Slashdot Mirror

One of a billion or so: "In a detailed briefing for analysts in New York on Friday, executives at Advanced Micro Devices painted the company as making "irreversible progress" into new architectures, specifically multicore microprocessors and 64-bit processing. Executives confirmed that the company plans to enhance its Opteron enterprise processor line to four cores in 2007, adding focused optimizations to manage power and improve throughput." -- LINK

This story was on Fark yesterday. I mean, come on....
Fark links the story on ExtremeTech by the same author:

http://www.extremetech.com/article2/0,1558,1922776 ,00.asp?kc=ETRSS02129TX1K0000532

and TheGrapeApe is using the 1Up link:

http://www.1up.com/do/newsStory?cId=3147826

They're the same exact story.

I have an idea: to make /.'s editors' lives easier, why don't you just use a crawler and every hour, you copy one of the Fark/Joystiq/Kotaku/Gizmodo/Engadget links? That way, you still have some 'news', and they might actually be a little less behind.

(BTW, I think you missed the OP's point...)

• resting it in an abnormal position on top of the VerticalMouse 2, thereby completely negating the advantages of a vertically oriented mouse,
  
• the use of your arm muscles, leading quickly to fatigue and muscle strain,
  or,
  
• clinging to the vertical surface of the mouse with your fingers and/or thumb, again leading to fatigue and muscle strain.

"IBM researchers said Monday that they have created a low-power chipset that will compete with ultrawideband technology, offering data rates at around 630 Mbits/s.

"The chipset conforms to the IEEE 802.15.3c specification, which IBM refers to as "millimeter wave" or "mmWave" technology, according to Brian Gaucher, a research manager with IBM Research. "

The bit where it talks about how the 630-Mbits was the limits of their test equipment was cool...

You've got it backwards ;-)

OS X uses some OpenGL stuff; a lot of 2D compositing. It doesn't totally bury the system, however, and it can move a lot of that to software rendering as well; that's why it works just fine on my Powerbook with a GeforceFX 5200, 32 MB ram.

Vista, on the other hand, uses boatloads of 3D, everywhere. Lots of texturing. The main issue with Vista is not having enough graphics ram. For the full "Avalon" "experience", you'll need 256 MB in a 32-bit environment, and possibly more in a 64-bit environment. Fill rates will also be important, in order for you to keep your windows flying around the screen in 3D.

God knows why so much is needed; Project Looking Glass provides a similar display with far more modest requirements, and thats a JAVA window manager. Not to mention that Xorg is getting really, really close to alot of these things. Xgl is currently running with all kinds of interesting shader/geometry effects, and KDE's got the window manager refraction/reflection (take a look at the CrystalGL, the big cousin of Crystal, which does it in software).

Ultimately, Linux will get there, but the problem is integration; most of these features are avaliable on X, but few of them play nicely with OpenGL, and they often don't play well together. We'll have to see a big, combined push between the KDE 4 effort, GNOME's next generation Metacity, the freedesktop XGL/Xorg 7+ people, and NVIDIA/ATI. As I understand it, much of this is occuring now; but we probably won't see releases till near the time Vista is released, and we won't see proper integration into distributions till late 2006/early 2007.

The best part is, however, that once it DOES get into Linux, it'll run just fine on 32/64 MB cards, and most likely will degrade much more gracefully than Vista; there'll be a finer set of non-functional options, rather than 3/4 main settings.

I have no fear that we'll see plenty of desktop eye candy in the near future on Linux; this is mainly attributable to the freedesktop people, who have saved X with Xorg, a product that is making progress now after years and years of stagnation.

I'm much more worried about DirectX 10 (WGF 2.0). Will OpenGL keep up? I hope so, otherwise we'll see the few Linux/Mac gaming houses there are out there (in addition to Transgaming) fail completely as they become unable to port over Windows graphics features. NVIDIA, ATI and Apple seem to be keeping the OpenGL group moving, though.

http://www.extremetech.com/print_article2/0,1217,a =170327,00.asp

Oh, right, TFA.

Surprisingly, the prices of these two cards are very close: ATI's X1800 XT & Nvidia's 7800 GTX.

I'm guessing that they used an X1800 XT with 512MB of GDDR3 while most 7800 GTXs only have 256MB GDDR3. They come to be about the same price but I attribute their release dates ... remember Moore's Law.

Newegg has a great datasheet regarding all mainstream cards.

http://www.extremetech.com/article2/0,1697,1914692 ,00.asp

As of right now, you are correct. But what I was referring to was this (actually a future incarnation of this- note the DNF reference?) new BadBoy on the Block:
(http://www.extremetech.com/article2/0,1697,190949 7,00.asp). Here is an excerpt FTA:
  "The Athlon 64 FX-60 is an AMD dual-core CPU running at 2.6GHz--200MHz faster than the Athlon 64 X2 4800+. Rather than calling the CPU something like "Athlon 64 X2 5000," AMD has chosen to brand the CPU as one of its FX-series line, with a higher model number than the FX-57."

UWB Standards Group Calls It Quits "

Unable to resolve a deadlock between two competing proposals, the IEEE working group responsible for the ultrawideband technology threw in the towel Thursday.

The Institute of Electrical and Electronics Engineers (IEEE) 802.15.3a task group (TG3a), which oversaw the formation of the UWB standard agreed to withdraw the Jan. 2003 project authorization request that formed the group. Instead, the two competing technologies - MB-OFDM, championed by the Intel-led WiMedia Forum, and DS-UWB, promoted by Freescale Semiconductor and its UWB Forum - will be left to fight it out in the marketplace.

I don't think you really grasp what virtualization is capable of. Once the software support is there (i.e. the new VPC, VMWare, Xen, etc), it is almost a certainty that e.g. BF2 for Windows will be able to run at near-native speed on virtualized Windows (though perhaps not til vista), with full support for any graphics card features.

See: http://www.extremetech.com/article2/0,1558,1791683 ,00.asp for more.

-justinb

IBM can incorporate this into the World's Smallest Computer.

Microsoft don't distribute other people's software with their software, simply because they're a commercial entity and don't want to put themselves out of business. Combined with the competitiveness pressure, meaning they can't distribute much of their software with their OS, you get Windows as more or less a blank slate.

I think you are missing the point. It is not that Microsoft don't distribute other people's software - it is that they come to arrangements with PC vendors to prevent the vendor distributing other people's software, or even alternative versions of Windows. Here is an example:

http://www.extremetech.com/article2/0,3973,760198, 00.asp

It is about Microsoft controlling what the PC vendor can do.

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

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

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

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

That reminds me an interesting part of Extremetech's review of Dell's new 30" LCD:

Having the ability to open lots of windows and being able to see most of them was a real productivity enhancer. On the other hand, we did occasionally suffer from the "tennis game spectator" effect--you have to move your head sometimes to take in the entirety of the display.

It's a nice unit. No embedded controls, except for brightness, so you need to use your graphics card control panel to make adjustments. Some minor uniformity problems with the backlight, but good D6500 color temperature tracking.

Oh, and Civ4 looks great at 2560x1600 ;-)

I'd really like to have a laptop with a few GB of flash memory that acts as a read and write cache for the hard drive. With a good caching algorithm, it should be possible to keep the hard drive spun down most of the time and save a ton of energy

You are not the only one thinking of that.

BTW, the article layout is disgusting - 11 pages!

Here is a link to the print version of the article, less clicks and only one ad! http://www.extremetech.com/print_article2/0,1217,a =168194,00.asp

Link to the "print" version.

http://www.extremetech.com/slideshow/0,1206,l=1670 88&s=200&a=167099,00.asp

MF: It's not spyware, like that stuff from what, Gator? It's a consumer-prefence program and digital wallet from Claria. It makes sure that users only see ads directed to their particular tastes.

That's priceless. I got some added stuff with mi Dell, but not quite that evil, and nothing that I could think of from Claria.

On the contrary, my demos of Quicken stuff came with an "Intuit Internal Printer" that can print to a PDF file, and (unlike Ghostscript!) can save links in the documents too. I'm extremely satisfied and await their presshhiouss*.

*from an Ebayer--but a reliable Ebayer, one with a long, flawless record. (crosses fingers)

The Radio Shack's TRS-80 Model 100.

date: Funny argument specified
$ _

While this article may bring some new ideas (hoping for it to be the panecea for converts is kinda silly) a very similar article was out in Feb 2005: Hacking a PSOne Screen

How about the Power Squid? Looks pretty nifty.......

I was not talking about the green marker BS. This was a seperate deal see this article for one example, for more.

http://www.extremetech.com/article2/0,1558,87100,0 0.asp

Bush also inadvertently coined a great spoonerism about power-stealing vampires when talking about this initiative.

I don't get it -- what's the spoonerism? It's not "wall wart" is it? It can't be "wall wart" because that article is from July, 2001, and the term "wall wart" was around many years before that. Go search Google Groups -- I found two uses of the term as far back as 1989.

And for what it's worth, a transformer that is inline (as opposed to hanging off the wall socket) is called a "line lump".

http://www.extremetech.com/article2/0,1558,87100,0 0.asp

George Bush campaigned for this stuff back in the early days. I may not like the guy much, but he was right about this. Companies consistently make their products more power inefficient just to make them cheaper, because very very few people pay attention to efficiency of appliances. They save a few pennies on day 1 and give it back and then some every year.

Energy Star has been incredibly effective. The cheapest refrigerator you buy is within 80% as efficient as the most efficient models. This is definitely not true with many other classes of devices (like lights!).

Bush also inadvertently coined a great spoonerism about power-stealing vampires when talking about this initiative.

Kontiki *IS* spyware. I remember when Gamespot used them, it was basically a spyware download manager that used DRM to hide *other* spyware.

http://www.extremetech.com/article2/0,3973,365073, 00.asp

Because nVidia wasn't the only company cheating. ATI was also found to be "optimizing" for benchmarks, too. Yes, it was a couple of years ago. But you give too much credit to think that people have stopped being stupid. It won't surprise me at all if it happens again.

It seems that any time ATI or nVidia releases a new card, they've also got some drivers that "optimize" for the 3DMark benchmarking software. So I figure it must just mean that nobody's found out how they're doing it yet.

It's kind of sad to think that when they announce some obviously kick-ass hardware that all I can think of is "how did they cheat this time?"

Earlier today I was mentioning how it was understandable that Nintendo would avoid High Definition support because of the potential Performance problems; many people (some of which would probably be XBox 360/PS3 fans) claimed that it was a weak excuse because those systems had the power necessary to render anything regardless of the resolution.

By looking at the following benchmarks you can see how resolution effects performance:

FEAR:

1280x960 http://www.extremetech.com/article2/0,1697,1883897 ,00.asp

1600x1200 http://www.extremetech.com/article2/0,1697,1883898 ,00.asp

4xAA 8xAF http://www.anandtech.com/video/showdoc.aspx?i=2575 &p=5

soft shadows http://www.anandtech.com/video/showdoc.aspx?i=2575 &p=6

As you can see, even with the best graphics cards a person can buy a new game has difficulty maintaining a playable framerate at 1280x1024 (and is mostly unplayable at 1600x1200); at 800x600 they have frames to spare.

I'm not trying to say that higher definition is a bad thing; but it needs to be considered in its place. Higher definition is great for PC games because as performance drops you can simply upgrade your system, on a console (which is a static platform) higher resolution is not that important; eventually when we have the performance to spare it will be a good time to produce games at 1080p or what not.

Earlier today I was mentioning how it was understandable that Nintendo would avoid High Definition support because of the potential Performance problems; many people (some of which would probably be XBox 360/PS3 fans) claimed that it was a weak excuse because those systems had the power necessary to render anything regardless of the resolution.

By looking at the following benchmarks you can see how resolution effects performance:

FEAR:

1280x960 http://www.extremetech.com/article2/0,1697,1883897 ,00.asp

1600x1200 http://www.extremetech.com/article2/0,1697,1883898 ,00.asp

4xAA 8xAF http://www.anandtech.com/video/showdoc.aspx?i=2575 &p=5

soft shadows http://www.anandtech.com/video/showdoc.aspx?i=2575 &p=6

As you can see, even with the best graphics cards a person can buy a new game has difficulty maintaining a playable framerate at 1280x1024 (and is mostly unplayable at 1600x1200); at 800x600 they have frames to spare.

I'm not trying to say that higher definition is a bad thing; but it needs to be considered in its place. Higher definition is great for PC games because as performance drops you can simply upgrade your system, on a console (which is a static platform) higher resolution is not that important; eventually when we have the performance to spare it will be a good time to produce games at 1080p or what not.

NetBSD's fast portability is due to its Modular Portability Layer (MPL). With the MPL, the driver is completely isolated from the hardware platform, I/O instructions or no I/O instructions, interlocking, retry error recovery, bounce buffers, memory type boundaries, scatter/gather maps in host bridges, even peripherals are transparently handled beneath the driver layer.

Umm... yeah this is how its done in Linux too. They don't even need a fancy TLA for it - it is simply the obvious way to write a portable operating system.

With Linux, however, device driver code must be reworked for every new architecture.

No, this is patently, 100% wrong. You probably got this from Wasabi Systems, they spread FUD about Linux like it is going out of fashion. No, *YOU* made the claim that NetBSD is more portable, so I want to see *YOUR* reasons why, not some crap cut and pasted out of a website you read.

Wasabi Systems engineers ported NetBSD to the SuperH processor core in under six weeks; Linux took three months. NetBSD was ported to the AMD x86-64 architectures in about a month; Linux took six months.

I don't know how they know the details of the Linux port because it was initially done under NDA and even before the hardware came out (ie. on emulators).

But if it did take longer then I wouldn't be particularly surprised. Linux has advanced features like NUMA and SMP that is key for good Opteron support.

NetBSD supports over fifty architectures from the same source tree meaning any change to the machine independent code is available to all architectures as soon as the build system can churn them out.

This is "clever" wording. Most of the "architectures" are just platforms (or system architectures) rather than CPU architectures. Linux is ported to about 50% more CPU architectures and more platforms than NetBSD from the same source tree.

And likewise any change to MI code is available to all architectures... by definition, right?

Even AMD agrees: ...

This is nice, but nothing to do with NetBSD portability compared to Linux though. Linux was obviously 64-bit ready as well, as it was ported to ia64 and ppc64 -- incedentally, the current 2 fastest GP CPUs on the planet (Itanium2 and POWER5+) neither of which NetBSD supports.

"Due to NetBSD's ease of portability, Wasabi was able to get up and running on AMD's x86-64 platform several months sooner than other teams could port Linux."
- full article: http://www.extremetech.com/article2/0,1558,1150126 ,00.asp

Again, Wasabi are trolls and probably wrong because Linux was running on emulators before the hardware was released.... but why are you so fixated on speed of the port? Hey, Linux was ported to ia64 several *years* before NetBSD could port to it (and it is still pretty broken), likewise for PPC64.

Again, I am not saying that NetBSD is better that Linux, just an inherently more portable opertaing system.

I don't think you've proved your point at all.

It is funny that the machine dependant portions of code for a single port are generally much bigger on NetBSD than Linux (for those with equivalent functionality and support). Combined with the fact that Linux actually is ported to many more architectures than NetBSD that build from the same source tree, there is a strong argument that Linux is more portable.

Another thing, NetBSD cannot support a CPU architecture that doesn't have an MMU. It simply cannot support such a port. Linux can. Linux is obviously "inherently more portable" in this case, right?

I don't really see much of a point in arguing with you as I obviously can't change your mind, but if you insist:

NetBSD's fast portability is due to its Modular Portability Layer (MPL). With the MPL, the driver is completely isolated from the hardware platform, I/O instructions or no I/O instructions, interlocking, retry error recovery, bounce buffers, memory type boundaries, scatter/gather maps in host bridges, even peripherals are transparently handled beneath the driver layer. Several embedded systems using NetBSD have required no additional software development other than toolchain and target rehost.

With Linux, however, device driver code must be reworked for every new architecture. Wasabi Systems engineers ported NetBSD to the SuperH processor core in under six weeks; Linux took three months. NetBSD was ported to the AMD x86-64 architectures in about a month; Linux took six months. NetBSD supports over fifty architectures from the same source tree meaning any change to the machine independent code is available to all architectures as soon as the build system can churn them out.

Even AMD agrees:

"According to Ed Gasiorowski, director of developer relations for AMD's Computation Products Group, the port to AMD's Opteron processor was extremely fast. "I think the initial port was probably about two days [and] might have been as short as one day. That was a full 64-bit port." He told NewsFactor that the speed of the port was due to the quality of NetBSD's codebase. "We were quite impressed with how clean the code was.... This code was ready for 64-bit before it ever got to porting over to Opteron. Not only from having done it before [porting to a 64-bit system], but also good programming practices." - read full article at: http://www.newsfactor.com/perl/story/21244.html

"Due to NetBSD's ease of portability, Wasabi was able to get up and running on AMD's x86-64 platform several months sooner than other teams could port Linux."
- full article: http://www.extremetech.com/article2/0,1558,1150126 ,00.asp

Again, I am not saying that NetBSD is better that Linux, just an inherently more portable opertaing system.

Does RAM Latency Matter?

How's about this one? 2004, ExtremeTech
http://www.extremetech.com/article2/0%2C1558%2C163 7762%2C00.asp?kc=ETRSS03039TX1K0000564

ExtremeTech Article

Check out http://www.extremetech.com/article2/0,1697,1869093 ,00.asp for an explanation.

The third future project at ATI is dramatically improved support for the GPGPU scene. These are researches, mostly academic, that are tapping into the massive parallel computing power of graphics processors for general computing tasks, like fluid dynamics calculations, protein folding, or audio and signal processing. ATI's new GPU architecture should be better at GPGPU tasks than any that has come before, as it provides more registers per pipeline than either ATI's old architecture or Nvidia's new one. This is a sore spot for GPGPU developers but not really a limitation for game makers. The improved performance of dynamic branching in the new architecture should be a huge win for GPGPU applications as well. Developers working to enable general purpose non-graphics applications on GPUs have lamented the lack of more direct access to the hardware, but ATI plans to remedy that by publishing a detailed spec and even a thin "close to the metal" abstraction layer for these coders

Audit report (PDF)
Breaking story

To at least make the article into one continuous stretch of nothing, rather than nothing broken by ads every so often, click here for the printer-friendly version.

A couple other views of the Vista control panel:

Category view from build 5112. Another full screen shot with the menu bar on the left.

Icon view

Personally, I use the classic view in XP, setting it to show a menu in the Start Menu. I know where I want to go, and it's nice that the options haven't changed much since 95. It's nice that they don't go renaming and regrouping the different options every release.

I do like the category view in Vista better than XP. The "quick links" to oft-used functions in each category are listed under the main category name, a big improvement. The fewer sub-menus I have to open, the better.

Unforuntately you'r wrong about this-- AMD's fab partner, Chartered Semiconductor, is still a long way from being able to produce CPUs to help AMD meet demand. It's going to be about a year, from what I read, before Chartered can begin to make parts for AMD, assuming all goes well-- and AMD has a history of things not going well with foundries (though with Chartered, they seem to be ready to share the crown jewels of their fab technology in the name of having the extra capacity on a "variable cost" basis, where they're not stuck sinking multibillions into a mega-fab before the first chip comes out). http://www.extremetech.com/article2/0,1697,1826658 ,00.asp reports that Chartered can make maybe 10M CPUs in 2006, 30M in 2007. Not insignificant but a while off.

Dell is not the be-all and end-all of the marketplace anyway-- my thinking: Sun, HP, Fujitsu-Siemens (who just announced product) and IBM (especially Sun, being exclusive and having some really clever server designs Dell can't touch) can move as much Opteron as AMD can make, at wonderfully plump prices. Dell would get AMD into certain accounts they're not in now, but there's time enough for that later-- possibly much later.

Argh. I'd forgotten that 64-bit XP was *still* only XP Pro.

As for the Pentium 4's 64-bit capability, a quick Google search turned up a couple of sources and dates. The first Pentium 4 processors with EM64T support -- Intel's version of AMD64 (because there's no way they'd actually use a name with AMD in it!) -- came out in March, and apparently the "majority" of the Pentium 4 line has been 64-bit since June. Maybe PR is keeping quiet because the average user will be using XP Home?

January: The Pentium 4 adds 64-bit Extensions
March: Intel's 64-Bit Pentium 4s Hit The Streets
June: Intel Shifts Pentium 4 to 64-Bits

Actually, I don't think it would help much. Most games now don't benefit from 2 way SMP, so the benefit from 64 way is debateable to say the least. Still for servers, this thing might help. I suspect that most server applications/os's will have servere scaleability problems once you go this far SMP though.

BTW, Has anyone heard of the MLX1. Makes you wonder what would happen if you put a bunch of these on a chip with some clever caching and the mother of all memory controllers. x86 Niagra anyone.

I agree completely. I remember what it was like "in the dark days" before Windows came along. Hardware vendors supplying operating systems for their systems. No two systems even remotely alike. Yes, Apple was the first with a GUI for the masses (copying from Xerox PARC), but Apple was just another hardware vendor that also did software - they just did a much better job than the others because they considered "the rest of us" when they designed their systems.

Anyhow, I think Microsoft is almost singularly responsible for the incredibly cheap and powerful machines we all enjoy today. And yes, they made Linux possible by forcing all the hardware vendors to adhere to common standards.

Check out this article for Lloyd Case's view of what the world would have been like had Microsoft not emerged:

http://www.extremetech.com/article2/0,1697,1869093 ,00.asp

In some ways, Microsoft's history reminds me of AT&T. Perhaps the benefits of having a monopoly while a new technology is "growing up" might outweigh the disadvantages. They can bring order to a chaotic marketplace. Of course this then makes me wonder if we have reached the tipping point where the benefits are no longer greater than the drawbacks?

A few nitpicks:

I don't think a new version called "Media Center Edition 2006" was ever planned or announced. The article's introduction referred to the "massive Rollup 2 patch," which is a free update to MCE 2005. This update was supposed to include Cable Card support, but this feature was cut (according to Microsoft's Matt Davis). Here's the blog entry where I read about this: "More details on Rollup 2."

Also, the next "new" version of MCE (Longhorn version) will not be called "Media Center Edition" anymore. It will be called "Vista Home Premium Edition" (details here). This will probably be the first Windows version with Cable Card support. Ugh.

obviously you missed the fact that when gaming at 1600x1200 and are using 4x antialiasing and 8x anisotropic filtering that the x1800xt beats the shit out of the gf7800.

if you're buying a 500 dollar card, are you seriously worried about benchmarks that are run without aa+af? this card even does HDR (hi dynamic range) plus AA, something that the gf7800 can't.

this card is way more sophisticated and highly refined that the brute force 7800. the 7800 isn't bad but that this card can do with 16 pipelines what the 7800 can't do with 24, says a lot.

and that's just raw performance with todays games. never mind the fact that the 1800xt comes with 512megs of super fast ram... ready for well into the next generation of games, whereas 256meg 7800's are already obsolete for the high end of the next generation. sure 256 will be enough if you pare down the resolution and lower the texture detail. one example is the game F.E.A.R... on the 1800xt it absolutely trounces the 7800 in performance.

my advice... read ALL the reviews you can get your hands on. there are too many discrepencies if you only read one or two. if you want to get a more full picture, get to reading.

http://www.anandtech.com/video/showdoc.aspx?i=2552

http://www.beyond3d.com/reviews/ati/r520/

http://www.driverheaven.net/reviews/r520reviewxvxv /

http://www.extremetech.com/article2/0,1697,1867116 ,00.asp

http://www.firingsquad.com/hardware/ati_radeon_x18 00_xt_xl/

http://www.guru3d.com/article/Videocards/262/

http://www.hardocp.com/article.html?art=ODIy

http://www.hexus.net/content/item.php?item=3603

http://www.hothardware.com/viewarticle.cfm?article id=734&cid=2

http://www.neoseeker.com/Articles/Hardware/Reviews /ati_radeon_x1800_x1600preview

http://www.pcper.com/article.php?aid=172

http://www.tbreak.com/reviews/article.php?id=407

http://www.techreport.com/onearticle.x/8864

and check out the wicked new 3d tech demos... both are very impressive but the toystore demo is jawdropping.

http://www.ati.com/designpartners/media/edudemos/R adeonX1k.html

wmv9 hi def format but plays fine in mplayer or VLC.