

Can most hobbyists afford its 400 USD price tag?

Maybe you should inform Gordon Moore, since he argues exactly that it translates into lower cost.

More Intel employees should say in public what they have told me in private: Intel CEO Paul Otellini is not a competent leader. He lacks social ability. He has added to the adversarial management style in which employees are pitted against other employees, rather than encouraged to do their best.

It's very, very sad to see Intel on the way down. Intel processors have literally helped the world become a better place.

Self-destructive behavior at Intel did not start with Otellini. Long ago, Intel closed its consumer division because it could not manage it effectively.

Do I think I could be a better CEO of Intel than Otellini? Yes, I do. I told one Intel employee that I thought of applying for Otellini's job, not because I thought I could get hired, but because I might possibly be able to educate the Intel board of directors about what needs to be done to pull Intel out of its long-term slump.

First, I would re-organize Intel's marketing, which has become Zombie-like in that it has been minimally connected with reality. Last year Intel sent me several email messages offering an "Intel BunnyPeople(TM)" doll if I would be involved with one of their marketing efforts. Possibly there were those who make purchasing decisions who were not offended by an offer of a free doll. However, I've never known one who would be influenced by such an offer. There are plenty of other examples of the scary disconnection of Intel marketing from Intel's needs.

For example, I would re-organize Intel's web site. I haven't checked recently, because we stopped buying Intel motherboards. But before, the part numbers for Intel motherboards were not immediately available. To get the part numbers necessary to place an order, it was necessary to jump through hoops with an online registration, and then have an Intel employee direct you to the proper web page. The result was that it was difficult to call a distributor and order Intel motherboards. The salesman at the distributor would offer plenty of choices, but there was no way to know which one to choose unless you had done Intel's hoop jumping.

Second, I would change the culture at Intel. Business is NOT a place for top executives to act out their anger. Top managers should go into therapy if they need to deal with their anger. They should NOT make their anger a problem for the corporation. But that's what has been happening. I think it was perhaps 15 years ago that I became acutely aware of this. I could give several examples, and I've heard others from Intel employees.

Third, very important, I would take good care of the technical staff. At present they are not treated sufficiently well. If I were Intel CEO and I saw a floor that needed to be sweeped to give the technical staff a better work environment, I would sweep it myself if necessary. Intel's business DEPENDS on creativity. It's difficult to be creative in a poor social environment.

There's more, but this is is not a complete proposal, of course. It is just a Slashdot comment.

Google says you can encrypt your data with an 8 character password so that "not even Google" can see it [1].

Quick math. 26 lower case letters + 26 upper case + 10 numeric characters. (should cover most users)
62^8 possibilities. Google probably has about 100,000 servers [2], so that's about 2 billion combinations per server [3] - chump change.

AYPABTG.

8 character passwords work because servers can throttle bogus logins - few seconds delay after 3 failed attempts for example. There's very little security against an "adversary" like Google who is able to try all the combinations unabated.

Thanks for playing!

[1] http://www.google.com/tools/firefox/browsersync/fa q.html#q10
[2] http://www.intel.com/cd/ids/developer/asmo-na/eng/ 202679.htm?page=3
[3] http://www.google.com/search?hl=en&lr=&q=(62%5E8)% 2F100000

Not when UWB's output power is limited so that its range is only 30 feet.

The hardware just isn't there. You throw on a game with texture and lighting on a mac mini with that Intel GMA 950 and I'll show you a mac mini that is doing t&l in software. It absolutely dies. Same demo/game, crap graphics chip.

http://www.intel.com/products/chipsets/gma950/:
"Microsoft* DirectX* 9 Vertex Shader 3.0 and Transform and Lighting supported in software through highly optimized Processor Specific Geometry Pipeline (PSGP)"

Wow, through software. What's the other option, it just exits? Hardware > Software > Just quits

I'm sorry for the flame. Demos go from 30fps to 4fps. It's just really, really not even worth discussing T&L on these mac minis or Macbooks (the Macbook Pro has real vram and a "real" gpu).

A regular Pentium 4 draws on the order of 65 watts. At $0.12/kWh, that's .065kW*24h/d*.12$/kWh*365d/y = $68 dollars per year. Even if the custom NAS's CPU runs on air, you'd have to run the Pentium system for a long time to make up the difference in cost.

Don't they kidnap Paul Otellini and tell him that 18 months ago he screwed their wireless networking domination to allow free standards?

Don't worry, Craig Barrett will rescue him, while Andy Grove masterminds the coverup, all while on a "Moore's Law " tour of east Asia.

Don't they kidnap Paul Otellini and tell him that 18 months ago he screwed their wireless networking domination to allow free standards?

Don't worry, Craig Barrett will rescue him, while Andy Grove masterminds the coverup, all while on a "Moore's Law " tour of east Asia.

Don't they kidnap Paul Otellini and tell him that 18 months ago he screwed their wireless networking domination to allow free standards?

Don't worry, Craig Barrett will rescue him, while Andy Grove masterminds the coverup, all while on a "Moore's Law " tour of east Asia.

(a) lend themselves extremely well to optimisation

(b) lend themselves extremely well to incorporation in subroutine libraries

(c) tend to isolate the most compute-intensive low-level operations used in scientific computation

SGEMM

If you read the article, you will find (among others) a reference to a operation called "SGEMM". This stands for Single precision General Matrix Multiplication. This is the sort of routines that make up the BLAS library (Basic Linear Algebra Subprograms) (see e.g. http://www.netlib.org/blas/). High performance computation typically starts with creating optimised implementation of the BLAS routines (if necessary handcoded at assembler level), sparse-matrix equivalents of them, Fast Fourier routines, and the LAPACK library.

ATLAS

There is a general movement away from optimised assembly language coding for the BLAS, as embodied in the ATLAS software package (Automatically Tuned Linear Algebra Software; see e.g. http://math-atlas.sourceforge.net/). The ATLAS package provides the BLAS routines but produces fairly optimal code on any machine using nothing but ordinary compilers. How? If you run a makefile for the ATLAS package, it may take about 12 hours (depending on your computer of course; this is a typical number for a PC) or so to compile. In this time the makefile will simply run through multiple switches and for the BLAS routines and run testsuites for all its routines for varying problem sizes. And then it picks the best possible combination of switches for each routine and each problem size for the machine architecture on which it's being run. In particular it takes account of the size of caches. That's why it produces much faster subroutine libraries than those produced by simply compiling e.g. the BLAS routines with an -O3 optimisation switch thrown in.

Specially tuned versus automatic?: MATLAB

The question is of course: who wins? Specially tuned code or automatic optimisation? This can be illustrated with the example of the well-known MATLAB package. Perhaps you have used MATLAB on PC's, and wondered why its matrix and vector operations are so fast? That's because for Intel and AMD processors it uses a specially (vendor-optimised) subroutine library (see http://www.mathworks.com/access/helpdesk/help/tech doc/rn/r14sp1_v7_0_1_math.html) For SUN machines, it uses SUN's optimised subroutine library. For other processors (for which there are no optimised libraries) Matlab uses the ATLAS routines. Despite the great progress and portability that the ATLAS library provides, carefully optimised libraries can still beat it (see the Intel Math Kernel Library at http://www.intel.com/cd/software/products/asmo-na/ eng/266858.htm)

Summary

In summary:

-large tracts of Scientific computation depend on optimised subroutine libraries

-hand-crafted assembly-language optimisation can still outperform machine-optimised code.

Therefore the objections that the hand-crafted routines described in the article distort the comparison or are not representative of real-world performance are invalid.

However ... it's so expensive and difficult that you only ever want to do it if you absolutely must. For scientific computation this typically means that you only consider handcrafting "inner loop primitives" such as the BLAS routines, FFT's, SPARSEPACK routines etc. for this treatment, and that you just don't attempt to do that yourself.

Currently, the biggest difference you can make comes from the CPU. Go with a P4, and you might as well abandon power consumption as a design constraint. On the opposite end of that spectrum, if you don't need a lot of horsepower, the Via Epia boards... For my "real" machines, I currently have Athlon64s... A Pentium-M would give more bang-per-watt, but they cost a hell of a lot more. And as I mentioned, the next-gen Core Duos look very promising.

Note that Core 2 Duo (Merom) is supposedly compatible with the current Core Duo chipset and socket. However, it might be worth waiting 2-3 months for the really good Core 2 Duo CPU/motherboards.

I haven't seen any reasonable-priced full-ATX motherboards announced for Core Duo. It seems like there should be more Core Duo desktop motherboards since Intel makes a desktop chipset specifically for low-power Core Duo desktops (945GT).

...it also sounds strangely familiar, somehow...

From TFA (emphasis mine):

"M2Z's goal is ... provide free high speed connections to 95 percent of U.S. consumers without any recurring fees. This is a grand undertaking."

There should be dozens of 3rd tier Core motherboards out there by now for people to use to make their own PCs. Instead there is one, and it costs $300. Intel is not doing a great job indicating to the market which way they are headed.

I think Intel has purposely restricted their promotion of Core Duo for desktops. The only type of Core Duo desktops that I've seen promoted (by Intel) are pre-built, small, quiet ViiV desktops. Since Core Duo is designed to be a notebook processor (low TDP), maybe they think the best way to promote its desktop use is to show off how fast and quiet it can be in tiny computers.

If they wanted to promote it as an all-around desktop CPU, they could have easily made faster versions for bigger computers. However, Core Duo's current weaknesses (32-bit and average floating point performance) would be more exposed when compared to desktop processors. I think they're waiting for Conroe, with its EM64T and improved floating point, to make their big bang on the desktop.

That said, I don't understand why more motherboard manufacturers are not offering ViiV-oriented motherboards like Asus's. Intel has been offering a desktop chipset for Core Duo for a while now. AOpen does offer a Tivo-sized Core Duo "media center" barebones using that chipset, but it's not cheap (around $400 without CPU, hard drive, and memory). It looks fantastic for building a HTPC with no external parts, though.

I assume the "new" ultra-low power chips the OP is talking bout are the Intel will have the L2400 running 1.6Ghz at 15W and the U1400 running 1.2 GHz ~6W http://www.intel.com/products/processor_number/bod y_view_core.htm AMD has for some time now had available a 1GHz processor running 6W, and a 1.4GHz running at 16W as part of their Geode line. http://www.amd.com/us-en/ConnectivitySolutions/Pro ductInformation/0,,50_2330_9863_10837,00.html So AMD was ahead of the game on this one, but their chips are getting "consistently hotter"? I just think that comparison reeks of easily refutable fanboy dogma.

Interesting, their datasheet seems to be out of date. They don't mention the low power part.

On the other hand, you're still looking at 62W for a single core part compared to 55W for a dual core. And looking a little further, it looks like AMD has added a new "EE" (energy efficient) line which run at 1.4GHz and draw only 30W.

And the memory bandwidth numbers aren't particularly interesting. They're advertising 10.6GB/s on desktop Pentium 4 boards as well (see the I975X chipset, for example), and the AMD AM2 socket chips coming out this month will have a 12.8GB/sec memory controller on-die.

The memory bandwidth provided by AMD solutions also scales with the number of processors, since you're adding an additional memory controller per processor. There are boards on the market now - like the Tyan Thunder K8QW - which boast 51.2GB/s aggregate memory bandwidth.

And the reason for different models is because, unlike Intel processors which plug into a shared bus, Opterons are have dedicated Hypertraport links between the processors. There's three lines - 1xx which has a single link and is designed for uniprocessor configurations, 2xx which has two links and is designed for dual processor configurations, and 8xx which has three links and supports four and eight processor configurations. On current parts, the individual links at 8GB/sec, which means 8xx parts have 24GB/sec in aggregate bandwidth.

That's going to change some next year, though, when they move to Hypertransport 3 (current chips use 1), which scale to a whopping 41.6GB/sec per link.

You should probably call up Intel and tell them to fix their spec sheets, because they seem to think that they're selling Itaniums clocked from 900MHz to 1.66GHz and draw anywhere from 90-122W.

This is where I got the info from: http://www.intel.com/products/processor/itanium2/i ndex.htm

62, 99, and 130 Watts, and check out the new memory bandwidth numbers: 10.6 GB/sec.

I am not brand loyal, but at this time I prefer Intel over AMD. One thing I like about Intel is that they are a more mature company and offer a wider range of products. Intel makes damn good compilers, motherboard chipsets, CPUs, NICs, etc. AMD pretty much just makes processors at this time. And one thing that confuses me about AMD, is why do they have so many different processors? They have a web form on their site to do a search on them. I understand there is a difference between the 1xx opterons that are single CPU boxes and 2xx models that are SMP, but then there is the 4xx, and 8xx models as well. I don't know the down and dirty here, I'm not a computer engineer or chip designer, but to me its always been SMP or not SMP. I've found that looking for an AMD chip is almost impossible, and I guess its just worthwhile to benchmark some tests on a vendors box and see what happens. There are 191 models of Opterons listed on their website today. That is nuts!

http://www.intel.com/network/csp/resources/white_p apers/4631web.htm This is a standard for sending fax data over VoIP (or any packet) networks. Many VoIP systems have DSPs that can detect fax tones and automatically switch to T.38 over RTP mode where non-lossy fax data is bundled into RTP packets. I know that at least the Linksys WRTP54G wireless Vonage router supports T.38 and other models may do as well.

Subject: Important Question

Will the current Generation of Macs meet these requirements?

Actually, you can use IPMI to power-off, power-on, reboot, reconfigure the BIOS, etc all done remotely. More info at: http://www.intel.com/design/servers/ipmi/

look at ... how many years ago "Moore's Law" said a 4.5 GHz replacement should have been available

Nit: Moore's law has nothing to do with clock speed, it's about transistor count. The "law" still holds btw.

You can get the specifications for this graphic card here:
http://www.intel.com/products/chipsets/gma950/

And also a nice review here:
http://www.extremetech.com/article2/0,1697,1821811 ,00.asp

Why would I want a laptop to have a great graphics card? To eat more power and give me less batery life? Do you really need to play doom 3 in a 13' laptop?

Seriously, that graphics card helps to lower the price and power consumption and is more than enought to decode any kind of DVD/video and play any game that doesn't requires too much graphic power. It even supports HDTV, and several acceleration capabilities (pixel shaders and all that crap).

If I have to choose between a cheap and power-savvy low-end card and a high-performance but expensive and power-hungry one, I choose the first. For 1.099 $ it's exactly what I want.

Paul is that you ?????

http://www.intel.com/pressroom/kits/bios/otellini. htm

Ex-MislTech

And hopefully correct.

Well, Intel seem to be saying that they're not yet shipping Intel Core Microarchitecture chips ("We expect processors based on Intel Core microarchitecture to start shipping in the third quarter of 2006."), and that the Core 2 chips use the Intel Core microarchitecture ("the Intel® Core(TM)2 processors for desktop and mobile computers are based on the Intel® Core(TM) microarchitecture"), so the Core 2 chips are the first "Intel Core" chips to use the new microarchitecture. (Whether they'll be the first chips at all to use it depends on when the first Intel Core microarchitecture Xeon chip (Woodcrest) ships.)

And hopefully correct.

Well, Intel seem to be saying that they're not yet shipping Intel Core Microarchitecture chips ("We expect processors based on Intel Core microarchitecture to start shipping in the third quarter of 2006."), and that the Core 2 chips use the Intel Core microarchitecture ("the Intel® Core(TM)2 processors for desktop and mobile computers are based on the Intel® Core(TM) microarchitecture"), so the Core 2 chips are the first "Intel Core" chips to use the new microarchitecture. (Whether they'll be the first chips at all to use it depends on when the first Intel Core microarchitecture Xeon chip (Woodcrest) ships.)

So, we can expect the intel version of the G5 desktops out in July or latest August 2006, since that is when the next Apple Worldwide Developer Conference (WWDC) takes place.

Even though Pentiums and Xeons have shared the same core since P6 (Pentium 2), Intel has reserved some workstation/server features (like multiple processors) for the Xeon platform (AMD does the same for Athlon/Opteron).

So if Apple still wants dual-processor options for the PowerMac's successor, then they will very likely use Woodcrest (Xeon). However, the emergence of dual-core and the availability of Intel "Extreme" CPUs might mean Apple does not need more than one processor anymore for their "pro" desktops. Also, Intel offers "low end" workstation platforms that use Pentium CPUs with worstation chipsets (ECC memory, worstation graphics cards, PCI-X, PCIe x4/x8). Example: E7230 chipset.

My revised predictions: iMac will use Core 2 Duo. PowerMac (Mac Pro?) will use Core 2 Extreme (no more dual-processor) with a workstation chipset (ECC, workstation graphics cards, PCIe x4 and x8 slots). Xserve will use dual-processor Xeon (Woodcrest). I'm hoping Apple will offer another non-pro desktop (besides iMac and Mac mini) that uses Core 2 Duo.

The one you've got there is over 5 months old. Try this one.

Quote: With its own unique brand name, the Intel® Core(TM)2 Extreme processor, also to be launched later this year, will be the world's fastest processor and the new first choice of gamers and multimedia professionals worldwide. Core2 Yet it will also be the world's premiere choice for ripping off your wallet.

Last time there was a move from G4-based devices to Core Something devices (Solo or Duo) was the Mac Mini. And the price went up quite a lot for that.

I'm often wrong on my Mac predictions, but I think Apple might offer a low-priced MacBook based on the new Celeron M CPU based on the Core Solo (Yonah) core (model number 420, 423, and 430). These Celerons differ from Core Solo by having less L2 cache (1MB), slower bus (533MHz), and no Enhanced SpeedStep. I think the lack of SpeedStep is the only thing a MacBook buyer might miss because it significantly improves battery life. However, many other parts of this CPU architecture and chipset were designed with power-saving in mind. If battery life can be made "good enough," these new Celerons should offer great performance for a low price.

I think it would make a lot of sense for Apple to offer a new $500 (or less) Mac mini using this new Celeron M CPU. SpeedStep is not needed on the Mac mini, of course. They also use the same socket as Core Solo/Duo.

"Then, once I get the system up, the network keeps getting thousands of "too many interrupts" - Even unplugged from the switch!"

http://www.google.com/search?hl=en&lr=&q=e1000+int erruptthrottlerate&btnG=Search
ftp://download.intel.com/design/network/applnots/a p450.pdf
http://lwn.net/Articles/152989/
http://sourceforge.net/projects/e1000

True, but think of all the money and resources lost by Intel for a rare error that would not effect the vast majority of the customers using the chip. How many perfectly decent chips were just thrown away, passing the cost onto the consumer who had to make up for the money lost in remanufacturing these chips. Here is a report done by Intel on how often an average user might see an error: http://www.intel.com/support/processors/pentium/fd iv/wp/6.htm It's certainly bad PR for AMD and they will most likely offer an exchange program like Intel, but the practical need for exchanges isn't really there (if what I am reading in other comments is correct).

The XScale PXA-255 dissipates ~600mW.

http://www.intel.com/design/pca/products/pxa255/te chdocs.htm

Believe it or not, sometimes the marketing message trails what engineering does, especially when there is less whining to be had. The prototype used a less efficient 486 core. But it was still embedded.

Incredibly enough I own a Latitude that is about 8lbs and 7510t BB. Good call.

Enjoy utility deregulation. :-)

Please, no more myths about Intel power hungry designs. Core Duo fixes this for good:

AMD Athlon64
Freq: 2.0 GHz
Tcase: 70 degC
TDP: 89 W

http://www.amd.com/us-en/assets/content_type/white _papers_and_tech_docs/30430.pdf

Intel Core Duo Destkop (Yonah)
Freq: 2.16 GHz
Tcase: 100 degC
TDP: 31 W (scale to 70deC like AMD puts you at about 20W).

http://download.intel.com/design/mobile/datashts/3 0922102.pdf

About the chip, you can use cheap p2 chips that take 10 watts. It's actually not completely stupid. Maybe have the controller monitor prices to take advantage of on-the-fly pricing. The plant I work at pays continually variable pricing. Intel even has info for embedded systems.

I'm not really a fan of this sort of design - it seems to duplicate the purpose of the existing kernel/userspace security architecture, but I can appreciate the pickle we're in with de-facto standard kernels that allow anything to be loaded into them. Windows Vista 64 bit requires all kernel drivers to be signed: correctly so, in my opinion, but this doesn't help the huge 32 bit userbase today.

These might help :

http://www.amd.com/

http://www.intel.com/

The general consensus is that the 12" Macbook Pro will be released when Merom hits the market. With the heat problems the Macbook Pros have been having, a 12" wouldn't be viable at this time.

Here's some links if you want em':

• Intel Core Processor Number Details (includes Low Voltage Core Duos)
• Lenovo ThinkPad X60s review (March 24)
• First Look: Lenovo ThinkPad X60s (April 19)

Sounds neat; just out of curiosity, what are you using for the virtualization?

Parallels Workstation. Takes full advantage of Intel VT, too.

And perhaps I'm revealing my ignorance here, but how does the guest OS on a virtualized system handle networking? There must be some sort of psuedo-device driver that you install so that it can talk to the virtualized 'hardware,' or else the virtualization software must emulate some kind of commonly-supported networking hardware. I guess I'm just curious what you have to do to the guest OS, if anything, to get it to work inside the sandbox.

It can either be NATed behind the host OS, or it can have its very own IP on the same network as the host machine. The virtual machine has an emulated network interface with its own MAC address. You don't have to "do" anything in the guest OS. It's a very generic network card that has had drivers available (and in my case, always included) in every OS I've installed so far.

Right now all my equipment (Mac stuff, anyway) is PPC based; however I find the whole virtualization concept really intriguing and I'm hoping that by my next round of upgrades, it'll be sufficiently mature to make dual-booting (or having a separate PC for Linux connected with a KVM switch, my current solution) unnecessary.

Well, you'll likely be very happy then, because, even at this early stage, that's exactly what I'm using it for right now. All on one machine that I can carry around with me.

- FireWire 800 (9-pin) is included, in addition to FireWire 400 (6-pin) (so no, FireWire, and particularly FireWire 800, is not dead, as some like to continually predict)

- 3 USB 2.0 ports are included; 2 on the left, 1 on the right

- The left side ports are: power, 2 USB 2.0, analog and digital optical audio in and out, ExpressCard/34; the right side ports are: DVI (supports VGA, S-Video, composite), 10/100/1000 ethernet, FireWire 800, FireWire 400, 1 USB 2.0, security port

- An 8x dual layer SuperDrive is included (unlike the 15" MacBook Pro)[1]

- While this is known by many, it bears repeating that the wireless chipset in all Intel-based Macs supports 802.11a/b/g, though Apple doesn't advertise 'a'

- The 1680 x 1050 resolution of the 17" display is the same as many desktop 20" widescreen LCDs such as the Apple 20" Cinema Display and the 20" Dell 2007WFP

- Retail $2799, Education/government $2599 with 2.16 GHz Core Duo, 1 GB RAM, 120 GB 5400RPM Serial ATA drive, 256MB ATI Radeon x1600, and 8x dual layer SuperDrive

- For detailed specs, see here

I'd also note that for some people who might think that the 15" MacBook Pro looks like a bad deal next to this, the 17" is simply too large for many people, and many of those same people have no need for the faster dual layer SuperDrive, nor for FireWire 800.

And the Apple we site does not have to explicitly say it for us to know that, yes, of course the 17" MacBook Pro will support "Boot Camp" (and triple booting[2]), which is simply an umbrella marketing name for a collection of technologies that support booting Windows on Intel-based Macs:

- A Compatibility Support Module (CSM, BIOS compatibility layer) for EFI: this is already a non-beta, supported component of the recent rounds of firmware updates for Intel-based Macs, which the 17" MacBook Pro will ship with

- The ability to live-resize partitions on a GPT formatted volumes: this is already a non-beta, supported component of "diskutil" as of 10.4.6

- A collection of Windows drivers for the hardware in Intel-based Macs: almost all of these are non-beta, preexisting third party drivers

- A setup assistant that brings everything together: this is the only part of the solution, from a technical standpoint, that is "beta"

[1] Some may note that the new 17" MacBook Pro, at the same thickness of the 15" MacBook Pro (1.0"), includes an 8x dual layer SuperDrive versus the 4x single layer drive in the 15" model. It might be recalled that the reasoning for not including a faster, dual layer SuperDrive in the 15" MacBook Pro was because of the necessary space not being available inside the case; the 15" MacBook Pro could only use a 9mm tall mechanism as opposed to the 12mm mechanism currently required for dual layer capability and the greater speed. How, then, can the 17" MacBook Pro (or even the previous 17" PowerBook), at the exact same thickness, include this drive? Does this mean Apple was holding back? Is the 8x DL drive due in a 15" MacBook Pro imminently? The answer is no: the reason why the drive didn't (and still doesn't) fit in the 15" MacBook Pro is because the wider trackpad mechanism Apple chose to use encroaches internally on the space needed for a 12mm drive by about 1/8" laterally. However, this is not the case on the 17" MacBook Pro.

[2] Who wants to dual boot, much less triple boot? I'd rather have all of my environments running side by side in virtualization. And yes, I know there are some specific reasons people may want to dual boot (such as games for native 3D graphics support), and that's fine...but other than for those specific tasks, who would really prefer dual/triple booting over virtualization, especially given the excellent benefits Intel VT now offers for virtual machines?

Actually you don't need a CoreDuo to be VIIV complaint, a Pentium D or EE would also fit within spec: Viiv requirements