• Hey Guest. Check out your NeoGAF Wrapped 2025 results here!

Beyond3D Xenos Article

most of it is too much details for me but some quotes..:

One element that has been reported on is the number of 150M transistors in relation to the graphics processing elements of Xenon, however according to ATI this is not correct as the shader core itself is comprised from in the order of 232M transistors. It may be that the 150M transistor figure pertains only to the eDRAM module as with 10MB of DRAM, requiring one transistor per bit, 80M transistors will be dedicated to just the memory; when we add the memory control logic, Render Output Controllers (ROP's) and FSAA logic on top of that it may be conceivable to see an extra 70M transistors of logic in the eDRAM module.
Click to expand...

so xenos is more like 382M transistors.

With its unified shader pipeline Xenos has a fundamental difference with virtually all current shader capable graphics processors and whilst shader processing are enveloping both the geometry and raster pipeline, until now they have done both those elements distinctly. On Xenos there is a logical disconnect between the old OpenGL pipeline, which is basically the evolution path most graphics processors followed, as now the geometry and pixel shader processing are moved on to a single processing element of the chip as all the shader ALU's can dynamically be tasked with either vertex shader programs or pixel shader programs.
Click to expand...

MEMEXPORT expands the graphics pipeline further forward and in a general purpose and programmable way. For instance, one example of its operation could be to tessellate an object as well as to skin it by applying a shader to a vertex buffer, writing the results to memory as another vertex buffer, then using that buffer run a tessellation render, then run another vertex shader on that for skinning. MEMEXPORT could potentially be used to provide input to the tessellation unit itself by running a shader that calculates the tessellation factor by transforming the edges to screen space and then calculates the tessellation factor on each of the edges dependant on its screen space and feeds those results into the tessellation unit, resulting in a dynamic, screen space based tessellation routine. Other examples for its use could be to provide image based operations such as compositing, animating particles, or even operations that can alternate between the CPU and graphics processor.
Click to expand...

ATI also believe that Xenos specifically has the most advanced power management features of any chip they have produced so far. There is a top level power management system that can be controlled by the OS that allows for various elements of the pipeline to be turned off for various operations, such as DVD playback for instance. There are low power modes that regulate the speeds and voltages and, when inactive, the data is held in stasis rather than just switching transistors on and off to keep the data. However, in the graphics core itself there localized power management techniques applied at the block level to minimize power consumption during idle or low usage periods.

When we factor in the savings for both power and die size savings we can see that this potentially has some advantages over traditional architectures. In the case of the XBOX 360 not only does this result in a relatively smaller die size for a fairly high performance ratio but also means that the graphics need only be air cooled, without the use of its own additional fan. Beyond the immediate application we can see that unified designs that are bound for the PC could have smaller die sizes for equivalent performances as current discrete solutions or more silicon dedicated to either more ALU's for higher performance, or other transistors dedicated to other functionality.
Click to expand...


Now the real question is who is going to take advantage of this gpu?
 
Blimblim said:
Waiting for "Dave Baumann is a sell-out" comments :D
Click to expand...

It's not a controversial article at all, I don't think :p

A good article, I think. Perhaps would have liked a little more detail in parts, but it clarified some things. As expected, it can't really come to any definitive conclusions re. performance etc. which may disappoint some who were expecting it to be the be all and end all of Xenos articles (or to provide fanboy fodder :p - there's not so much at least), but certainly it's the best article on it yet (and likely the best we'll have before final hardware gets into "talkative" hands).

thorns said:
so xenos is more like 382M transistors.
Click to expand...

That seems to be their speculation, not something they were told by ATi. It doesn't really change much from what we knew before since the transistors beyond 232m are still being spent in the same way we were aware of before.
 
gofreak, kaching: You have missed a very nice topic a few days ago with lots of speculation about this (by that time) upcoming article. If you had seen it you'd have understood why I said that ;)
 
sounds good. The geometry only Z-pass, combined with the potentially very efficient shader layout and edram makes it sound like Xenos isn't to be trifled with.

I think the key in PS3 Vs X360 will be this Xenos being fully utilzed, Vs the RSX/CELL combination.
 
Sems like the SIMD units can do different functions, which is impressive also the memexport appears to be important, although I'm still trying to get my head round it. By the way although 232 is dedicated to logical functions the 10MB edram is fundemental to the graphical output of the GPU. 382 transistors is absoultely massive.This GPU must be costing a fortune.
 
Good read. Sounds like ATI isn't pulling any punches with C1. I would be very suprised if RSX is as custom a part (not to say it will be less powerful). MEMEXPORT sounds cool as hell! I can't wait to see developers really push this console. I am very excited for Xbox360 at this point.
 
akascream said:
MEMEXPORT sounds cool as hell! I can't wait to see developers really push this console. I am very excited for Xbox360 at this point.
Click to expand...
It will be interesting to see what shaders specialists like Masaki Kawase will be able to do with this MEMEXPORT thing. Basically it's a cumulative shader system if I understood correctly, so each pass makes the output even more difficult to "predict" and thus debug. This will be a very interesting coding challenge for shaders experts.
 
It will be very interesting to see the performance and quality of graphics it is able to produce once developers have had decent access to development kits based on the final hardware, however we suspect that it won't be until the second generation of XBOX 360 titles before we see developers being able to seriously scratch the surface of understanding the processing capabilities of Xenos and the XBOX 360 as a whole. That being said, though, much of the architecture is transparent to the developer and they shouldn't need to concern themselves much with the types of workloads they are handing to the graphics processor as this will all be handled automatically, and without stalling any part of the pipeline.


this pretty much confirms what the ATI rep said in the bit-tech interview
Good stuff indeed.

RSX has a much more traditional pc-card like architecture right?
 
Hajaz said:
RSX has a much more traditional pc-card like architecture right?
Click to expand...

Maybe. We just don't know very much about it. There is some strong two-way linkage between RSX and CELL though, which could be the real trick up PS3's sleeve.
 
Blimblim said:
It will be interesting to see what shaders specialists like Masaki Kawase will be able to do with this MEMEXPORT thing. Basically it's a cumulative shader system if I understood correctly, so each pass makes the output even more difficult to "predict" and thus debug. This will be a very interesting coding challenge for shaders experts.
Click to expand...

I don't see what's so special about MEMEXPORT. The ability to write back to a vertex buffer should of been in the shader specification in the first place instead of the replacement of a stage in the pipeline. Writing back post-processed vertices to VU RAM after transformation/vertex processing on VU1 has been an extremely awesome ability on PS2. It has allowed do many things just not possible on PC/Xbox because of this limitation. I just hope MS or ATI don't try to hide the memory via some stupid addressing scheme. Just give me linear memory, damnit!
 
Zeenbor said:
I don't see what's so special about MEMEXPORT. The ability to write back to a vertex buffer should of been in the shader specification in the first place instead of the replacement of a stage in the pipeline.
Click to expand...
MEMEXPORT is much more than that, It gives you the ability to make random writes everywhere in a shader. This is the first time a GPU has a real random scatter feature.
 
Zeenbor said:
Writing back post-processed vertices to VU RAM after transformation/vertex processing on VU1 has been an extremely awesome ability on PS2. It has allowed do many things just not possible on PC/Xbox because of this limitation.
Click to expand...
Like what? Are you talking about an "only on PS2" end result I could see in a particular game, or about a more efficient technique that gives the same end result you might see on PC/Xbox?
 
CrimsonSkies said:
Whoa. That article is so good nobody can even understand it enough to bitch and moan about it ahah.
Click to expand...

An exercise in reasonable neutrality, I think :) Obviously he couldn't touch on potential downsides too much since that's not something ATi would entertain discussion on, but it's a very level-headed article I think. Good job indeed.
 
Hajaz said:
It will be very interesting to see the performance and quality of graphics it is able to produce once developers have had decent access to development kits based on the final hardware, however we suspect that it won't be until the second generation of XBOX 360 titles before we see developers being able to seriously scratch the surface of understanding the processing capabilities of Xenos and the XBOX 360 as a whole. That being said, though, much of the architecture is transparent to the developer and they shouldn't need to concern themselves much with the types of workloads they are handing to the graphics processor as this will all be handled automatically, and without stalling any part of the pipeline.
Click to expand...

I can only imagine what the next Panzer Dragoon game would have looked like on the 360 with the Orta team developing......
 
BlimBlim said:
It will be interesting to see what shaders specialists like Masaki Kawase will be able to do with this MEMEXPORT thing. Basically it's a cumulative shader system if I understood correctly, so each pass makes the output even more difficult to "predict" and thus debug.
Click to expand...
The most exciting things about MEMEXPORT isn't about shaders. Like Nostromo says, we're looking at a new kind of computing with GPU, this goes way beyond simple graphics applications.

slidewinder said:
or about a more efficient technique that gives the same end result you might see on PC/Xbox
Click to expand...
Basically, although we're talking very different levels of efficiency on non VU architectures.
 
Shorthand version for those who don't want to read 19 pages:

ATI ownZ!

It's true, though. ATI does rock indeed.
 
Kleegamefan said:
Sounds like a monster....
Click to expand...

For seriously. It seems like they got themselves a great little machine here, which is not in any trouble of being rendered obsolete by the competiton in terms of power.
 
I agree...


*rant*

It seems to me that both Sony and MS are bringing us some rediculous technology with some amazing ablilites....

Nintendo says they don't want to go down that road because they don't believe technology/performance are the keys to innovative gameplay.....that (to me) implies Revolution will not have PS3/X360 levels of performance even though Nintendo claim they will blow the competition away with innovation....

I understand what Nintendo is trying to say but I dont 100% buy the notion of innovation(gameplay or otherwise) and technological performance being mutually exclusive...

I believe that along with greater tech comes a bigger and better "canvas" on which games can be created...

The better the performance, the more things can be simulated in a three dimentional space which can lead to greater gameplay possibilites...

Yeah, although it may be easy to label this tech as simply a shallow indulgence for graphic whoreism, but I thing greater tech can lead to better gameplay because more things can be calulated and simulated if you have greater power....

All this, of course, is just IMO

*/rant*
 
After reading the article, the info about the Xenos unified shader technology does sounds great. I think shading is the big thing that is going to occur next gen, so if X360 lacks some polys in games compared to PS3, it could more than make up for it by having a better shader technique.
 
With the capability to fetch from anywhere in memory, perform arbitrary ALU operations and write the results back to memory, in conjunction with the raw floating point performance of the large shader ALU array, the MEMEXPORT facility does have the capability to achieve a wide range of fairly complex and general purpose operations; basically any operation that can be mapped to a wide SIMD array can be fairly efficiently achieved and in comparison to previous graphics pipelines it is achieved in fewer cycles and with lower latencies. For instance, this is probably the first time that general purpose physics calculation would be achievable, with a reasonable degree of success, on a graphics processor and is a big step towards the graphics processor becoming much more like a vector co-processor to the CPU.
Click to expand...

Damn, this thing sounds like a beast!
 
I was really suprised at the extent to which, at least conceptually, they worked in PowerVR's tile-based rendering. Just another part of the "efficiency" puzzle that's nice to know about.

Also, did I read it wrong, or has MS bascially chosen to move some of the control logic from the CPU cores onto the Xenos?
 
Kleegamefan said:
I agree...

Nintendo says they don't want to go down that road because they don't believe technology/performance are the keys to innovative gameplay.....that (to me) implies Revolution will not have PS3/X360 levels of performance even though Nintendo claim they will blow the competition away with innovation....

I understand what Nintendo is trying to say but I dont 100% buy the notion of innovation(gameplay or otherwise) and technological performance being mutually exclusive...

I believe that along with greater tech comes a bigger and better "canvas" on which games can be created...

The better the performance, the more things can be simulated in a three dimentional space which can lead to greater gameplay possibilites...
Click to expand...

That's kind of what I'm thinking. Nintendo's idea of a base for innovation is seriously flawed, IMO. Nintendo claims that more powerful hardware and intuitive software is not a solid base for innovation, something I completely disagree with. Powerful hardware and intuitive software allows developers to think less about what they can't do and more about what they can do. Inspiration through freedom, or something. Nintendo's idea (if the DS is any indication) is to develop a set framework for "innovation", and thereby limiting the potential from start to finish. It seems to defeat the basic idea behind innovation.
 
slidewinder said:
Like what? Are you talking about an "only on PS2" end result I could see in a particular game, or about a more efficient technique that gives the same end result you might see on PC/Xbox?
Click to expand...

I was mainly talking about writing a general vertex program on the VU that could do skinning, UV transforms, and then any other post process. Another example would be to completely generate particles with only a few inputs to VU. With current shaders on PC/Xbox, you need an input, like a vertex position. This means you have to lock the vertex buffer and send that to the GPU for the shader to process.

I'm not saying PS2 is superior... I was just pointing out that the ability to write to RAM and read from RAM is a very nice advantage.
 
Nostromo said:
MEMEXPORT is much more than that, It gives you the ability to make random writes everywhere in a shader. This is the first time a GPU has a real random scatter feature.
Click to expand...

I didn't see that. Damn cool!
 
With tile accelerated rendering, there's no need for an embedded framebuffer, especially if display list rendering is used. The amount of eDRAM they chose for an embedded framebuffer, though, doesn't completely fit the target frame, so they're ending up using both approaches. Perhaps this combination method will be the best balance.
 
With tile accelerated rendering, there's no need for an embedded framebuffer
Click to expand...
What do you think Tile buffers are then? ;) The whole point of tiling is to use embeded buffers - small enough to make them affordable.
 
Nostromo said:
MEMEXPORT is much more than that, It gives you the ability to make random writes everywhere in a shader. This is the first time a GPU has a real random scatter feature.
Click to expand...

So enlighten me some more here. What is the significance of this? Is it in any way analogous to getting rid of the traditional memory stack FIFO deal?
 
Aww what the hell... they go to all the trouble of having edram and it still has to write the frame/tile to the main ram... and then write that to a seperate display processor? And they were trying to say the PS3 would have bandwidth troubles... Grrr...
 
You must log in or register to reply here.

Similar threads

The "deprofessionalization" of the gaming industry - article
2
89
3K
PlayStaytion replied
  • Featured
Warhammer 40,000: Space Marine 2 ≼|℧T|≽ Xenos Saga 2
|OT|  6 7 8 9 10
479
31K
The_hunter replied
GTA VI article coming out in the next couple of weeks by Tom Henderson
Opinion  2
53
2K
Audiophile replied
Las Vegas Weekly (Local Vegas Magazine) Has Article On Fallout: New Vegas Most Iconic Locations
11
571
Darkmakaimura replied
  • Locked
PlayStation has lost over 1 billion dollars between Concord & Marathon (UP: Screenshot posted is fake. No article like this exists on Forbes)
2
90
2K
digimaster7 replied
Top Bottom