Slashdot Mirror

I don't remember everything it is used for, but it is more then just the LED ;) One thing it does is sit ready for a "power-on" command. Which you would need your battery in the laptop for anyway. Intel's site explains it in full somewhere. I do know that the government supposed to lower it to 1-watt but it has not gone into effect yet. Intel explains in the 1-watt initiative: The 1-watt measurement applies to the S5 standby mode, in which the computer is plugged into the wall, but the power is off on the computer. There is some more info on Intel's site about the 3,2, and 1 watt power consumptions. But, here is one about the bill to lower it to 1-watt: intel article

Note: most documents refer to "power-down" as "stand-by", and windows/linux "stand-by" modes as "sleep".

WTF? Please, just have a look at some IA-64 assembly code! It's NOT pretty, especially if you want it to go fast. You've got to do the whole explicitly parallel thing, manually pack together independent instruction according to what pipelines you want to run them in.

Itanium is NOT a RISC machine like Sparc, not in the least. Sparc is much more closely related to x86 than it is to IA-64. The Itanium is a VLIW chip, or EPIC in Intel-speak. It's a whole different animal altogether.

FWIW, here's a brief article where Intel talks about implementing a bubble-sort in IA-64 assembly vs. the original C. In particular, they start with the code that the Intel C compiler generates and optimizes it. Their final, optimized version of the algorithm is on page 5, and it's anything but easy.

uh, maybe when the pentium came out but a 486 sure as hell couldn't do 64 bit math.

hrm...1997...Pentium II was released then...

quite the hacker you were with your 1989 486.

hrm...1997...Pentium II was released then...

quite the hacker you were with your 1989 486.

(I'm not trying to sound like I have a big ego but here goes)


First we need a protocol somewhere between Bluetooth & 802.11b (I'm sure it exists, I am just to lazy to look for a link), that offers low-bandwidth + medium range (50m) and ultra-low battry usage (1 week of usage).

Next we can embed network aware sensors that WE can program. (level of detail to provide and/or services offered)

(You are a middle-aged computer geek who loves Thai food, hockey, and biking) You can program the device to "broadcast" your likes & dislikes this can also help when looking for a job, post your resume and be alerted whenever a company is hiring. Meanwhile companies that want to sell to your demographic can advertise to you when you get near (50m) the shop. (Like Minority report but less invasive because you can turn all of this off if you want.)

It can store all of your personal media to be displayed whenever/wherever you are.

Devices like this with an LCD in the glasses combined with something like this and a terrabyte storage in a form-factor the size of a deck of cards.

It can help with meeting other individuals that share your same tastes, AND store vital data about people you meet (incase your memory is so bad that you can't remember information ie. wifes's name, kids birthdays, etc..)

But most importantly, it's up to YOU to decide how much or how little you provide.

The history of Dragon Naturally Speaking and ViaVoice is pretty interesting. Dragon Naturally Speaking was started by a company called Dragon, several years ago. Originally, they marketed a program called Dragon Dictate, which was designed for command and control, more so than for document dictation. They stumbled on a new technology using phonemes, and were able to create Naturally Speaking, which, like the name, allows someone to speak at normal conversational speeds and the software will still recognize it (Dragon Dictate pretty much made you speak like William Shatner on a bad day).

Fast forward a little...Naturally Speaking got sold to a company called Scansoft when Dragon went under (I forget why). Scansoft, within the last year bought ViaVoice from IBM, which had dropped sales and further development of their Linux version of the software, which is now nowhere to be found.

Ultimately, Scansoft pretty much controls the market in terms of Voice Dictation.

However, there are some interesting things still going on. For instance, Intel has been working on the Audio/Visual Speech Control/Recognition engine, which is part of Intel's OpenCV (face recognition stuff).

Ultimately, try out this link. There's quite a bit of information there.

The history of Dragon Naturally Speaking and ViaVoice is pretty interesting. Dragon Naturally Speaking was started by a company called Dragon, several years ago. Originally, they marketed a program called Dragon Dictate, which was designed for command and control, more so than for document dictation. They stumbled on a new technology using phonemes, and were able to create Naturally Speaking, which, like the name, allows someone to speak at normal conversational speeds and the software will still recognize it (Dragon Dictate pretty much made you speak like William Shatner on a bad day).

Fast forward a little...Naturally Speaking got sold to a company called Scansoft when Dragon went under (I forget why). Scansoft, within the last year bought ViaVoice from IBM, which had dropped sales and further development of their Linux version of the software, which is now nowhere to be found.

Ultimately, Scansoft pretty much controls the market in terms of Voice Dictation.

However, there are some interesting things still going on. For instance, Intel has been working on the Audio/Visual Speech Control/Recognition engine, which is part of Intel's OpenCV (face recognition stuff).

Ultimately, try out this link. There's quite a bit of information there.

When reliable, open source centrino support exists.

here is intels official site: Intel LCoS

The article in NYtimes made claims that this would produce sub 1000 dollar RPTVs by Christmas 04' I think this is a bit optimistic however it does bode well for us avgeeks.

LCoS is Liquid Crystal on Silicon. The best way to describe it would be a cross between DLP and LCD without the colorwheel and micromirrors of a DLP system.
I would also like to note that this technology has nothing to do with flat panel displays(LCD/Plasma)
More Links
Yahoo Article
CNET Article

From my description of SMT, you might conclude that it is able to make the fullest use of CPU core resources of any multithreading approach. Then why are some benchmarks slower with HyperThreading enabled? I have three ideas. First of all, since OSes tend to treat a single CPU with hyperthreading like a dual-CPU system, they may schedule lower-priority processes on one of the threads, resulting in less total CPU time available for the main benchmark process. Secondly, there is some overhead to multiplexing all of these resources amongst two threads, so perhaps Intel had to "tone down" superscalar operation or memory queues to meet clock timing when in HyperThreading mode (I know this is vague). Thirdly, perhaps in ostensibly non-HyperThreaded mode the second thread is actually used for speculative execution, and hence slightly boosts the performance of a benchmark.

What, like the Intel 875P and 865 family chipsets have had for the last 8 months or so, you mean?

What, like the Intel 875P and 865 family chipsets have had for the last 8 months or so, you mean?

For those more technically inclined I would suggest reading Intel's Hyper-Threading Technology Architecture and Microarchitecture whitepaper instead.

I was hoping for the ipod mini to be without a HD, but this one still has a HD (moving parts), so NOT good for running (eventhough apple says its ok for running). I guess I will have to stick to my oldskool Intel Pocket Concert.

Every Intel processor since the Pentium has had 64-bit memory access and datatypes. Any SSE chip can also support 128-bit datatypes for certain operations. Making the Pentium '64-bits' only requires 64-bit memory addressing(which entails a 64-bit program counter and stack pointers). x86 instructions already range from 1 to 10 bytes in width(14 bytes including the immediate field) so memory bandwidth will not change as a result of the new instructions existing.

Every Intel processor since the Pentium has had 64-bit memory access and datatypes. Any SSE chip can also support 128-bit datatypes for certain operations. Making the Pentium '64-bits' only requires 64-bit memory addressing(which entails a 64-bit program counter and stack pointers). x86 instructions already range from 1 to 10 bytes in width(14 bytes including the immediate field) so memory bandwidth will not change as a result of the new instructions existing.

So, you'll have a 64-bit module sitting on top of your 32-bit CPU?

I've been speculating (here and elsewhere) that this stackable thing is not going to be Intel's next big thing. I believe that the stacked module will simply contain NVRAM and not a 64-bit coprocessor. Why NVRAM? Well, it opens up some interesting possibilities. For example, if you had enough NVRAM on-chip (or reasonably close in terms of latentcy and bandwidth), you could simply shut down portions of the processor on-the-fly to save power. You could also stick the entire operating system on the stuff. The possibilities are amazing. If you haven't looked already, see my journal for much information on the subject as it relates to Intel.

Of recent interest are some presentations by Intel on NVRAM. Of interest is that they've announced that they've found that OUM will take them beyond transistors in one presentation while another presentation actually shows a transistorless cell that is quite simple (two electrodes and a programming material sandwiched in between).

A transistorless storage device could be the piece that stacks onto the P5.

So, you'll have a 64-bit module sitting on top of your 32-bit CPU?

I've been speculating (here and elsewhere) that this stackable thing is not going to be Intel's next big thing. I believe that the stacked module will simply contain NVRAM and not a 64-bit coprocessor. Why NVRAM? Well, it opens up some interesting possibilities. For example, if you had enough NVRAM on-chip (or reasonably close in terms of latentcy and bandwidth), you could simply shut down portions of the processor on-the-fly to save power. You could also stick the entire operating system on the stuff. The possibilities are amazing. If you haven't looked already, see my journal for much information on the subject as it relates to Intel.

Of recent interest are some presentations by Intel on NVRAM. Of interest is that they've announced that they've found that OUM will take them beyond transistors in one presentation while another presentation actually shows a transistorless cell that is quite simple (two electrodes and a programming material sandwiched in between).

A transistorless storage device could be the piece that stacks onto the P5.

Intel is currently working furiously on a cellular chip that will seamlessly roam to WiFi networks. They also want it to carry prodigous amounts of data. Known as the Digital Briefcase specification, any compliant PC will automagically recognize the phone and allow you to log into the PC as if it were your own. Mail, favorites, documents/music and even wallpaper and settings will appear seamlessly. Check out my sig for more...

POTS will die a quick death unless the big TelCos start lobbying for taxes. Slashdotters move these companies up on the list of Evil entities.

that nothing much really came of it.

Of course nothing ever came of it, because it was killed off.

Only a small percentage of machines have it, and those that do generally turn it off. Therefore no one tries to use it. If it had survived and been incorporated into all new PC's then turning it off would be much like trying to surf the web with cookies and Javascript turned off. Possible to do, but it becomes a massive pain in the neck.

That may well be a good indication of what will happen if they try anything similar this time round...

I certainly hope you're right, but they are putting up a huge smoke screen advertizing unrelated benefits and slipping in this poison pill. They are also getting very sneaky and scrubbing out any direct refference to Trusted Computing from all related projects.

This is at least the third time I've caught a hidden Trusted Computing connection in a slashdot story - even the techies and geeks here are completely missing the connections.

In this case they didn't scrub hard enough and I can directly prove the connection. First of all Google's cache has a link to an Intel paper titled TCPA Technical Overview for EFI. The link is dead - that paper must have "accidentally" fallen down the Memory Hole. I did find another Intel paper on EFI, on page 20 it states that EFI's Security Support Protocol 'maps well' to the the Trusted Computing Groups's Trusted Platform Module. It is specificly designed to run on top of Trusted hardware. If you look at page 21 it even has support for a Remote Take Ownership ability.

Another story was on the Content Reference Forum. Their technical specs go into great detail on "Rights Expression Language" and "Contracts". They clearly expect these contracts to be enforced, they state that thier system is designed to run on top of some sort of enforcement system, their system is an exact match for Trusted Computing, they use all of the key Trusted Computing terminology, the group is a successor to the failed Oasis group which was working on Trusted Computing, any direct verifiable link to Trusted Computing is conspicuously missing.

The really scary one it Cisco's Network Admission Control system. Every single news site that covers it hails it as Cisco declaring war of viruses and worms, or Cisco blocking viruses at the router. However these routers do nothing to block viruses. What they do is deny an internet connection to any computer that is not compliant with the security policy. The computer must be running a Cisco Trust Agent, that Trust Agent then verifies that the computer is running (or not running) any specific software, souch as an approved firewall, virus scanner, OS, and OS patch level. Again, the system is an exact match for Trusted Computing, they use all of the key Trusted Computing terminology, Cisco is a founding member of the Trusted Computing Group, yet after spending several hours researching the system I still can't find a single concrete link proving it is based on Trusted Computing.

I'm not sure whether to be elated that Trusted Computing is getting a bad reputation such that they need to make such efforts to scrub out any reference to it, or whether to despair that they have succeeded so well that even Slashdot is completely oblivious to the connections.

-

First off - if you haven't already, read the EFI specs: http://www.intel.com/technology/efi/index.htm

This has a good overview of what efi is and entails, as well as the specifications for it.

There are some good things about it - hardware drivers are easier to develop for it, it allows booting off of non-standard devices, and in some ways very similar to openFirmware. There is also linux support for efi (at least on IA64)

However, it has some serious drawbacks:
The potential is there to implement DRM, and attempt to "lock out" non-signed binaries, etc...
It requires a 100 Mb efi (FAT) partition (so it appears useless for diskless servers)
It appears to me at least that there are some potential serious security flaws to the implementation

Overall, EFI doesn't add anything that LinuxBIOS doesn't (except that EFI has been "blessed" by Microsoft), and it appears to be intel's way of locking in the BIOS market.

Link to presentation

The whole thing is designed so the OS can use EFI firmware interface or a legacy BIOS interface, so one platform can support both. Even though it adds some size to the firmware image, it makes the industry transition easier.

The download page requires a fake name and email, but you can skip that and get the latest version (1.10-001) here. (Total karma whore link: EFI homepage)

The license isn't actually too bad - it just says that if you provide them feedback, then you also grant them the right to implement your idea.

The download page requires a fake name and email, but you can skip that and get the latest version (1.10-001) here. (Total karma whore link: EFI homepage)

The license isn't actually too bad - it just says that if you provide them feedback, then you also grant them the right to implement your idea.

The download page requires a fake name and email, but you can skip that and get the latest version (1.10-001) here. (Total karma whore link: EFI homepage)

The license isn't actually too bad - it just says that if you provide them feedback, then you also grant them the right to implement your idea.

The excuse "WEll, current BIOS systems is just patch written upon patch written upon patch. ITs a mess."

But it works. Is an EFI system going to be markedly faster? When you tell me you are loading device drivers at the BIOS level, that tells me "No"- you are creeping the OS lower.

So whats the deal?
from Intel's EFI web site: Together, these provide a standard environment for booting an operating system and running pre-boot applications.

AHhhh! Running PRE-BOOT operations! This sounds like a lame way to shoe-horn in DRM or something similar onto my machine before it loads up.

Maybe I'm acting paranoid, but the slowest thing on my windows computer is WINDOWS, not the bios- that runs pretty fast.

I'm guessing, This.

No IDF keynote link (I didn't want a kiss anyway) but here's an Intel page about EFI:

http://www.intel.com/technology/efi/efi.htm

According to this info, EFI is a replacement for the old PC BIOS standards. For example they have a replacement for VGA called UGA, and it looks like they are improving the expansion card ROM access methods. I certainly agree with the idea of fixing the old broken cruft.

How is MS planning to use EFI to hurt Linux? I don't see anything obviously evil on this page.

Anything sold as a PC that doesn't run Linux properly will get flamed by the techs, will have trouble competing in the market, will have lots of "broken" returned systems, etc.

Stop faking things like the contents of this post.

Cowards are just that. Why don't you go purchase a clue? Also check out the story here for references on Intel's quality control. You also might want to check out who's designing your new Intel chips (can you spell alpha moron). For a story related to that, check here then of course there is the denial that the alpha design has influenced Intel here

You may also want to check out sources like EEtimes, shannon knows DEC/Compaq/HP and of course the intel web site. BTW VMS is running on IA64 hardware. Do you know what VMS is little girl?

AngryPeopleRule

AGP 8x can move 2.1 gigabytes per second (GB/s), according to Intel.

In the forward for "The Universal Computer" (by Martin Davis) there are a couple of quotes:

"If it should turn out that the basic logics of a machine designed for the numerical solution of differential equations coincided with the logics of a machine intended to make bills for a department store, I would regard this as the most amazing coincidence I have ever encountered."

Howard Aiken in 1956

Let us now return to the analogy of the theoretical computing machines... It can be shown that a single special machine of that type can be made to do the work of all. It could in fact be made to work as a model of any other machine. The special machine may be called a universal machine..."

Alan Turing in 1947

NOTE: In Mr. Aiken's defense, he is probably referring to a differential analyzer (which was an analog computer)

When I was in high school, my (supposedly) CS teacher read an article that stated, "the world would no longer need programmers". She attempted to persuade me from becoming a programmer because in the future no one would need programmers. It would be a dead profession. The year was 1994. Okay, she was half right (there won't be anymore jobs for a while, and they'll all go overseas...), but still.... You can't extrapolate. My teacher never would have imagined (or actually just read the other article) about the internet.

What if AI takes off? I think in the future even the soft sciences will become more computational. Look at fields like bioinfomatics or computational linguistics. There are all kinds of new areas opening up. The problem is that the world doesn't revolve around computers, but all the phenomena of our universe may be one really grand one. Programmers have to learn other skills. I see biologists, actuaries, and engineers (outside of EE/ECE) write code all the time. You need to attach an extra skill to your code.

All this just goes to prove, you shouldn't extrapolate about science or computing, unless your one of these guys:

Alan Turing
Albert Einstein
Kurt Godel
Nikola Tesla
Gordon Moore
Jules Verne

Of course, I'm extrapolating (and as you can guess, I'm not one of these guys...), so if you're a good philosopher you can safely ignore my post. Nothing to see here.... Carry on.

Most benchmark tests I've seen show icc producing faster code than gcc, which isn't surprising given that Intel's people presumably know preciesly what optimisations work best on their chips. Since icc produces faster code than gcc, compiling from course with icc will produce a faster system than compiling from source with gcc. Of course, there are fairly obvious downsides, too - using icc limits you in what platforms you can use, and (for those who care about such things) it's most definitely non-Free.

I'm looking at Mandrake's two-page year-end "Newsletter to Investors" and I can qualtitatively say that there's no way one could definitively say their financial health is improving.

I'm not sure if it's just rigorous US accounting standards have kept me from the harsh realities of international investing, but I have no idea about Mandrake's debt position, their return on investment, where exactly they're generating cash flow (operating, investing, or financing activities--they're very different) and about fifty other such ratios and line-items and on average fifteen pages of notes that are given for you or very easy to figure out on companies that follow U.S. Generally Accepted Accounting Principles (GAAP).

Compare the annual report of any publically traded U.S. company (here's Intel's annual 2002 report--the PDF is 102 pages) and you'll notice that a lot more information is given to investors and shareholders. We have, off the top of my head, the usual letter to shareholders from the CEO, some "PR fluff", the balance sheet, income statement, statement of cash flows, notes to consolidated financial statements, a signed auditors report indicating you can actually trust the data, segmental data, and thorough management discussion and analysis (MD&A) in which the company's head honchos actually talk about their company's financial health.

I'm not dissing MandrakeSoft in any way, I think their software is top-notch and with the disappearance of Red Hat from the consumer line I think Mandrake has a critical role.

I think it's important for /. readers to know that accounting and investing or disgustingly complex topics, and most shareholders don't read the annual reports or know enough to make sense of the number and subsequently get caught up in the bandwagon without a further analysis. It's very easy to lose your money in this market simply by not looking at the books.

Mandrake, for example, could be earning all their money from external financing and losing money from operations. That looks good on your income statement but if you don't check the statement of cash flows, you wouldn't know about that and you'd bee royally screwed when those external lenders come to collect. Plus, all I know about their debt situation is that they're in chapter 11--how much debt do they really have? I could think of a hundred other questions not answered by their newsletter.

Mandrake's "newsletter" does not give me the numbers I need to make that sound analysis.

Oh, and before some of you wiseguys respond to this, realise that Enron, et al. are the EXCEPTIONS, not the rules.

IA-32: A system based on a 450 MHz Intel(R) Pentium(R) II processor or greater, Intel Pentium 4 recommended # 128 MB (256MB recommended).

IA-64: A system with an Intel Itanium(R) or Itanium 2 processor or greater

Corrected link here

coriordan, please stop trying to convert the unwashed masses here at slashdot and lwn. Anybody who is at either site is well aware of GNU and their "philosophical" positions and has heard these regurgitated arguments before, and if we aren't members of the GNU/Cult it's because we've examined their positions and found Stallman and Co. to be frothing-at-the-mouth crazy. Your naive, ignorant, and overconfident opinions aren't the gospel you take them to be. Your sermons at LWN expounding the GPL and the GFDL to the non-believers are downright ludicrous (especially since you don't understand the actual natures of the licenses and software development models you discuss).

So intel releases a compiler. So you don't approve of its being protected by regular copyright law. Then a) refrain from buying it and b) sit down and shut up.

For those who have real interest in the compiler instead of fantasized complaints, Intel offers zero-cost downloads of both the c/c++ and fortran compilers for noncommercial use.

coriordan, please stop trying to convert the unwashed masses here at slashdot and lwn. Anybody who is at either site is well aware of GNU and their "philosophical" positions and has heard these regurgitated arguments before, and if we aren't members of the GNU/Cult it's because we've examined their positions and found Stallman and Co. to be frothing-at-the-mouth crazy. Your naive, ignorant, and overconfident opinions aren't the gospel you take them to be. Your sermons at LWN expounding the GPL and the GFDL to the non-believers are downright ludicrous (especially since you don't understand the actual natures of the licenses and software development models you discuss).

So intel releases a compiler. So you don't approve of its being protected by regular copyright law. Then a) refrain from buying it and b) sit down and shut up.

For those who have real interest in the compiler instead of fantasized complaints, Intel offers zero-cost downloads of both the c/c++ and fortran compilers for noncommercial use.

• ICC supports two-pass compilation. So does GCC.
  
• ICC supports data prefetching. So does GCC.
  
• ICC can do code coverage testing. So can GCC.
  
• ICC can do interprocedural optimizations. Released GCC versions can't, but work is in progress.
  
• ICC can do automatic vectorization, GCC can't. Advantage ICC.
  
• ICC supports something called "processor dispatch". I'm not even sure what that is.
  
• ICC supports a number of optimizations that might be interesting if you happen to have an Itanium 2 sitting around.
  
• ICC supports parallel programming, GCC doesn't (not very well anyways). Advantage ICC.
  
• ICC's debugger supports "GDB syntax and a subset of GDB commands". Why not just use GDB?

The Linux kernel specifically requires gcc because it uses gcc extensions. But it should be possible to recompile the usermode code with the Intel's compiler. Even glibc may be worth trying (if you bought the compiler).

The Intel compiler has supported compiling the kernel with it for a few versions now, I believe. It may require a few kernel patches, but I think some of them have been accepted into the kernel tree. Intel has a web page about it.

My first thought was, "does this mean it can finally compile the Linux kernel?" But the website says "with a few kernel modifications, [icc] can build the kernel." Since gcc can compile it without modifications, doesn't this mean they are not 100% compatible? Also, there is no link to these patches anywhere, just this article on icc 7. Do you have to figure out the problems and fix them yourself?

Obviously there is other software to compile besides the Linux kernel, but since the icc is so tuned to the Intel hardware, and Linux interacts so directly with the hardware, people believe that icc would give great benefits to the kernel. At the very least, nothing can claim 100% gcc compatibility unless it can compile Linux unmodified.

Anyways, semiconductor process chemicals are treated (at least in Europe / North America, and they're getting better than they used to be in India, Taiwan, etc). HF is easily neutralized. Look at the environmental permitting at Intel's fab 12 in Arizona - waste discharge is a huge issue; they don't just dump HF out a pipe. Once neutralized, fluoride salts are about as toxic as toothpaste (toothpaste is approx 0.243% NaF, which is nasty stuff).

this articles a good read so take the time to go through it as it summarises innovation from the early internet years to date.

innovation. The trick is finding that one crazy idea. The problem with crazy ideas, though, is that for every one good crazy idea, there's a thousand bad crazy ideas

the eternal quest for an idea. you better start with a good idea. if you don't, no matter how hard you try it wont pan out.

the Internet community back then, the key technical people, didn't want the Internet to become easy to use or graphical, ... Only smart people could use the Internet ...so we needed to keep it hard to use

what other examples can you think of right now?... only smart people can use [insert you own example]

Mosaic started with 12 users in February 1993. It had 1,000 users within three or four weeks. About 10,000 users by spring. It was up to 1 million by early 1994

Posters who question why Andreessen has such prominence should reflect on this. No Mosaic (mozilla), no Microsoft Internet Explorer (IE Based on NCSA Mosaic code base licensed from Spyglass), no World Wide Web in the early to mid 90's. No doubt someone else may have invented the browser but how much longer would it have taken?

At first that makes you like a little bunny rabbit ... Everybody wants to play with you ....within a year ... fearsome competitors shooting at your head with high-powered ammunition

Larry, Sergi do you feel the hot breath of the MS juggernaut as you approach your IPO. Will google will be a repeat of Netscape/MS tussle?

Oracle database was a huge success ... Larry's spent the last 25 years trying to come up with the next product

it sure helps when the government (CIA) is your preferred backer. Why does oracle feel the need to keep trying to re-innovate or create the next best idea?

innovation comes from companies that are 2 years old, populated by 19-year-olds ... preposterous that Marc should think that innovation is .. the province of little entrepreneurial companies.

In fact it's both. The technical revolution was spurred on the back of the transistor. This was the combined effort of Bardeen, Brattain and shockley at Bell Labs - no small comany there ... but look at Intel, though a big company now, it was started with the (not so young) Noyce, Moore and Grove. What about the Linux kernel, third person shooters and that other search engine, Yahoo?

this articles a good read so take the time to go through it as it summarises innovation from the early internet years to date.

innovation. The trick is finding that one crazy idea. The problem with crazy ideas, though, is that for every one good crazy idea, there's a thousand bad crazy ideas

the eternal quest for an idea. you better start with a good idea. if you don't, no matter how hard you try it wont pan out.

the Internet community back then, the key technical people, didn't want the Internet to become easy to use or graphical, ... Only smart people could use the Internet ...so we needed to keep it hard to use

what other examples can you think of right now?... only smart people can use [insert you own example]

Mosaic started with 12 users in February 1993. It had 1,000 users within three or four weeks. About 10,000 users by spring. It was up to 1 million by early 1994

Posters who question why Andreessen has such prominence should reflect on this. No Mosaic (mozilla), no Microsoft Internet Explorer (IE Based on NCSA Mosaic code base licensed from Spyglass), no World Wide Web in the early to mid 90's. No doubt someone else may have invented the browser but how much longer would it have taken?

At first that makes you like a little bunny rabbit ... Everybody wants to play with you ....within a year ... fearsome competitors shooting at your head with high-powered ammunition

Larry, Sergi do you feel the hot breath of the MS juggernaut as you approach your IPO. Will google will be a repeat of Netscape/MS tussle?

Oracle database was a huge success ... Larry's spent the last 25 years trying to come up with the next product

it sure helps when the government (CIA) is your preferred backer. Why does oracle feel the need to keep trying to re-innovate or create the next best idea?

innovation comes from companies that are 2 years old, populated by 19-year-olds ... preposterous that Marc should think that innovation is .. the province of little entrepreneurial companies.

In fact it's both. The technical revolution was spurred on the back of the transistor. This was the combined effort of Bardeen, Brattain and shockley at Bell Labs - no small comany there ... but look at Intel, though a big company now, it was started with the (not so young) Noyce, Moore and Grove. What about the Linux kernel, third person shooters and that other search engine, Yahoo?

Most likely if you have a PC built within the last 5 years and a compatible Ethernet card, you can use Preboot Execution Environment (PXE) to boot from the network.

(Though the article referenced refers to Linux, the spec is OS independent. It is, in fact, used by Win2k's RIS, and can be used on computers without any hard drive at all.)

...will continue indefinitely.

Even Intel will only say that they expect it to continue "at least through the end" of the decade. They're running out of space.


p

You are aware that Moore's Law is about the doubling of density of transistors and not "computing power" or some such undefinable quantity? Moore's law will be broken simply because physical entities cannot follow an exponential growth for very long. Computing power will still increase.

Most of you know this, so please just bear with the sermon for those who do not.

Moore's Law is a marketing term which was coined by the press, not Gordon Moore himself. It's not a law in the scientific sense, like the Law of Gravity. The 'law' simply states that the number of transistors on IC's roughly doubles every 18 months. People have been predicting the death of Moore's Law for many years, and probably will for many more.

If it truly were a law, it could not die. But eventually it will fail. In the mean time, it's a 'law' that keeps sales and marketing people busy, ensuring there will always be faster processors to run the latest bloatware.

Copyright(c) 1999 - 2003 Intel Corporation. All rights reserved.

This program is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the Free
Software Foundation; either version 2 of the License, or (at your option)
any later version.

This program is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
more details.

You should have received a copy of the GNU General Public License along with
this program; if not, write to the Free Software Foundation, Inc., 59
Temple Place - Suite 330, Boston, MA 02111-1307, USA.

The full GNU General Public License is included in this distribution in the
file called LICENSE.

Contact Information:
Linux NICS
Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497

That driver is a GPL implementation meaning that the OpenBSD developers are more than welcome to port it at their leisure.

Oh, you want real documentation too. Take a look at developer.intel.com:
Intel 82551ER Fast Ethernet Controller Networking Silicon Datasheet
Intel 82559ER Fast Ethernet Controller Networking Silicon Datasheet
Intel 82559ER EEPROM Map and Programming information

True, Those are for their FastEthernet chipsets, not the Gigabit chipsets that, "Intel steadfastly refuses to provide us with documentation"

Well, what about these?
Intel(R) 82541ER Gigabit Ethernet Controller Networking Silicon Datasheet
Intel(R) 82547GI(EI)/82541ER EEPROM Map and Programming Information Guide

Ok, so the IPSEC chip on the NIC isn't supported nor is there any data on that chip forthcoming. However, there are a number of papers that show that IPSEC and TCP offloading (not to be confused with TCP fragmentation/checksumming) are not efficient. Specifically, the "hardware" IPSEC is done by firmware downloaded to a small embedded processor on the NIC. This small, embedded NIC is not very fast, in fact, its rather slow.

Result:
Processor utilization drops marginally (modern processors can encrypt 10 megabytes/s trivially)
Latency shoots up (It takes the embedded processor longer to encrypt a packet than the host processor would.)

There are a number of papers corroborating that latency has a huge effect on maximum bandwidth.

[I think the paper regarding TCP offloading not being worthwhile is by Mudge. The IPSEC offloading not being worthwhile is my hypothesis, untested, but I feel logically founded.]

My point is that IPSEC offloading is not an advantage - it probably was in 200MHz K6 days, but it certainly is not in 2.0 ghz K7 days.

Other notables, for example, the 3Com 3CR990 still doesn't have IPSEC offloading, despite promises from either the openbsd txp driver or the linux typhoon driver.

Frankly, as far as Gigabit

Copyright(c) 1999 - 2003 Intel Corporation. All rights reserved.

This program is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the Free
Software Foundation; either version 2 of the License, or (at your option)
any later version.

This program is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
more details.

You should have received a copy of the GNU General Public License along with
this program; if not, write to the Free Software Foundation, Inc., 59
Temple Place - Suite 330, Boston, MA 02111-1307, USA.

The full GNU General Public License is included in this distribution in the
file called LICENSE.

Contact Information:
Linux NICS
Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497

That driver is a GPL implementation meaning that the OpenBSD developers are more than welcome to port it at their leisure.

Oh, you want real documentation too. Take a look at developer.intel.com:
Intel 82551ER Fast Ethernet Controller Networking Silicon Datasheet
Intel 82559ER Fast Ethernet Controller Networking Silicon Datasheet
Intel 82559ER EEPROM Map and Programming information

True, Those are for their FastEthernet chipsets, not the Gigabit chipsets that, "Intel steadfastly refuses to provide us with documentation"

Well, what about these?
Intel(R) 82541ER Gigabit Ethernet Controller Networking Silicon Datasheet
Intel(R) 82547GI(EI)/82541ER EEPROM Map and Programming Information Guide

Ok, so the IPSEC chip on the NIC isn't supported nor is there any data on that chip forthcoming. However, there are a number of papers that show that IPSEC and TCP offloading (not to be confused with TCP fragmentation/checksumming) are not efficient. Specifically, the "hardware" IPSEC is done by firmware downloaded to a small embedded processor on the NIC. This small, embedded NIC is not very fast, in fact, its rather slow.

Result:
Processor utilization drops marginally (modern processors can encrypt 10 megabytes/s trivially)
Latency shoots up (It takes the embedded processor longer to encrypt a packet than the host processor would.)

There are a number of papers corroborating that latency has a huge effect on maximum bandwidth.

[I think the paper regarding TCP offloading not being worthwhile is by Mudge. The IPSEC offloading not being worthwhile is my hypothesis, untested, but I feel logically founded.]

My point is that IPSEC offloading is not an advantage - it probably was in 200MHz K6 days, but it certainly is not in 2.0 ghz K7 days.

Other notables, for example, the 3Com 3CR990 still doesn't have IPSEC offloading, despite promises from either the openbsd txp driver or the linux typhoon driver.

Frankly, as far as Gigabit