The CPU
Of course, the Central Processing Unit, the heart of every computer or console. Most of the calculations take place here. The XBox has a Intel processor which runs at a clock-speed of 733MHz. That's a lot higher than the 300MHz at which the PS2 CPU is running. But does that make the CPU better? Not at all...
Here's why the PS2 CPU (Emotion Engine) is a lot more powerful:
-Data bus, cache memory as well as all registers are 128 bits on the PS2 CPU while the XBox CPU is 32 bits.
-It has a max. performance of 6.2GFLOPS while the XBox CPU can only do a bit over 3 GFLOPS.
-It incorporates two 64-bit integer units (IU) with a 128-bit SIMD multi-media command unit, two independent floating point vector calculation units (VU0, VU1), an MPEG 2 decoder circuit (Image Processing Unit/IPU) and high performance DMA controllers. Yes, this is all on the emotion engine itself.
Okay now what does this mean? It means that the PS2 can handle heavier physics and 3D engines (and can do more accurate realistic visual effects like splashing water and explosions). It also means that the PS2 can handle a lot more sophisticated Artificial Intelligence programming so that you have intelligent human-like opponents. And with a floating point calculation performance of 6.2GFLOPS/second, the overall calculation performance of this new CPU matches that of a super computer. This is a completely new CPU architecture especially designed for sophisticated graphics and physics while the architecture of the XBox CPU is pretty old and simple (it's a little less powerful than a standard PentiumIII processor). The architecture of the Emotion Engine really is very sophisticated so I'm not going to explain it in detail here. But simply put the main advantage of the PS2 CPU is that it is subdivided into lots of other tiny powerful processors, all of them designed to do a special task and almost all of them can work independently from eachother. And another thing... the processor inside the box does not say "Pentium III" anywhere. It simply reads "Intel". The XBox's processor is NOT an Intel Pentium III, as Microsoft would have you believe, but in fact a Celeron II. It is a 700mhz Celeron, complete with 128kb of L2 cache (P3 coppermines actually have 256kb L2 cache), but overclocked to a 133mhz FSB, resulting in PIII/Celeron hybrid. What makes it a Celeron II is the fact that it is still using a Coppermine Core, with 8 way set associative L2 cache rather than your typical Celeron 4 way set Level 2 cache. What it ultimately comes down to is that this Coppermine core, which allows Microsoft to market the XBox as a PIII Coppermine, is about a 10% speed increase over the Celeron equivalent of this processor. Is the XBox CPU a Celeron? Not really. Is it a Pentium III CPU in the sense that everyone thinks of a PIII Coppermine? Nope. It's somewhere in between and while it is great to run standard applications and some PC games on the side, it's not a good thing to have it as a CPU in a gaming console...
The Graphics Chip and VRAM
This is where the images are rendered. The XBox uses an Nvidia Graphics Processing Unit running at 250MHz and the PS2 uses the Graphics Synthesizer running at 150MHz. Again, judging by these specs the XBox looks better. The XBox GPU has a few advantages (or maybe not) over the PS2 GS, for example:
-The XBox GPU can do 125 million polygons (according to Microsoft) while the PS2 GS can only do 75million polygons
-The XBox GPU has a max. resolution of 1920x1080 and the PS2 GS can do 1280x1024, the rest of the graphics chip will be comparable to NV-20 chip.
There are alot of neat effects the XBox GPU can do with its hardware, but all those effects can be done by the Emotion Engine in software too (while the XBox' CPU is not powerful enough to do complex visual effects in software). But the catch is that these advantages (talking about higher resolutions here) don't make a lot of difference on a TV screen, even on an HDTV screen the difference would be barely noticeable (when the console's hardware is used properly). So, is the XBox Graphics Processing Unit better than the PS2 GS? It doesn't look like it, the architecture of the PS2 GS looks far more advanced. For example, PS2 has a parallel rendering engine that contains a 2,560 bit wide data bus that is 20 times the size of leading PC-based graphics accelerators. The Graphics Synthesizer architecture can execute recursive multi-pass rendering processing and filter operations at a very fast speed without the assistance of the main CPU or main bus access. In the past, this level of real-time performance was only achieved when using very expensive, high performance, dedicated graphics workstations. There is a 48-Gigabyte/sec memory access bandwidth achieved via the integration of the pixel logic and the video memory on a single high performance chip. The quality of the resulting screen image is comparable to high quality pre-rendered 3D graphics. (that is once the game developers have learned how to use it properly) There has also been a misunderstanding about the VideoRAM on the PS2. The VRAM is included in the 32MB of main RAM on the CPU (the developer chooses how much of it he wants to dedicate to VRAM). Everyone thought the 4MB of memory on the GS was the VRAM while that is just a buffer in which all the rendering is done so no external bandwidth is needed (only for texture streaming). Another rumor that's been spread by several gaming sites is that the XBox is capable of texture compression and full scene anti-aliasing while the PS2 isn't. This is simply not true. The PS2 can compress/decompress textures and do full scene anti-aliasing without causing as much slow-down as on the XBox. And although the XBox GPU can do a lot of effects that are not 'built-in' in the PS2 GS, the PS2 can do all these effects and more in software mode (but at least at the same quality) through the Emotion Engine. XBox fanboys will probably tell you that the XBox GPU is more powerful because of its vertex shaders while the coprocessors on the Emotion Engine of the PS2 can be used to get the same effects as the XBox' vertex shaders (but the vertex shaders can't do everything that the EE's coprocessors can do).
Now let's take a look at how Microsoft got the idea that their graphics chip can do 125 million polygons, because this is a little unclear... (I'm going to go in some tech details now) The PS2's Graphics Synthesizer has the highest pixel fill rate of the next generation of consoles. Most remeber the 4.0 GPixels on Microsoft's spec comparence sheet. Well, Microsoft was nice to include a "(anti-aliased)" next to it. What does "4.0 GPixels (anti-aliased)", mean? It's misleading. The Xbox has hardwired 4x FSAA, when this is turned on the actual total of 1.0 GPixels is re-rendered 4 times to remove aliasing. Another possible reason for Microsoft to say Xbox's fill-rate is 4 GPixels per second. Is that the 1 GPixels is with 2 texture layers, if it is NOT used Xbox would not gain any performance and if it is used Xbox wouldn't lose any performance. It remains 1.0 GPixels w/ 2 textures, so what MS possibly did was it doubled the fill rate twice. Trying to compare it to PS2's fill rate w/ no texture. What MS did was it came up with misleading numbers. The Xbox can't go higher than 1 GPixels per second. The NV2a in the Xbox has 4 pixel units running at 250 MHz, that's 1 billion pixels/second. While the GS in the PS2 has 16 pixel units running at 150 MHz, which is 2.4 billion pixels every second. Now let's talk about polygons. Right here I'm talking about polygon rendering and not polygon transformations. To calculate polygon rendering performance, you take the pixel fill rate, and write it in millions. So PS2s pixel fill rate is 2400 Million. When Sony says polygons, it is refering to 32 pixel polygons. Divide 2400 Million by 32. You get 75 Million (32-pixel) polygons per second. That is raw and doesn't include textures, they use up pixels also. Now let's take Microsoft's allegged pixel fill rate of 4000 Million, which MS has on it's spec sheet and divide it by 32, you get, yes you guessed it, 125 Million (32 pixel) polygons per second. Here's the problem, the NV2a doesn't have a 4000 M fill rate but a 1000 M fill rate. So it's 31 Million (32 pixel) polygons per second. This isn't raw, since there's also 2 texture units for each pixel unit. So that's 31 million with 2 texture layers, the PS2 is around 38 Million with 1 texture layer and 20 million with 2 texture layers. The Xbox maxes out at 31 MPolygons per second, if textures aren't placed on those polygons- Xbox will not gain a polygon rendering increase in performance. The PS2's Graphic Synthesizer could render 75 MPolygons per second with no texture. The NV2a in the Xbox can't render higher than 31 MPolygons per second at all.
Okay now take that all into account and then check out the following... The XBox graphics chip does not really give you the same power you get out of a GeForce3 3D accelerator card. It is only a graphics chip, similar to the one on that card but it shares its memory with the XBox's system RAM and has a 250mhz RAMDAC. While a standard GeForce3 accelarator card gives you an additional 64MB of video memory with 350MHz RAMDAC. The NV2A compensates for this by having a second vertex shader, as opposed by the GeForce3's single vertex shader. However, Microsoft claims that this second vertex shader instantly bumps the XBox's theoretical max poly count from the 31 million that Nvidia lists for the GeForce 3, all the way up to 125 million pps. According to most experts, the area that will actually see the most improvement from this will actually be in Bump Mapping. Microsoft has yet to explain how the second vertex shader yields an additional 94 million polygons per second." I don't know enough to go more in detail about this but this is definately an interesting point, and it seems that the XBox doesn't have the advantage here.
I can understand that this is all a bit confusing if you're not a real tech-freak. It comes down to this: when developers have learned how to use the power of the PS2 GS properly they'll get a lot more out of it than XBox developers will get out of the XBox GPU. The PS2 GS combined with the EE can do a lot more advanced visual effects than the XBox GPU combined with its CPU.
The RAM
This is the main memory of a console or computer system. There isn't much to say about the RAM. XBox has twice as much RAM as the PS2. Will this give the XBox a huge advantage? Not really, let's take a look at how the PS2 accesses the RAM:
-32MB Direct RDRAM 2 channels at 800MHz
This means that the PS2's powerful Emotion Engine can manipulate the data stored in the RAM fast enough to compete with the XBox' memory access time. This is very important cause all data is stored there (even the graphics because the VRAM is included in those 32MB of RAM). Judging by the information that Microsoft and Sont have released it looks like the PS2 can also compress and decompress images faster than the XBox because of the implementation of the MPEG2 decoder on the CPU. And you really have to take the extremely fast (48GB/sec) VRAM of the PS2 Graphics Synthesizer into account because this eliminates a lot of bandwidth problems developers could have on the XBox, also most parts of the Emotion Engine can independently transfer data to the Graphics Synthesizer. But then again the XBox has 6.4GB/sec memory bandwidth on the RAM while the PS2 has 3.2GB/sec there. So it's difficult to say which console has the real advantage here, some developers say they like the XBox more because it has more memory, others say that even if the XBox had 128MB of RAM they would still prefer PS2's memory system.
The Sound Chip
This is where the XBox does beat the PS2. The XBox has 64 sound channels while the PS2 has 48. This won't make the noticeable difference though. What will is that the XBox sound chip is designed for interactive and variable CD quality music which means that the music in games can (if the developers use the feature) change and adapt itself to the gameplay. The PS2 sound chip is more simply designed for 'precalculated' CD music. It's possible to do interactive music on the PS2, it's just a bit harder to program than on the XBox which is why you'll probably see more XBox games with interactive music. The XBox also has a few sound features like Dolby Surround and stuff but all of this has been done on the PS2 in software through the Emotion Engine too (just listen to the PS2 version of SSX Tricky, same interactive music, same Dolby Surround as on the XBox).
Backwards Compatibility and PC ports
As you probably already know, the PS2 can play all old Playstation games and even enhance the graphics and speed up loading times. This gives the PS2 a huge advantage of course. But the XBox has another advantage over the PS2: PC ports. It's very easy for a PC game developer to create an XBox version of his game because the XBox hardware design looks a lot like that of a PC. And while the PS2's hardware architecture is superior to that of the XBox, it's a lot more difficult to make a PS2 version of a PC game because of the huge difference in hardware designs. If a PC developer wants to port a PC game to the XBox he just has to make some modifications in the program code. But if a PC developer wants to port a PC game to the PS2 he has to rewrite most of the program code. But a lot of developers have gotten used to the hardware now and we're seeing some great PC to PS2 ports (Quake3, Max Payne,...).
What it comes down to:
The XBox hardware design looks a lot like that of a powerful PC, which makes it very easy for developers to program for as most of them know the hardware very well and know how to get the most out of it. The PS2 has a much more innovative, effective and powerful design of hardware but it is also much more complex. The problem here is that the programmers have a lot to learn about the hardware. If we compare top-quality XBox games like Halo and Dead or Alive 3 with top-quality PS2 games like Metal Gear Solid 2 and Final Fantasy X, we see that the graphical quality is slightly better on the PS2. Some details are done better on the PS2 games but then again other details are done better on the XBox. But many PS2 programmers have stated that even the best looking games on the system that are out right now don't show the true potential of the machine. For example, the game Gran Turismo 3 on PS2 looks more photo-realistic and has more detail than Gotham Racing (XBox's prettiest racing game). But Sony's research team has developed a 'Performance Analyzer' that shows that Gran Turismo 3 doesn't even use 25% of the PS2's power. So you see there's still a lot of room for improvement and the programmers still have a lot to learn while the XBox programmers have little room for improvement because they're already using more optimized methods of coding.
What the PS2 developers have to say:
(excerpts from interviews from IGN.COM)
Jason Rubin from Naughty Dog:
Question: Have you learned any new technical tricks on the PS2 that you would like to share since we last talked?
Jason Rubin: Yes I have. I've learned a bunch of them, and some I don't even understand. The system is extremely deep, and in this generation all of the systems are deep. I have never been a bit person, you know a person who quotes bits and all that, but this generation of systems is incredibly powerful. We're currently doing tests on all sorts of things, including collision detection, artificial AI, camera positions, backgrounds, etc. The PS2 is unbelievable. It provides great simulations of the real world but also of fantasy worlds. So, for this next generation of games, people will learn things steadily. Like for instance, Crash 1 had no Z-buffer, Crash 2 had a Z-buffer. Crash 3 had new techniques in it, and Crash Team Racing had heat waves, much like those in GT3. Each new game that appears on PS2 will provide new ways to reach deeper into the system's capabilities.
---
Question: Have you had any trouble working with the PS2 hardware? And what kind of tools have you been working on?
Jason Rubin: We were very early on PS1. I think that we were the 30th project started. We were also EA's second Genesis title. Both of these starts happened before the hardware was on the shelf in Japan, and well before America was going to see the systems. This is a hard fight. New hardware is never complete. And even as it completes, you still have the uphill battle of fighting to do things that have never been done.
The PS2 is hard as well. But this is not because the hardware is difficult, it is because there are more new things that can be done than ever before.
We think that the jump from Genesis to PS1 is SMALLER than the technical hurdles between PS1 and PS2. This is NOT because the hardware is specifically harder, but because technology is really changing
We went from trying to simulate everything cheap on the PS1 to actually being able to do it on the PS2. But we have never done these things before, and that is challenging the industry.
Question: Do you think there will be any major advantages to the Dynamic hardware setup of the PS2 as opposed to traditional PC-esque hardware, it seems to making mainly problems for developers?
Jason Rubin: Yes, I do think that there are advantages to the hardware setup of the PS2. The multiprocessor setup of the system allows the PS2 to do so many things at the same time (if programmed well) that the hardware is actually far more powerful than its spec. They say the X-Box is going to be either 600mhz or 1000mhz. The PS2 is only 300mhz. But if you can have your background engine running in 10% CPU and 50% VU1 time, then you have far more power.
That is how we are doing our programming for the next game. I think that the problem with the current games (the ones you have seen) is that they have opted to take the easy way out and they are only doing CPU code.
So think of it like this. They are actually leaving a significant part of the power of the system on the table. We plan to bring that power into the game.
And yes, of course there will be more powerful systems in the future. That is how technology works. But the X-Box may not be that system.
Question: In what ways do you think the PlayStation 2 will change gaming as we know it today? Do you think that it'll be revolutionary?
Jason Rubin: The PS2 will first be evolutionary... and then later, revolutionary. First, you will see games that are similar to the past, but a whole lot better looking and smoother. Later, you will see games that do things that have never been done before.
Broadband will have a lot to do with the Revolutionary part of the cycle. We will go beyond deathmatch, and beyond the glorified chat room of EverQuest and get true multiplayer online gaming with interaction. That is something we haven't seen yet.
Question: How do you feel about the VRAM in the Ps2? Is it better than much of what's in PC's today because of its high speed? Or is it worse because of its tiny amount?
Jason Rubin: The VRAM on the PS2 is both better and worse than the PC. Sure it is smaller. A LOT smaller. But it is also much faster, which allows for more polygons, and hence better usage of the texture space.
I am not sure whether or not low polygon/high texture is better than high polygon low texture. And I guess that that is going to be the developers challenge on all systems this generation.
We are not having a problem with the texture space at Naughty Dog. I guess only time will tell whether or not this is a problem for other developers.
Question: Can you give us some tech info about Ps2?Is it really possible pushing 10+ million polygons/sec while applying FSAA, high quality textures and running a highly advanced physics engine?
Jason Rubin: To answer your question directly: Yes, I do believe that that is possible. You are talking about 166,000 polygons a frame at sixty frames a second, and I not only believe that to be true. I KNOW it to be true.
Question: Can you explain exactly why the lack of anti-aliasing and shimmering backgrounds seem apparent in the early batch of PS2 software?
Jason Rubin: Yes: The PS2 gives you a raw frame buffer. It does not blend past frames with the current frame to create the blurry effect the Dreamcast gets. For a bad TV, the Dreamcast method is great. But on a good monitor, or with a RGB cable, the Dreamcast method is limiting because it tends to take away from your ability to get small details.
The texture capabilities of the PS2 are very complicated. We have spent over 2 months simply looking at our options in regards to texture usage, and have found a good solution to the mip-mapping and filtering issues of the system.
The early PS2 games did not have time to do this. That is why these games flicker horribly. With time, and effort, you can overcome the flickering issues to a great extent.
I think that you see this with MGS2. There were not as many issues in this game as there were in the early release titles. Time will improve them all.
Question: Over the past months.. we've heard several figures for the maximum amount of polygons the PS2 can push. Where would you place the max?
Jason Rubin: At this time, I can't give a fixed number on the max for the system. That depends on many things, such as how big they are, or how much transparency there is, etc. But I will say that today... ok a few months ago... we hit the 10 million mark that you already mentioned. With time, and generational movement upward in our knowledge of the system, there is no reason to believe that this won't go MUCH higher.
bioware:
Question: About the math capabilities mentioned earlier, could you talk in more detail about them and what you'll be doing with them?
Charles: The PS2 math capabilities are like Intel's SSE... but on steroids. The math processors can be used to optimize any kind of math operations, which, in a 3d game, means a lot. So they will basically be used to speed up a lot of the 3d transformations and lighting.
David: You can also do a lot of math operations in parallel, if you order your operations right. That's something you just can't do on most consoles or PCs.
Question: What do you think of the multiple-processor structure of PS2? Bandwidth issues involved with splitting things up between the different processors seem to be an issue some developers are having trouble coming to grips with.
Charles: Well, it definately takes a while to get used to to the multiple processors, but its really just a matter of organizing your data accesses across processors. Most of the processors can run 'mini programs' without being controlled by the main processor, which allows for a lot of asynchronous processing. So while the main processor is busy doing AI, and animations, etc, the other processors can render, play sounds, and access data.
David: There are several different render paths in our engine already. There's one path optimised for static geometry renderning, one for dynamic geometry etc.
---
Reverend-IGN: Tech fun again!: In the last chat you said that low video RAM was limiting things like the textures and anti-aliasing. Isn't there some way you can take some Main RAM and assign it to be video RAM? And if not, why?
DavidBioWare: Yes, the video memory situation has improved dramatically since last time.
DavidBioWare: The problem was that there was too little video memory to fit all our textures, and the machine can't use a texture unless it's specifically in video memory.
DavidBioWare: What we've found since then is that the PS2 has enough bus bandwidth to transfer each texture from main memory to video memory as it's needed.
DavidBioWare: That's on the order to 100s of Mb per second. We hadn't anticipated that the PS2 had that kind of brute horsepower on its bus. No other machine I've used does, including any PC or the Dreamcast.
DavidBioWare: We had to reorient our thinking after that.
So now we have almost more texture memory than we know what to do with.
I guess your dumbass will say that's false. Look it up on a tech site, not a pc tech site. That's your prob idiot, you need to read up on this so you know something.
.... i dunno.... well all I could say is i'll just let him belive what he wants... true everyone has there opinion.. but somtimes you just have to face the fact that you could be wrong every now and then and obviously he hasnt learned that yet.. like the saying goes "the world doesnt revolve around you" take that as you will but its the truth...
