Slashdot Mirror

A list of resolutions on consoles: http://forum.beyond3d.com/showthread.php?t=46241 As you can see, many titles don't support 720p. And some have both low resolution and no AA.

Not even close. Out of the hundreds of games available for the platform there are only a handful that use 1080p. Full details here.

> C can't really be hugely optimized either,

You've never worked on a C/C++ compiler have you? :-)

C most definitely can be sped up. Why do you think we even have the "restrict" keyword?
http://cellperformance.beyond3d.com/articles/2006/05/demystifying-the-restrict-keyword.html

Function temperature provides hints to the compiler on what code should be inlined.

The back-end of a compiler has a lot of room for generating optimal code IF it understands the target hardware. i.e. Minimizing register spill is extremely important on x86 with its lack of GP registers, reordering instructions to minimize pipeline stalls, known when/where to issue read/write barriers, etc.

One of the major features / problems in C is that arrays = pointers. This is a MAJOR headache WRT to optimizations in a C compiler. In a "clean" language without pointers the compiler is able to strongly infer a lot of assertions; once you throw pointers in the mix the potential for optimizations goes right out the window since the compiler is no longer able to "pre-calculate" offsets.

The lack of strong type inference in C also limits its optimizations.

The LISP / Haskell guys don't have pointers and as such they force the burden on the compiler to "churn" through the type inference system in order to generate optimal code. This leads to a smaller, cleaner compiler, at the expense of a developer needing tons of RAM and CPU to just get the same level optimization a C compiler can do in half the time.

At the end of the day it is all about instruction ordering and caching. Keeping track at what time are what variables "live" and where.

Almost every PS3 game I have is native 1080p, all the way up to Mortal Kombat.

Man you're a stupid one. Very likely under the age of 20 given the sheer lack of knowledge you possess, which is evident in your comment history.

Again, completely shows how LITTLE you know about making games on these machines (TBH, "little" is actually too much to describe the knowledge you have on this subject). Let me guess (like many others have): did you look at the top right of your TV and took that as the native rendering resolution of your games? LOLOLOLOLOLOLOLOLOLOLOLOLOLOL. PLEASE FAIL MOAR.

I think it's conclusive to everyone reading that you're a PS3 fanboy. Nothing wrong with being a fanboy, but you should know when to stop. And don't worry, I'll answer your question (if it wasn't clear enough from my comment history): I'm a dev who actually loves the strengths and weaknesses of each platform (yes, weaknesses too, because that forces us to innovate, such as using the SPU to do post processing!).

P.S. For anyone else who's bothering reading these posts, don't take the list as set in stone...even thought it's close enough, but it's still a few dozens of pixels off here and there (mainly due to memory alignment issues).

Each new console generation is a major leap over the last, given the years that go by between upgrades, but I wouldn't call them vastly overpowered - each upgrade brings them about up to scratch with the GPU technology of the day. In particular, the vast majority of games on the Xbox 360 and PS3 don't render at 1080p, and sometimes can't sustain 30fps despite this. Developers get more out of the hardware when they can target a specific platform, but they're looking decidedly long in the tooth now.

Having said that, the Ouya isn't even supposed to be released until next year, and it's already not on the fastest SoC in a market that's moving fast. I have a Tegra 3 device myself driving a 1200x800 display, and I'd say that its rendering can be about on-par with current consoles (given that they usually render at 720p at best) as long as things aren't too complex... but the current consoles are years old and past due for replacement. The Ouya is okay for today, but it'll be decidedly shoddy in five years' time, and if they release incremental upgrades you lose the simplicity and hardware-targeting that are the advantages of consoles in the first place.

So what? STL isn't suited for all possible uses, sometimes you need your own string and container classes.

Don't be a zealot, pragmatic programmer should find the right trade-off between reusing code and writing an optimal one for a specific problem/area. Nothing can be optimal in all cases - sometimes you need to be as close to hardware as possible at the expense of unreadable/inflexible code (for me, those are the most interesting and challenging areas), and sometimes you only care about maintainability of your code by a disposable programming drone.

Of course they wouldn't publically admit to this problem, but many reviewers were puzzled by the choices that ATI made at the time to implement fast path AA resolve when samples were fully covered (for example here). These choices were basically because that path in the ROP was not pipelined properly (and thus slow)...

Anyhow, they made a diving catch for that and that's water under the bridge now days...

Modern C++ compilers put the unwinding logic in other pages. The C++ one can be faster if errors are rare, but slower if errors are common.

I've looked at this myself. It is very nice that in C++ the "default" way of approaching error handling with RAII plays nicely with the compiler's static branch prediction logic. However, it's not a sufficient argument for truly high-performance code. Either your compiler provides conventions on static branch prediction or it provides intrinsics such as __builtin_expect that give you control. Both of these will give you the same benefits as the C++ RAII style, and (in my experience) neither approach gums up the logic enough to hurt readability.

> Who does this help?
People who are curious about "How Things Work", specifically, "How Games Work"

Perfect example:
quake 3 fast inverse square root
http://www.beyond3d.com/content/articles/8/

> will anyone read through this code?
Most definitely:
  Game Programmers - Sub-Topics: Rendering, Networking, Audio, Physics, AI,
  Future game programmers
  Hackers

I've shipped numerous games. It is always a pleasure reading other people's game code -- just to see how they did things, read the comments, etc.

--
"Necessity is the Mother of invention, but Curiosity is the Father"
    -- Michaelangel007

Lair = 800x1080 (2xAA) - AA buffers are merged to produce 1600x1080 for further scaling
(sounds odd but might be accurate?)

Wipeout HD = dynamic-framerate-dependent 1080p framebuffer (1280x1080 to 1920x1080)

Lair = 800x1080 (2xAA) - AA buffers are merged to produce 1600x1080 for further scaling
(sounds odd but might be accurate?)

Wipeout HD = dynamic-framerate-dependent 1080p framebuffer (1280x1080 to 1920x1080)

Metal Gear Solid 4: Guns of the Patriot = 1024x768 (2xAA, temporal)

I can't believe people are settling for the current generation's graphics when they could be so much better. Think about a Grand Theft Auto type of open world game with character models as good as a one-one-one fighting game, that would look amazing. It's not necessarily about photorealism - it's about pushing the boundaries of current technology.

And the new consoles should at least render games natively in 1080p. The current generation can't even do that with the vast majority of games: http://forum.beyond3d.com/showthread.php?t=46241 Indeed, most games render natively at 720p, when virtually no TV available does 720p natively. So you end up with a horribly stretched out picture with most displays.

That was proven false
http://forum.beyond3d.com/showpost.php?p=1549251&postcount=491

Running up-to-date software would probably be a good start. The rest isn't rocket science either. Creating secure networks is not some esoteric art. I mean, plenty companies out there run their servers for years without having issues like that. Some even do it on *gasp* Windows servers! Maybe Sony needs to hire some of people who manage that?

There are good evidences that their servers were up to date:

http://forum.beyond3d.com/showpost.php?p=1549251&postcount=491
http://www.quartertothree.com/game-talk/showpost.php?p=2673715&postcount=961

Noboby has fully assessed what happened. Nobody but the usual mythomaniac guys that crowd the big net.

While there may be a few, the vast majority do NOT run at 1080p natively. Not even close. Even many games that are commonly thought to run at 1080p native, such as Gran Turisimo 5, only have certain parts of the game that run at 1080p native (menus, etc).

Here's a list of PS3 games and resolutions. I'm only counting a few sports games that honestly run at 1080p.
http://forum.beyond3d.com/showthread.php?t=46241

Here's the master list.

Here's a short list of recent games:

Alan Wake = 960x540 (4xAA)
Aliens Vs Predator = 1120x630 (custom AA)
Alone in the Dark = 1040x600 (2xAA)
Call of Duty: Modern Warfare = 1024x600 (2xAA)
Call of Duty: Modern Warfare 2 = 1024x600 (2xAA)
Call of Duty: World at War = 1024x600 (2x AA)
Darksiders = 1152x648 (no AA)
Don King Prizefighter = 1024x630 (2xAA)
Far Cry 2 = 1280x696 (2xAA, black borders)
Final Fantasy XIII = 1024x576 (2xAA)
Marvel: Ultimate Alliance 2 = 1024x576 (2xAA)
MK vs DC = 1040x624 (2xAA)
Overlord II = 1280x612 (custom edge blur, black borders)
Saint's Row 2 = 640p (2xAA)
Splinter Cell: Conviction = 1024X576 (2xAA)
Split-Second = 1280x672 (2xAA)
Star Ocean 4 = 1248x702 (2xAA + edge blur/DOF, in field), ~936x512 (no AA, in battle)
Wanted = 1120x640 (no AA)

It seems to be a prevailing trend that games either get more complex and lose resolution, or else they end up looking dated (just look at Red Dead Redemption, which looks the same as anything released on the PC in 2006).

It's time for a hardware upgrade, damn it. People are buying 1080p TVs left and right, but hardly any new games do more than 720p, let alone reach that benchmark.

http://cellperformance.beyond3d.com/articles/2006/06/understanding-strict-aliasing.html

A good list here.

You'll be disappointed about MW2.

I think you are referring to US Patent 6300956 - Stochastic level of detail in computer animation, from 2001. So maybe Pixar has a monopoly on a specific stochastic technique, but you can always invent a better one or go around it using non-stochastic optimisation. I'll have to find it, but some time in the last week I saw an infinite zoom, infinite object demo which allowed this in real-time without using stochastic anything.

Pixar is a movie company - they don't even need optimization except on the dailies which don't have to be perfect. The final render on an animated movie doesn't have to be efficient, just good, so it doesn't even make sense why they would be pursuing this. If one idea is patented, invent around it and surpass it. GIF was patented, so PNG was born. If not for patents, we wouldn't have PNG. There's your innovation.

Unfortunately, algorithms are often dictated by mathematical properties of the world we are living in. When two people solve the same mathematical problem, it's not all too surprising that they tend to arrive at the same solution.

You're not solving a mathematical problem with a set answer. How many algorithms do we have for calculating PI? More than I ever thought possible. It's just the simple measure and divide problem, so why is it so complicated? Because there are many solutions to the same problem. When ID Games invented the optimized inverse square root for their game engines, they let CPUs do what they could never have done without it, and gamers rejoiced. It is strongly dependent on the "first approximation" for Newton-Raphson iteration, which with a clever "first approximation" does not need iteration. There are lots of ways to optimize inverse square root in software, some better than others.

http://www.beyond3d.com/content/articles/8/

PS The companies that maintain usable, marketable products are typically the ones that take good design and merge it with existing patents, usually cross-licensed. Apple tends to try to buy little companies for their patents, but iPod is based on a patent minefield, just done better than the competition and thus more profitable. So you can in theory invest in marketable companies rather than smart ones that produce intellectual property, but either way the IP companies get paid - royalties or being purchased, it doesn't matter.

All of software comes down to an algorithm ultimately - there's one that detects which key I'm pressing and tells the BIOS a key code, and one that recognizes the key code, and translates that into ASCII or Unicode, and the FireFox window responds by adding characters to this text box, and then FireFox collects them all and sends them to Slashdot. It checks for updates at a regular interval, renders images predictably given standard input - it understands HTTP and FTP communication - it's an algorithm.

But there's a reason you can't patent algorightms, even complicated ones. But a cotton gin is patentable, even though it's basically "use an automated arm to brush seeds out of cotton." That's a boring program to write, more exciting than farmville, but revolutionary for its time. So we give a temporary monopoly, and now it's free for anyone to use. Or improve on, or invent a better one.

I wonder if we will see NVidia in 5 years at all in the PC market they might end up being a second PowerVR still healthy in the embedded sector but not at all present on the PC side of things.

Ironically, if the folks on Beyond3D are right PowerVR are still in the PC market in the Intel Lincroft chipset with dedicated 3D, video decode and video encode IP.

Some ps3 game already uses the Cell for video stuff

You are comparing the PS3 CONSTANT speed, to the MAXIMUM speed of the 360.

The 360 DVD, like most DVD reads tracks on the outer edge of the DVD faster than it reads them on the inside.

BD discs are read at a constant speed, from start to finish.

The 360 *average*, on a full disk, around 64Mbps.

and finally, Xbox uses 12xDVD, not 16 as you used.

http://forum.beyond3d.com/showthread.php?t=42157

They say that one of the benefits, if not an outright goal, of some Linux distributions is to be a great platform to develop software on.

At one time Linux looked downright competitive as a platform (if certainly not market,) so what happended since 2002?

I do think one thing that would help is for OSS games to have much better tools. Make it easy for people to add assets, build levels and so on. Maybe more people would be willing to do so.

Well, games are not just software. The software is simply there to make the game go.

Perhaps the reason that there are so few (or in some opinions no) good games on Linux is that for developing games, Linux sucks?

Perhaps it is time to admit that OpenGL is a not the only kid on the block and start providing another popular API that other developers want to use?

Perhaps it is time to stop throwing away all that boatload of artwork with each release and start saving anything under a usable license to an appropriate gathering spot?

Perhaps it is time to put down that cumulative-xml2pd-custom-package-colored-pretty-printer patch and answer some basic questions in such a way that new people don't hate us?

What happened to those Open Source game engines that were going to let you MOD your way for WoW 2.0? Perhaps they are still there, waiting for content.

Perhaps what Linux Gaming needs is a little less CompSci and a little more Bachelors of Arts?

For some years ago Slashdot had a news article about this http://www.beyond3d.com/content/articles/8/

So let's go from the other direction... WHICH APPLICATION wouldn't mind if it randomly did random things. You know, a simple IF-THEN will evaluate wrong, or a simple GOTO will go to somewhere else (or not go to at all) or a WHILE...DO will quit before being done...

I have no idea which application (either software or real life institution) would go for this.

There are different kinds of instructions. Even today, floating point operations are done in a different area than jumping instructions like IFs and GOTOs.

You can imagine a faster or cheaper chip that requires less energy that does floating point operations with 99% accuracy - or maybe even with an accuracy of your choice. So in your example, the code part that decides what frame to encode runs on the 100% accurate chip, the code part that creates the waveform or whatever runs on the chip with 99% accuracy. My ear probably couldn't tell a perfect waveform form a slightly different one.

Or think about the brightness or color in a video clip - if the errors are small enough, it will just look like film grain.

In video games, approximations are already being used, because you probably don't care if your rocket is flying with 10 units per second or 9.996 units per second.

Actually no. You're wrong. Try knowing something about what you're talking abouut.

I do, which is why I'm right.

DirectX 10 would require fundamental changes to the XP driver model and kernel to work, after which you'd have done a significant chunk of the work to get Vista.

This will help you:

You'll have to excuse me if I don't consider the tech equivalent of the Weekly World News as a reliable source of information. Especially when they're talking about Microsoft.

Here, try something that actually has a bit of real information and testing behind it. Here's more.

The 3D is licensed from PowerVR,

but the driver seems to have been written by a third party.

ID appears to be done with Linux: John Carmack not as interested in Linux (and OpenGL)

iPhone 3.0. Actually the current iPhone uses Power VR MBX and the new one is rumored to be using the Power VR SGX graphics. The Power VR VXD video IP core can supposedly "supports 1080p H.264 Main/High Profile decoding, as well as VC-1 and a variety of other standards" http://www.beyond3d.com/content/news/638 http://www.appleinsider.com/articles/08/04/30/apples_bionic_arm_to_muscle_advanced_gaming_graphics_into_iphones.html

Just like with current rasterization? Even current mainboards and laptops have rasterization accelerators (modern gpu) built in. So why not go ahead and make ray tracing units in gpu? They have already vertex shaders, pixel shaders, ray shaders are next logical step. Of course there are software solutions for rasterization too, like Swift shader. And they are comparable to raytracing solutions in speed :D

Quoting from the excellent article posted by the parent:

A serious problem that is often overlooked when demonstrating a ray traced game like Quake 3 RT is that while ray tracing scales well for static scenes, it does not for moving scenes. The most important operation in ray tracing is to determine the closest triangle from the origin of any particular ray. This is generally done by building a partitioning tree which allows you to quickly disregard most of the triangles in the scene without actually doing the expensive intersection test. The problem is that computing these trees is expensive, and any change to the location of triangles inside the tree means recomputing the entire partition tree.

The author of that article (Dean Calver) repeatedly gives examples (like the one above) to how doing ray-tracing in real-time for motion scenes scales much worse than rasterization.

I think that this post is just another one in a series of FUD tactics by Intel to set the stage for Larrabee. Besides nothing prevents current one chip GPGPU systems (like CUDA from NVIDIA and CTM from ATI) from being better than that really expensive cluster.

Although I have a hard time arguing in the realm of 3D lighting, I will direct attention to the Beyond3D article, Real-Time Ray Tracing: Holy Grail or Fool's Errand?. Far be it of me to claim that this article applies to all situations of 3D lighting, it may be that Ray Tracing is the best choice for games, but I for one am glad to see an article that atleast looks into the possibility that Ray Tracing is not the best solution; I hate to just assume such things. Indeed, the article concludes that Ray Tracing has its own limitations and that a hybrid with rasterisation techniques would be superior to one or the other.

This lends a bit of credence to the rumored NVidia G9 series launch, although I still think February is unlikely.

It was tracked down to Greg Walsh, with help from Cleve Moler.
http://www.beyond3d.com/content/articles/15/

As much as I would like to see it happen, there actually are technical challenges to just straight getting Direct3D 10 on XP. While we certainly now know that it is possible to see the same pretty graphics, Direct3D 10 itself is not a possibility on XP without major overhauls. Shiny DirectX 10 Graphics you get from Direct3D 10 are, but not Direct3D 10 itself, since Microsoft has zero incentive to do the kind of work necessary to get it to work on XP, no matter how small Vista usage is. This Beyond3D article explains why.

I can't comment on the quality of ATI's *nix drivers, but FWIW I've never had any trouble with their win32 drivers. My x800xt has served me well the last three years and it still ticks on nicely.

However, much to my regret the ATI of today seems to be a mere shadow of its former self. Given ATI's failure to meet expected release dates with the last two generations, the somewhat disappointing performance of both families when finally released, and the latest string of stories of senior employees signing off, I can't help but think of Netscape's fall from grace and Jamie Zawinski's letter of resignation:

The magic was gone, as the magicians had either moved on to more compelling companies, or were having their voices lost in the din of the crowd [...]

JWZ of course wrote the piece before Firefox saw the light of dawn, so...

Wow, I wonder how many people actually get that reference (which, by the way, wasn't created by Carmack)?

http://www.beyond3d.com/content/articles/8/

The only "Vista only" games that I have seen cracked were not DirectX 10 only games. The games supported DX9 and that is why it was able to be cracked to run on XP. They should have just released the games supporting both XP and Vista natively, but oh well. Here's an interesting article on some of the problems putting DX10 on XP.

I thought I read somewhere that TSMC was gearing up to do 32nm fabrication on contract for other companies. Here's a reference to assure me I have some sliver of sanity left. 45nm by September, and 32nm by Q4 of '09. So it seems that at least one company might be an option for outsourcing some fabrication.

Chartered Semi just signed another tech partnership with IBM, Samsung, Infineon, and Freescale. This one goes down to 32nm.

UMC and TI are working on 32nm together, too.

Fujitsu, although not especially known for fabbing chips for third parties, is working on getting down to 32nm as well. They do some fabbing for others now.

In any case, this story at Fabtech gives a much more reasoned and insightful look at the issues. They says it's likely AMD will outsource lower-end CPUs and continue to outsourc emuch of the GPU business as ATI already did. They may ramp up more outsourced work to Chartered than they currently do, and may share some fab space at Dresden and in New York. That's a far cry from going fully fabless.

Nintendo's warchest is peanuts compared to Sony and MS. That's what allows Sony and MS to hemorrhage at the seams; their game division is but one of many. They won't leave this game simply because that one division isn't turning a profit this quarter or year. Nintendo lives or dies by its games division since that's all it has.

Incidentally, this pretty much dictates that Nintendo has to be more conservative economically with their offerings. MS and Sony can, instead, gamble, be a bit more aggressive and hope it pays off in 3-4 years rather than next quarter.

I'm willing to bet that actually you know absolutely nothing about this. Nintendo's stock is traded over-the-counter (OTC) on American exchanges, and they do not file regulatory financial information with the SEC, not being an American company. So finding financial information about Nintendo is very difficult unless you speak Japanese. Try it: go to yahoo finance's page on them. There's like nothing there. This is why I have not invested in them, despite my enthusiasm for their business--too risky for me, trading in a company that I can't get financial info for.

The best English info I can find is here. If accurate, then Nintendo had over 7 billion in the bank at the end of fiscal 2005--two years ago. An English summary of the fiscal 2007 results are here, and they indicate the net income was 174.290 Billion yen, or roughly 1.43 Billion dollars, up from 809 Million dollars the year before. Add it all up, and it indicates that they currently have between 9 and 9.5 Billion dollars in cash. Cash. As in in the bank cash.

Now Sony's information is easier to find, so I won't provide links. Currently, Sony has about 10.5 Billion in the bank. And that's for all of Sony, keep in mind. Admittedly, Sony has a shit ton of other assets that Nintendo probably doesn't (but it's hard to say unless you read Japanese), but cash in bank is what you can spend/risk easily, which is why you brought it up. Now, Nintendo's year on year growth was 150% in operating income, while Sony saw about a 69% *decline* in operating income. In absolute numbers, Nintendo's operating income was almost 3 and a quarter times larger than Sony's (1.86 Billion dollars vs 579 Million dollars). Oops... Which means it is perfectly plausible, maybe even likely, that Nintendo will have a larger war chest than Sony within the next year. Think about that. Nintendo, who *only makes games* could have more cash in the bank than Sony.

However, MicroSoft is a different story. Even after giving out a bunch of dividends a while back, they've still got 30 some odd Billion in the bank. More importantly, though, is that their yearly revenue is *much* higher. Their operating income was a shade less than 16.5 Billion (almost nine times as much as Nintendo's) and is steadily marching upwards, though not quite as dramatically as Nintendo's. So in theory, yes, MicroSoft could give away an XBox 360 to every household in the major markets it targets that could possibly want one, say 100 Million for easy math, and let's say each cost $300 (I don't think they cost that much anymore), for 30 Billion dollars, which it has in the bank with room to spare. Problem is, they *still* wouldn't own the whole market. 100 Million free XBox 360s might kill Sony completely, but still wouldn't kill the Wii or the DS. As Nintendo shows, you can make 1 Billion a year on nothing but games--but a lot of that comes from portables, which MS doesn't offer. It would take a lot of years to make up a $30 Billion giveaway, and what happens for the next gen? Everybody's expecting a freebie, right? Seeing as how MS is already $4 Billion in the hole from the first XBox, you can understand that they're hesitant to keep hemorrhaging money on the division.

All in all, Sony

So you're saying that if DX10 was designed "properly", every video card would perform identically? I'm sorry, but that's a stupid thing to say, and Slashdot's the only place you'd get modded up for saying so just because it sounds like you're bashing Microsoft.

The bigger picture is that these cards are completely compatible with each other, sporting exposed feature sets that are practically identical from a software point of view, making it much easier to program DX10 games that work on both. This is a big step forward for the entire industry. The only important difference between the cards is performance, and that's exactly as it should be. Furthermore, the performance differences are rarely bigger than a factor of 2 in the real world. It's not as if one card is useless compared to the other, fanboy whining aside.

If it's detailed benchmark information you want, you'll get far more details than you'll even understand at Beyond3D (G80 review (AKA GeForce 8800 GTX), R600 review (AKA Radeon HD 2900 XTX)). But I guarantee you don't really want benchmarks that detailed. Analysis is impossible becuase these architectures are simply too complex to predict real-world performance from low-level hardware details. Like it or not, the only way to quantify real-world performance is game benchmarks. Just don't get your benchmarks from involved parties like ATI/NVidia. Get them from hardware review sites; it's not like there's a shortage of those. Most sites now do image quality analysis to combat driver cheating.

So you're saying that if DX10 was designed "properly", every video card would perform identically? I'm sorry, but that's a stupid thing to say, and Slashdot's the only place you'd get modded up for saying so just because it sounds like you're bashing Microsoft.

The bigger picture is that these cards are completely compatible with each other, sporting exposed feature sets that are practically identical from a software point of view, making it much easier to program DX10 games that work on both. This is a big step forward for the entire industry. The only important difference between the cards is performance, and that's exactly as it should be. Furthermore, the performance differences are rarely bigger than a factor of 2 in the real world. It's not as if one card is useless compared to the other, fanboy whining aside.

If it's detailed benchmark information you want, you'll get far more details than you'll even understand at Beyond3D (G80 review (AKA GeForce 8800 GTX), R600 review (AKA Radeon HD 2900 XTX)). But I guarantee you don't really want benchmarks that detailed. Analysis is impossible becuase these architectures are simply too complex to predict real-world performance from low-level hardware details. Like it or not, the only way to quantify real-world performance is game benchmarks. Just don't get your benchmarks from involved parties like ATI/NVidia. Get them from hardware review sites; it's not like there's a shortage of those. Most sites now do image quality analysis to combat driver cheating.

"Not many PCs around with the processing power of the PS3 or even the xbox360. Both have very high end graphic cards on dedicated buses. The only thing PCs excel in hardware against the current two nextgen consoles is memory and storage, but even then the memory on the consoles is incredibly fast."

I really don't want to go into a specific analysis of the difference in processing power when comparing the CPU's. So I'll give you a few details that lead me to arrive at my opinion that they aren't as fast.

1. Some notes about their processors:
The PS3 cell processor while it has 7 SPE cores there is only 1 PPE core, and no out of order execution. Which means that for a good chunk of game programming the cell is going to go very under utilized. Here is an article highlighting some of the issues with the cell as a console processor: http://arstechnica.com/news.ars/post/20070221-8896 .html

The Xenon processor from the xbox 360 while more at home in a gaming console then the cell still is not perfect for gaming. Due to the issues of having a deeply piped lines processor like the xenon and having to drop much of the branch prediction and having a smaller cache to keep die size down, the Xenon is poorly designed to deal with branchy code such as game control, AI, and physics code. Here is an article discussing the Xenon: http://arstechnica.com/articles/paedia/cpu/xbox360 -2.ars/6

So while the processors for these consoles may be more specifically designed for gaming. Due to cost savings on the die size they have had to drop much of the logic and caching that would have made them well rounded gaming processors like you get with a PC. The reality is that with the speed of development of processors by the time the consoles actually hit the market their processors were only marginally better for gaming then a PC processor, and required FAR more specialized programming in order to take advantage of that performance gap. The cell processor will likely never see its full potential used in a video game.

2. Some notes on their GPU's:
The Xenos GPU which is in the xbox 360 suffers a similar story to the Xenon, in that it is designed for the console but with cost in mind. The new line of ATI graphics cards that are directx 10 compatible surely borrow a lot of the design innovations of the Xenos in how the graphics pipeline is arranged for efficiency. However the new graphics chips are clocked higher have more ram at their disposal and your even able to double down on graphics cards if you still don't have enough power. Simply put it would make no since for ATI not to release a GPU that is as good or better for PCs when they make far more money off the PC graphics card market, and they are in stiff competition with nVidia. Here is an article discussing the design choices of the Xenos GPU: http://www.beyond3d.com/content/articles/4/11

In conclusion while consoles are closing the gap between PC and Console graphics, the PC market does offer graphics performance that is better then consoles and for a few reasons. The price spent on PC components is much higher then that of consoles so there are far less price restrictions on the size of the die for the CPU/GPU yielding much better general purpose processing. Consoles have a 5 year life span which enables PCs to surpass any advantage consoles may grab a very short time into their life span, and very easily because the same companies that the design the GPU for consoles also design GPU's for PCs. So until there is a huge surge in console gaming that decreases the required life cycle of consoles, and the price consumers are willing to spend on consoles goes up, console graphics just won't be able to best computer gaming for any meaningful length of time.

I tried that. And I heard of a texture tweak

http://forum.beyond3d.com/showthread.php?t=12732&h ighlight=humus+doom+3

Actually, even on the new card, a 6800, with all the tweaks I could find, I still had to run in medium quality and it was a bit choppy when anything happened.

But the fact that you need to do all this when contemporary games run ok make it seems that Doom 3 wasn't really that well optimized out of the box. In the end, I got bored with the gameplay and stopped playing.

Arun Demeure writes "Beyond3D has once again obtained new information...

If you are going to submit your own articles to Slashdot, at least have the decency to admit this instead of talking about yourself in the third-person.

As someone already said Intel's onboard stuff is the best out there. Especially if you are building a new computer, which propably don't have anything but pci-e slots.
With older hardware there were more options. I have an old Ati Radeon 9200 in my closet, just in case I need AGP graphics card. It's the last Radeon that works completely with open source drivers. (also 7000 and 8500 and 9000 work).

So it doesn't look too good, does it..

Well there is hope. Intel is working on discrete graphics chips. read more here and here.
I believe Intel has no reason to change their Linux friendly policy. So I hope they come up with a decent discrete graphics card and release open source drivers with it.
Since Intel is such a big player it just might encourage others to do the same.

You dumbfuck

you tried to save face by linking to a fucking zdnet.com article?

Run PS2 game on pre-1.5 on 480p set - perfect image quality.

There was no change to PS1 or PS2 emulation with any firmware update - due to the fact that PS2 emulation is DONE IN HARDWARE.

Don't bother replying dimwit.

The PS3 decodes Blu-Ray (and DVD) content in firmware. Firmware is updateable. Any problems you may consider exist in the firmware are entirely solveable. I know of at least one thread that goes into great detail as to what Sony engineers are doing to improve the playback. And that includes supporting various scaling options.

Besides, the PS3 has sold approximately 1.5 million consoles since launch, which doesn't even include Europe where Blu-Ray and HD-DVD market penetration is virtually zero but will be 99%+ Blu-Ray come March. It doesn't take much to see that the writing is on the wall for HD-DVD unless something pretty spectactular happens for the format.

Third, Sony's balking at allowing prawn into the format (like they did for betamax), which is (IMHO) likely to deal them another severe blow. It seems like they have developed an unmatchable expertise at shooting themselves in the foot.

Google "Vivid and blu-ray". Seems some porn companies intend to put out discs in the format. Not that porn companies even care these days what format to back. Porn is content. Porn companies are content providers, and they don't care what pipeline you get it through as long as you get it.

Except that some TVs can output in 1080p/24. So they can show the movie at the same frame rate as it appeared in the cinema. Getting a player to output in that is another matter. The PS3 can't (at the moment), but allegedly a firmware patch will add that support (see here for details).