Slashdot Mirror

The box AOL is using is the "Intel Dot.Station"

There is discussion about the box on Barriopunto (Spain's version of Slashdot) here.

Comments on it vary. Some people say it's OK for Joe Public. Others say it sucks and hangs up. One says that the Mozilla and Linux distribution is very heavily modified and it is difficult to tell which versions they are using.

Someone points out that the box is basically the Intel Dot.Station Web Appliance. The spec. for the box is here.

The spec. says:

* Custom Intel browser based on Mozilla-- the world's most standards-compliant browser technology.

And later:

* Custom Linux operating system for increased flexibility and innovation.

More information about AOL avant from Intel's web site here.

There is discussion about the box on Barriopunto (Spain's version of Slashdot) here.

Comments on it vary. Some people say it's OK for Joe Public. Others say it sucks and hangs up. One says that the Mozilla and Linux distribution is very heavily modified and it is difficult to tell which versions they are using.

Someone points out that the box is basically the Intel Dot.Station Web Appliance. The spec. for the box is here.

The spec. says:

* Custom Intel browser based on Mozilla-- the world's most standards-compliant browser technology.

And later:

* Custom Linux operating system for increased flexibility and innovation.

More information about AOL avant from Intel's web site here.

The new ABI for g++ is not standard...

Depends on what you mean by "standard". It's true that the C++ standard doesn't specify a C++ ABI. (It can't, since an ABI necessarily depends on the platform.) However, the new ABI for g++ does conform to an informal standard defined in http://www.codesourcery.com/cxx-abi. This ABI was defined cooperatively by a number of companies with an interest in C++ interoperability; it was originally defined for the Intel Itanium, and was later generalized to other platforms.

about this whole Transmeta thing is the level of speculation and un-clearness.

Talk all the shit you want about Intel, but I can tell you that I'm working on a board right now that uses a Mobile Celeron Mobile 400A: http://developer.intel.com/design/mobile/datashts/ 28365403.pdf. The datasheets says thermal power 10.1 Watt max. Well, we never _ever_ get that high. Also, the newer 500 Mhz ultra low power is 8 Watt max, 5 Watt under more normal conditions.

The thing is that TM _never_ published said figures (quickly: what's the MAX Watts a TM CPU can draw?), because supposedly all that we need to know is the power required to decode a DVD. Well, today that happens largely by the VGA controller now, doesn't it?

What suprises me even more is that Torvalds, if anyone, should know that using the simple HLT instruction in the idle thread, makes any Intel (or AMD) CPU draw a lot less power.

Even on paper I don't see the advantage of the TM CPU's. And I really hoped they would, believe me...

I'm curious about what information AMD and Intel have to back up their claims that they're catching up to Transmeta in power requirements.

The only thing I've heard about is the revolutionary new Intel Pentium(R) processor, described by a company spokesman as "Like the pentium III, but consumes much less power." Operating with an order of magnitude fewer transistors, and clock speeds of up to 200MHz, the performance is almost as good as the Crusoe.

The best news is they're already released and available from reputable dealers everywhere!

Most people did: SunOS was widely available, with sources, it worked reasonably well, and it ran on good and fairly cheap hardware. Linux took off because nothing quite like it existed for PC hardware.

I agree, though, about Java and the risks inherent in Sun's Java license. But the biggest problem with Java and open source is that Java is used so little in major open source projects. If the open source community used Java widely and developed their own standards (say, an "xAWT" or a "jGTK"), Sun would either become irrelevant or they would be quick to "donate" their libraries. And existing open source Java implementations will only get better if people actually use them.

Incidentally, Intel ORP is an open source and rather efficient Java VM, beating Sun's JVM on many numerical benchmarks.

Do a Google search on google cluster ide. The third result is an Intel customer profile on Google:

In Google's environment, disk I/O performance is an overriding factor, yet the cost of high-end SCSI disk subsystems is prohibitive. So the company standardized around inexpensive IDE technology, outfitting each server with a pair of internal disks storing either 22GB or 40GB apiece. "We did a lot of benchmarking early on, and we found that for the best price/performance we would set up two IDE hard disks, each on a separate controller," says Reese.

I like two IDE drives (one per channel), plus SCSI for the CD-RW and/or DVD.

This is right on track with Moore's Law.

Intel Promises a Cool Billion (Transistors)

"SANTA CLARA, California (Reuters) - Intel Corp., the world's largest chipmaker,
said Monday it's developed new semiconductor packaging technology that will
allow for microprocessors boasting 1 billion transistors and running at 20
gigahertz in the next six years.

"Currently, Intel's Pentium 4 microprocessor -- the primary computing engine of
personal computers and servers -- has about 42 million transistors and runs at
two gigahertz."

Moore's Law:

Moore predicted that the number of transistors per integrated circuit would
double every 18 months.

Intel has more info on this (both pdf's):

This backgrounder (4 pages, 17kb) has a basic diagram showing the change.
This briefing (18 pages, 2466kb) is a presentation, but actually has some nice detail. It has some photographs of the devices, better diagrams, and a picture of a naked man in the shower (really!).

I'll summerize:
PGA packaging (as used in many big processors) is basically a ceramic or fiberglass carrier board with pins on one side, wires in the middle (like a small PC board), and some method to directly attach to the chip. The chip is usually connected to the board with small solder balls, like BGAs, but on a smaller scale. The balls provide some flexibility and loose tolerances, but since they are bigger than the wires they connect they require a fairly large pad on the chip. This technology is a way eliminate these balls, allowing for smaller pads, freeing up more area on the die.

But you should check out the pictures -- they describe it better than I do.

Intel has more info on this (both pdf's):

This backgrounder (4 pages, 17kb) has a basic diagram showing the change.
This briefing (18 pages, 2466kb) is a presentation, but actually has some nice detail. It has some photographs of the devices, better diagrams, and a picture of a naked man in the shower (really!).

I'll summerize:
PGA packaging (as used in many big processors) is basically a ceramic or fiberglass carrier board with pins on one side, wires in the middle (like a small PC board), and some method to directly attach to the chip. The chip is usually connected to the board with small solder balls, like BGAs, but on a smaller scale. The balls provide some flexibility and loose tolerances, but since they are bigger than the wires they connect they require a fairly large pad on the chip. This technology is a way eliminate these balls, allowing for smaller pads, freeing up more area on the die.

But you should check out the pictures -- they describe it better than I do.

1) No, because it runs at 206MHz does not mean it comsumes a lot of power. It draws 0.7W.
2) It is RISC rather than CISC, and having used a 200MHz StrongARM desktop I can tell you it FLIES. Much faster than a P2-266
3) You use gcc to compile on StrongARM because Linux runs on StrongARM (well obviously). ARMLinux has been around for years running on Acorn machines. You can also cross-compile to StrongARM using a x86 box - just ./configure --target=arm-linux when compiling GCC.
4) You can even use them for Beowolf ;-)

Phillip.

Intel has several different versions of the Tualatin, the 256K cache version for the desktop, and the 512K cache versions, one for the mobile segment and one for the lower-end workstation/server market. The Tualatin processors require a different stepping in the 810 and the 815 chipsets to support the new voltages as well as a new revision of 1.25V GTL versus 1.5V GTL+ used in the original Pentium III processors, which aren't compatible. More information about this can be found in the Pentium III-S Datashet here.

As far as motherboard compatibility, I think Intel didn't want the 512K version of the Tualatin (aka the Pentium III-S) to flow through the retail/desktop channels because in a lot of cases, it performed better than their lower end Pentium 4 processor line. That's also the reason why Intel has slowed/stopped production of Pentium 4 processors below 1.6Ghz and will halt production of the desktop version of the Tualatin and shift the current desktop Tualatin into the Tualatin-based Celeron (but without data pre-fetch and only at 100Mhz FSB). More information can be found at the regular sites: Anandtech, Aces Hardware, Tom's Hardware.

Intel has something similar and should be shipping soon: http://www.intel.com/home/webtablet/index.htm

but for the price.. who really wants one?

This is yet another arena where there is lots of cool vapor and little actual product.

An example:

100 MHz [Intel Pentium] (166.3 MIPS, 3.30 SPECint95, 2.59 SPECfp95 on Xxpress 1M L2)

A quick Google reveals the following

Enormous FPU performance with three simultaneous instructions and one GFLOP at 500 MHz (1 billion floating-point number operations per second) with 80 bit floating-point numbers. Two GFLOP with MMX and 3DNow! instructions. That at least equals Pentium III's performance with full utilization of Katmai. The 3DNow! engine has even been improved comparing to the K6-3.

Yeah, they're talking about an Athlon, and it's only 1GFlop, but it's quite an old article, the Athlon still being in Slot 1 form.

Or, from Intels site (use lynx because it refreshes to a missing page)
The adder and multiplier were placed on different ports. This allows for simultaneous dispatch of 2-wide addition and 2-wide multiplication operations. This boosts the peak performance two more times when compared to the Pentium II, and hence, it allows 2.2 GFLOP/sec peak at 550 MHz. The new units developed on the Pentium III and modified P6 units are shown in color in Figure 3

Which also reveals that the PII was capable of 1GFlop.

Moral of today's story:
Search before flamebaiting, please.

Have a look at this, though: Intel Open CV library. The guy has a "no glove approach" to gesture recognition.

Especially, have a look at Gestscal.avi (AVI of detecting static hand position gestures using gradient histograms 5.1M) and of course the manual.

Hope you find some ideas in there (the lib is opensource).

Have a look at this, though: Intel Open CV library. The guy has a "no glove approach" to gesture recognition.

Especially, have a look at Gestscal.avi (AVI of detecting static hand position gestures using gradient histograms 5.1M) and of course the manual.

Hope you find some ideas in there (the lib is opensource).

Have a look at this, though: Intel Open CV library. The guy has a "no glove approach" to gesture recognition.

Especially, have a look at Gestscal.avi (AVI of detecting static hand position gestures using gradient histograms 5.1M) and of course the manual.

Hope you find some ideas in there (the lib is opensource).

The commercial version is not extremely unreasonable in price, the main disadvantage in it is that it is not completely compatible with gcc, so you can't go and recompile all the packages on your system with it.

Here's a link to Intel's page for it, I believe you can get an evaluation version from them here

A free version is here for non commercial use is here.

The commercial version is not extremely unreasonable in price, the main disadvantage in it is that it is not completely compatible with gcc, so you can't go and recompile all the packages on your system with it.

Here's a link to Intel's page for it, I believe you can get an evaluation version from them here

A free version is here for non commercial use is here.

As opposed to Intel CPUs, which have overheating protection and therefor run only with half the speed as labeled and sold.

The increase in power consumption and the resulting heat developement in current desktop microprocessor is not restricted to AMD. It is a common trend.

Intel has many server boards that have a dedicated chip for management over the serial port. Also, VA has\d a very cool box to plug all your servers into, may it rest in peace. ttp://developer.intel.com/design/servers/buildingb locks/boards.htm

All of the processors I've ever bought from a discount vendor or via mail order have been "tray" product--shipped as individual processors, on foam backing in static baggies. Dunno if AMD offers the same options, but since there seem to be a higher percentage of overclockers and bargain hunters in their customer base, it wouldn't surprise me if they didn't have the demand for a premium packaging offering like that.

I just got DSL from Qwest, and was provided with an Intel 3200 external DSL modem, which connects to my PC via a USB cable.

As far as I know, the Linksys and Netgear routers all use RJ-45 connections for the WAN side of the network.

Is there a router out there than I can connect to my DSL modem via USB?

Thanks.

IBM and Intel both support Linux, but also support fair-use restriction technologies (which subject one to the DMCA) that hurt Linux.

You're saying the compiler has knowledge of registers, and what branch will be taken? Further you're saying the compiler has knowledge of the *current* memory structure? Latency of a particular memory fetch/store (whether the data is in L1/L2/L3/L4/L5 memory?). When DRAM refresh is going to hit (if ever). Or that an interrupt may come in randomly. All this info is VERY useful for the processor to reorder its instructions to avoid a pipeline stall. But of course, you'd say the compiler knew all this detail ahead of time - right? ;)

And i'd really prefer to avoid optimized recompiles during VM page swaps. They take long enough as is. And really, on a decent system - you shouldn't be paging! The reason you do this during a page swap is because that is when the processor would normally be stalled/idle waiting for data or instructions anyway. If implemented properly, this requires no extra cycles to perform. Every system page swaps, which is why TLB's (translation lookaside buffers) exist...to translate between virtual memory addresses and local memory addresses.

Intel has a lot of smart people working for it (smarter than either you or I). They have done some dumb things, but overall they've been on the mark a surprising number of times (even the P4 looks pretty impressive now that clock-speeds have ramped up). It would be a serious mistake to underestimate them like that. If Intel is putting this product on the market, you can bet that they've fixed the compiler problem. Initial benchmarks of the Itanium seem to show that it can keep right up there with the Alphas and SPARCs in terms for performance (fp, at least).

As for compilers, don't discount Intel so easily. They make incredible compilers. The features of ICL for x86 make compiler designers cream their pants. Read this article for some info about Itanium's compiler design.

What's not commonly known is that the P3 and P4 also have dozens if not hundreds of registers. The trick is register renaming: the P3 and P4 speculatively execute instructions as fast as they can, and they assign the results to temporary registers. If the processor needs these results, they reassign them back to the real registers like EAX, EBX, and so on.

So, overall, I'm not sure where the 328 number comes from. :P

Another technical and practical article at Intel - Ultra-Wideband Technology for Short- or Medium-Range Wireless Communications

1. Your QNX numbers are out dated
2. QNX doesn't run stuff in kernel mode you dumb fuck
3. Your microsoft link doesn't exist
4. I am not aware of a 90MHz Pentium II ever existing. To my knowledge, the lowest clocked PII was the 233MHz
5. Who the fuck cares what Microsoft claims

as a true slashdot reader i must point out the PentiumIV runs faster:
Rapid execution engine:
The Arithmetic Logic Units (ALU) run at twice the speed of the clock, increasing the overall speed.

So if intel wants to play the Mhz game it still has some hidden mhz down there.

What makes you think that a 500MHz increase in CPU speed today is harder to achieve than a 50MHz increase 5 years ago? 5 years ago when the Pentium 200s were hot, another 50MHz would have been as big a deal as 500MHz today. It will take the same amount of time, too. Let me point you to Moore's Law clearly shows that CPU speeds increase at the same rate.

Let me tell you also, that if I'm running a maching on CPU bound tasks, even a 5% speed increase is worth buying. Especially if those tasks I'm running take large amounts of time to complete (weeks for scientific calculations!).

If you travel from point A to point B at an average speed of v miles per hour (where B and A are more than v miles apart), there will always be an interval exactly v miles long that you will transit in exactly one hour.

You haven't heard of Moore's law? Of course, not everything is chips, but prices do tend to fall and/or some components get replaced with cheaper alternatives.

The documents you are wondering about are here .

Have fun =)

I've never used them, but Intel does provide high-performance math libraries. So, their compilers probably have real technical optimizations as well (not just marketing fluff).

Uhhh... yeah. So I submitted this exact same story a few days ago when the original article appeared on IDG.net . It was rejected.

Guess I should go for CNN as the authoritative recycler of science content next time instead of getting it from the source. After all, we all know how CNN is always right. :/

Arbitrary decisions aside, at least this is some encouragement that the irresistable force of Moore's Law won't meet the immovable object of the physical limits of silicon, etc and our universe will continue to exist!

You could be much more conservative and still get a computer with better specs and in the same price range as the $299 Xbox.

That's part of the problem with consoles... They become obsoleted before they even hit the market. For example, the Xbox (which hasn't even been released yet) is supposed to run a 733MHZ P3, but Intel already has a 1.8GHZ P4 out in stores. The Xbox is supposed to contain an 8G hard drive, but 40G hard drives have been available cheaply for quite some time now. The Xbox graphics chipset also falls behind chipsets like the GeForce 3 and the Radeon 8500.

The real question is: Why would anyone want to spend so much for a cheap PC? (When most Americans already own one of those). It seems like most developers will use DirectX, so all of the games will be immediately 'ported' to the PC platform anyway.

God bless,
-Toby Reyelts

See, here's the thing... I make out okay in the money and beer department, but I don't always have the money to pay for cool geek toys. Find a cool geek toy that's roughly in the price range you're thinking, preferably one that you can verify or guess that they don't have. If its linux compatible and/or can help them continue developing, then that's even better!

Some quick ideas in the under $100 range:

• Drawing Tablet (most devs have always wanted one to play around with but not enough reason to actually pay for it)
• Some decent speakers
• Mice, keyboards, etc. with Geek potential (check out Intel's new wireless solutions for ideas)
• Palm m100
• Gift certificate to Best Buy / Circuit City / Half.com, etc.

That's it I could come up with... any other ideas?

I think people here who are saying "Big deal, it's how the industry works.. Old CPU's stop getting made" are missing the point.

The "embedded" world is a bunch of companies producing devices that are usually small, lower powered, small production runs, and generally get made for alot longer than most electronics you're used to.

The last company I worked for (Midway Games) made arcade(coin-operated) video games. For a brief time, I worked with the group called "Wavenet". Wavenet was an idea to link arcade games up in arcades all through the world, to allow real-time tournaments. The first game that was tried was Mortal Kombat 3. MK3 used a really weird processor called a 34010 from Texas Instruments. (Weird in that it had *BIT* addressable memory, funky graphics opcodes built in that we never used, etc) However, the game designers pretty much pushed the CPU to its max before we had a chance to make it a networked game. There wasn't enough RAM, CPU, or ROM (for networking code) left to do it, as well as this board didn't have an ethernet output on it to connect it up to the router.

Midway ended up designing a tiiiiny little board (running a small embedded OS that just translated game commands into TCP/IP and vice versa) that plugged into an expansion connector on the MK3 board. It had an Ethernet controller, some ram, more ROMs for the networking code and a 386SX CPU made by AMD on it. Why not use a Pentium, or Pentium Pro? (which was the newest CPU out at the time)

Cost. Right now, you can get 386SX CPU's for a couple of dollars.

Power. Compare the latest generation of 386 CPUs to even a slow PII. HUUUGE difference here.

Board space. The embedded 386's are a little bigger than an american nickle. Pentium class CPU's... well... are big.

Longevity. When we bought these, we got committments from our suppliers that the CPU would be around for at least X months/years. This is REALLLLY important to us. If we're going to spend a ton of cash designing a board based around a CPU, we don't want it to disappear next month when something better comes along.

Had the embedded world not existed, and we had to use a faster/newer CPU, the board cost would have doubled, it would have been a bigger board(again more $$), We likely would have needed to put a bigger power supply(or played tricks with regulators), and then had to redsign the board every time the trendy chip got unpopular. All for horsepower we didn't even need!

Take a look here. Intel is still supporting and selling 80186 CPU's, for embedded controller uses.

Many many companies depend on slower CPU's for things. I don't know if it's still true, but at one point nearly every computer-controlled traffic light system sold used an 80186 CPU. Intel(?) came up with a "hardened" version of it that tolerated extreme cold and extreme heat. Companies that produce products like that are even happy paying double price for an old CPU that can do that, than installing air conditioners and heaters in every traffic light box.

The embedded CPU industry is a place where normal PC economics do not apply. It's not unheard of to pay extra for a part just because you know it'll be around for 10 years, instead of a cheaper(sometimes better) part that will go away as soon as it's not trendy.

While I don't know the specifics of this deal, it sounds like AMD is breaking their previously announced EOL(End Of Life) dates. This is quite likely going to piss a lot of people off who built their product around one of these CPUs.

Interesting, hadn't given this one much thought, but it makes sense. Throw P4's out of the mix here for a moment ( since early reports suggest that P4 is a marketing tool, time will tell ). Check out this Intel document about managing increasing CPU power. It is interesting, if my boss gives me too much work, I'm going to shout "STOPCLOCK".

Instead of just factoring in performance when deciding whether to go with a 1.0 GHz PIII or a 1.3 GHz PIII, maybe we should starting looking at power consumption variations over the frequency range of a single architecture. Take the pIII for example: I wonder if the relationship between powerconsumption and frequency over the pIII line is linear ( probably not )? Anybody know?

Are you implying that Moore's Law is dictated by the laws of physics? Moore's "law" is not a law--not in the same sense as the "laws" of physics (e.g., gravity). Moore merely predicted the exponential growth of transistor density in ICs. This prediction is more of a sociological obsevation--a technology industry truism--and a self-fulfulling prophecy than it is a physical law.

Microprocessors with tens of millions of transistors were no less physically possible in 1965 (according to the unchanging laws of nature) than they are today. Man simply had not developed the expertise, tools, and vision to make them.

Moore's law may cease to hold true someday because of some physical limitation (though this may depend, in part, of how you choose to define a transistor--is it a silicon FET, an organic structure, or anything which functions conceptually as a switch?). However, there's no reason it couldn't end today if we simply chose to stop developing denser ICs.

Already done.

They even have a compiler plugin for VC++ that generates fast binaries. Check this out.

Adobe's a member of the BSA.

The BSA has an interesting statement on the DMCA here. This is a response to a Library of Congress rule available here.

Members of the BSA include Adobe, Apple Computer, Autodesk, Bentley Systems, CNC Software/Mastercam, Compaq, Corel Corporation, IBM, Intel, Intuit, Lotus Development, Macromedia, Microsoft, Network Associates, Novell, Sybase, Symantec, and Walker Digital; i.e. most of /.'s favourite hate companies, plus some extras.

These are the guys to line up against. They've been around since the '80s. I suspect that Adobe's lawyers are all BSA stooges. Certainly Adobe's PR department doesn't seem to be toeing the BSA line.

What does Moore's Law have to do with software? Moore's law just says that about every year, the number of transistors you can fit in a chip will double. He changed it in 1995, to say about every two years. 18 months to be exact. And he originally said this in 1965. And I do believe the software industry existed before microsoft entered the picture. See his bio here

Not only that, but Apple stole the damn slogan "The Center of Your Digital World" from Intel. My digital world only has one center, thank you very much.

I just don't pay for garbage web content. Sites like slashdot are just a waste of bandwith for the most part. They rarely provide any new and/or useful information.

On the flip side I do pay for access to assorted electronics databook and publication sites as well as buy a lot of books. For instance the IEEE page is worth the money. The Intel Developer site is worth paying for as is the National Semiconductor site but they are free because these sites make their money selling the product. The techical documentation is an aid in use.

There is another type of site that makes money on the internet. Its the type of site that provides a REAL service. These are sites like REALTOR.com, bill payer services, job search services, dating services, etc that work well on the web.

The problem with the web is this pay for bandwith scheme that is kicking everyone in the butt. Take internet radio for instance. You could be part of a band co-op and try to put your music on the web. If your site becomes popular though you could potentially be paying a whole crapload to serve the content. With IPv6 you only pay for a miniscule amount of bandwith required to send a single stream which gets broken up by the multicast routers. Sites like slashdot which are dynamic couldn't benifit from this because they try to tailor their views for every user. A nice idea, but completly useless when you compare the cost of running a site that doesn't provide a useful service and the tons of bandwith required to give ever user a unique experience. Slashdot could provide 99% of the user experience with just static content that could be cached in web caches.

then why does intel have it? just do a search on google for "gigabit ethernet over copper" and you will see all the hits that come up...