PC GPU comparable to PS3

[james.miller]

i think you are trying to argue a point a never made, drunkenmaster lol

 

[drunkenmaster]

I am of the opinion that Uncharted Drakes fortune is one of the best looking console games...

Honestly have no idea what title that is or how good it looks. But frankly gpu power isn't the be all and end all. A more powerful gpu will let you get away with better lighting effects and what not. But a lighting effect on a nasty texture with horrible level design and dull characters still looks like crap. While a beautiful level design with fantastically detailed textures with no lighting effects can look insanely fantastic and use very little GPU power. How games look is 98% dictated by the skill and experience of the design team coupled with what the games engine lets them draw with 2% improvements coming from newer better lighting effects.

AS gpu power scales, engines designed with the idea they will have more power to do what they want are made and designers find it easier to make better looking textures. Newer GPU tech rarely improves the looks of games instantly, but it paves the way for better games in the future in general.

 

[drunkenmaster]

i think you are trying to argue a point a never made, drunkenmaster lol

well honestly, thats just not true. The points you made were, PS3 can't be directly compared as the Cell can in the future and even now IS used for graphics so you can't know how powerful the ps3 is. Firstly yes you can know and secondly no the Cell is in no way used now and most likely in the future it will not be used for graphics in any way shape or form.

Mostly i'm just debating its interesting, it lead me to read some very interesting stuff on Insomniac which confirms the Cell isn't used for graphics in the slightest.

For the next gen console I could well imagine a double numbered cell core but with multiple gpu units along side spu's. This is really the direction probably Intel and AMD will go, At that point you would be closer to being able to canabalise the SPU's for more gpu power as the ondie communication could eliminate most of the latency problems. but you'd be better just having more gpu cores in the first place.

 

[james.miller]

sorry but thats wrong, the shaders on the SPU at the moment aren't for graphics work AT ALL thats the incorrect part of the statement. ALL they have done is split up normal stuff you do on the cpu into the spu's in "shader style". at the moment it has NOTHING to do with graphical power and NO games are using SPU's for graphical power, and its incredibly unlikely they ever will. the ONLY reason people think it might be possible is the use of the term shader, which they only did to let you know the rough idea behind how they work, IE spreading the work across multiple massively paralel cores. it has nothing to do with shaders in a GPU and the GPU power of consoles can be directly compared.

well honestly, thats just not true. The points you made were, PS3 can't be directly compared as the Cell can in the future and even now IS used for graphics so you can't know how powerful the ps3 is. Firstly yes you can know and secondly no the Cell is in no way used now and most likely in the future it will not be used for graphics in any way shape or form.

im sorry but no we cant. numbers are all well and good but its the sun of the parts that we should be interested in. and i dont think thats something that'll be realised for a few years yet. i have not and will not make any over claims on the subject. evidently i find it as interesting as you but that's no reason to be pointlessly arguing over such a small point

 

[Darkwave]

spu's aree being used to stream textures, which does have a direct influence on graphical 'quality' there's really no need to capitalize so many words to make a point lol

I was under the impression that all texture streaming is done by the shared memory.

Correct me if I am wrong.

 

[hernaldo]

lol, I might as well add a little bit I remember from way way back.

Definately remember Sony saying that the "cell" processor was going to demand that all PS3 users be connected to each other via a broadband connection so that when you play a game, data would get "uploaded" for processing by other peoples PS3's "cell" processors so your game would look better...

Basically, you've got the power of thosands of cell processors at your finger tips...

Oh how we laughed!

(Mind you this was said around 4-5 years ago when the PS2 was at its peak in some official PS magazine)

People don't half talk a lot of ******** I must say and Sony is terrible for it (Microsoft is no better but they do it in a different way, they don't actually make statements like the above but use marketing speak) maybe its a japan\us difference in culture in the way they talk about stuff.

So basically, this BS from Insomniac sounds exactly like the BS Sony would come out with like the stuff above about the "Cell" processor.

Total load of wacko rubbish, its almost as if the PS3 is an organic life form according to them. (Bear in mind Insomnica\Naughty Dog pretty much IS Sony)

Edit - A light bulb just went off in me head, I guess in reality the fruit of this "massive parallel processing universe" is.....uhhhh...what "Folding @ Home"....?

Sorry but you have to laugh, this is in no means a critisism, or console comparison, just a little addon to the Insomnic thing that you lot are discussing, and drawing a parallel...Maybe it was poor journalism by the PS magazine spinnign the story, who knows...

 

[drunkenmaster]

All CPU's stream textures to a certain degree. Info from hard drives on a pc are loaded into the cpu to be generated before being sent to the GPU to be stored and access as needed.

The cpu sends all data relevant to rendering to the cpu to do the actual rendering.

It does seem almost a deliberate attempt to mislead people using the word "shaders" to describe what they are trying to do on the SPU's. It has clearly lead to some confusion over if the PS3 is actually a lot more powerful graphically than the 360.

To be honest, the basics here are what i said, they split up what would be larger workloads on a single core into multiple loads on smaller cores. Its just what happened from games run on single cores to games run on quad cores, but to the next level. Its not hugely exciting or new, or unexpected, the only thing that makes it seem new is the whole Shader + cpu thing.

 

[drunkenmaster]

lol, I might as well add a little bit I remember from way way back.

Definately remember Sony saying that the "cell" processor was going to demand that all PS3 users be connected to each other via a broadband connection so that when you play a game, data would get "uploaded" for processing by other peoples PS3's "cell" processors so your game would look better...

Basically, you've got the power of thosands of cell processors at your finger tips...

Oh how we laughed!

(Mind you this was said around 4-5 years ago when the PS2 was at its peak in some official PS magazine)

People don't half talk a lot of ******** I must say and Sony is terrible for it (Microsoft is no better but they do it in a different way, they don't actually make statements like the above but use marketing speak) maybe its a japan\us difference in culture in the way they talk about stuff.

So basically, this BS from Insomniac sounds exactly like the BS Sony would come out with like the stuff above about the "Cell" processor.

Total load of wacko rubbish, its almost as if the PS3 is an organic life form according to them. (Bear in mind Insomnica\Naughty Dog pretty much IS Sony)

Basically, they almost, just almost have some grounds to call it shader work as the idea is very close. ON say crysis the physics will be probably on one cpu only, and take up a certain amount of power, with ai a separate unit and so on. The SPU's are basically used to further sub-divide that single physics unit into multiple small workloads which is good and hopefully will lead to better coding for future games on quad/octo cores(as i said its more a headstart with some people coding like that now for when we get to that point we need that paralelism in games). But thats it, calling is shaders as i said seems more like a deliberate ploy to mislead people on what the Cell can do.

WHen you mentioned it I do seem to recall reading something like that years ago, sounded ridiculous then, and still does now. Or maybe again they can still get away with that description, i mean, they did put effort into making folding at home working, that kind of fits in with the whole broadband + cpu work spread across hundreds of systems thing Its just the slowest most boring game ever made. But remember, you do get a score, and a leaderboard

 

[hernaldo]

Basically, they almost, just almost have some grounds to call it shader work as the idea is very close. ON say crysis the physics will be probably on one cpu only, and take up a certain amount of power, with ai a separate unit and so on. The SPU's are basically used to further sub-divide that single physics unit into multiple small workloads which is good and hopefully will lead to better coding for future games on quad/octo cores(as i said its more a headstart with some people coding like that now for when we get to that point we need that paralelism in games). But thats it, calling is shaders as i said seems more like a deliberate ploy to mislead people on what the Cell can do.

WHen you mentioned it I do seem to recall reading something like that years ago, sounded ridiculous then, and still does now. Or maybe again they can still get away with that description, i mean, they did put effort into making folding at home working, that kind of fits in with the whole broadband + cpu work spread across hundreds of systems thing Its just the slowest most boring game ever made. But remember, you do get a score, and a leaderboard

Who knows they may have actually meant something like "folding @ home", but emphasis they were putting across was certainly meant to imply...A graphical vista beyond your wildest imagination, but we all know that sort of thing doesn't actually exist, still would be nice if miracles of science actually happened fast enough to be impressive instead of the drib drab.

There we go I guess, hey thanx for your post pretty interesting

 

[mmj_uk]

well honestly, thats just not true. The points you made were, PS3 can't be directly compared as the Cell can in the future and even now IS used for graphics so you can't know how powerful the ps3 is. Firstly yes you can know and secondly no the Cell is in no way used now and most likely in the future it will not be used for graphics in any way shape or form.

I'm afraid quite a few developers disagree with you:

The SPUs are heavily involved in the graphics pipeline and do an enormous amount of work to eliminate inefficiency before anything arrives at the PPU and RSX. For example, the SPUs are powerful enough to decompress and check every triangle [polygon] before passing it on to the RSX. Triangles that are facing away from the player, or that are not on the screen can be 'trimmed' away by the SPUs, which hugely reduces the amount of redundant work sent to the RSX. This in turn lets the RSX get on with what it does best--drawing stuff on screen.

The SPUs can also be used to augment the RSX vertex shaders, making far more vertex-heavy tasks possible which is very useful for character animation. Additionally, the SPUs can be used to implement behavior very similar to geometry shaders--F1 CE uses them in this way to render seamless interpolated levels of detail for some scene elements. So in answer to the question "Do the Cell and RSX work together?" the answer is a resounding "Yes," and I think this is one of the real strengths of Playstation 3 that we'll see increasingly exploited by development teams going forward.

The PlayStation 3 has a lot of power. When we started Uncharted we were really ambitious and had no idea what the PS3 would give us. Once we got the first devkits, we realized quickly that we could do everything we had planned to. The three main points for me are the Cell, Blu-Ray and the hard drive. We’ve been using the Cell for pretty much all our systems: rendering, particles, physics simulation, collision detection, animation, AI, decompression, water simulation, etc … and to give you an idea of the power of the PS3, we're using only 30 percent of the Cell processor.

Let me say very clearly, we find the SPUs very flexible. We do in fact do some graphics pre-processing on them, including work for culling and clipping and RSX data building as I've mentioned elsewhere. At the same time, graphics alone don't make a game, and we've got lots of other systems running on the SPUs. The SPUs are absolutely central to how we write everything now - it's no longer a matter of "porting" things to the SPU - the SPUs are where everything should go by default. We've still a lot we can do, but we're definitely taking advantage of the Cell.

For us, the most exciting part of the PS3 has been the cell processor, the SPUs specifically. In our highest density scenes right now, we are currently using about 30 percent of the SPUs' capabilities--with the SPUs doing lots of heavy lifting for us on rendering, visibility, particle systems, skinning, animation blending, and so on...this with scores of pedestrians, cars, fires, etc., all going on. And the best part? We've not made any significant attempts to even optimize the SPU code. I think it's reasonable to guess we could put 10 times as much stuff on the SPUs and still make our frame budgets. It's really pretty amazing.

If by cooperative rendering you're referring to SPUs supporting the RSX, I strongly believe that this approach will become far more widespread. In addition to reducing the vertex load on the RSX through the use of culling and vertex pre-processing, this approach also provides an efficient mechanism to introduce procedural geometry.

Historically, CPUs have provided course grain scene culling using view frustums, occlusion planes, portal visibility and BSP-trees with GPUs left to perform fine grain rejection using guard band clipping, occlusion and backface culling. While such features improve fragment performance, they don't reduce vertex processing overhead.

The leap in performance provided by Cell gives us the bandwidth to significantly reduce RSX time spent processing vertices that don't contribute to the final scene. The favoured approach is to use SPUs to generate minimal scene/instance specific index and vertex buffers from compressed data.

In this talk, we will discuss our approach to face this challenge and how we designed a deferred rendering engine that uses multi-sampled anti-aliasing (MSAA). We will give in-depth description of each individual stage of our real-time rendering pipeline and the main ingredients of our lighting, post-processing and data management. We’ll show how we utilize PS3’s SPUs for fast rendering of a large set of primitives, parallel processing of geometry and computation of indirect lighting. We will also describe our optimizations of the lighting and our parallel split (cascaded) shadow map algorithm for faster and stable MSAA output.

We have all of our animations running on the S.P.U.s of the Cell's chip because you couldn't draw armies or basically animate armies of that amount and size without it. And our physics are completely on there. We are also doing fluid dynamics for the first time in a game, as far as I know. Water is not basically a sheet of a base surface, but completely animated and sub-divided, and you actually can direct with it thanks to the Cell. We actually do part of our rendering on the Cell. Simply because it's so powerful, we spent months and months moving more and more systems onto the S.P.U.'s.

David Kirk: SPE and RSX can work together. SPE can preprocess graphics data in the main memory or postprocess rendering results sent from RSX.

Nishikawa's speculation: for example, when you have to create a lake scene by multi-pass rendering with plural render targets, SPE can render a reflection map while RSX does other things. Since a reflection map requires less precision it's not much of overhead even though you have to load related data in both the main RAM and VRAM. It works like SLI by SPE and RSX.

David Kirk: Post-effects such as motion blur, simulation for depth of field, bloom effect in HDR rendering, can be done by SPE processing RSX-rendered results.

Nishikawa's speculation: RSX renders a scene in the main RAM then SPEs add effects to frames in it. Or, you can synthesize SPE-created frames with an RSX-rendered frame.

David Kirk: Let SPEs do vertex-processing then let RSX render it.

Nishikawa's speculation: You can implement a collision-aware tesselator and dynamic LOD by SPE.

David Kirk: SPE and GPU work together, which allows physics simulation to interact with graphics.

Nishikawa's speculation: For expression of water wavelets, a normal map can be generated by pulse physics simulation with a height map texture. This job is done in SPE and RSX in parallel.

The simple fact is that if it weren't for Cell then RSX would have far more work to do, just like in todays PC's where the GPU is required to do all of the graphics rendering. Cell is also especially suited to vertex work.

That said though the real benefit of PS3's Cell+RSX combo is in the area of physics, as pointed out above.

 

[hernaldo]

The SPUs are heavily involved in the graphics pipeline and do an enormous amount of work to eliminate inefficiency before anything arrives at the PPU and RSX. For example, the SPUs are powerful enough to decompress and check every triangle [polygon] before passing it on to the RSX. Triangles that are facing away from the player, or that are not on the screen can be 'trimmed' away by the SPUs, which hugely reduces the amount of redundant work sent to the RSX. This in turn lets the RSX get on with what it does best--drawing stuff on screen.

Uhhhh isn't that just a Z-Buffer? Which all GPUs since years and years have algorithms on the hardware level to sort out? Doubt there's any saving there getting the CPU to do it.

The PlayStation 3 has a lot of power. When we started Uncharted we were really ambitious and had no idea what the PS3 would give us. Once we got the first devkits, we realized quickly that we could do everything we had planned to. The three main points for me are the Cell, Blu-Ray and the hard drive. We’ve been using the Cell for pretty much all our systems: rendering, particles, physics simulation, collision detection, animation, AI, decompression, water simulation, etc … and to give you an idea of the power of the PS3, we're using only 30 percent of the Cell processor.

Thats just pure fluff, apart from the claim of 30% usage of the cell processor, which is also fluff but more of a lie. Yes they may be using 30% but thats because they're most likely only using 2 of the SPUs for that game....They'd use more if it would mean any benefit from using it or it wasn't too much hassle

I could only be using "25% of the power of my Q6600" to play ANY single threaded game, that doesn't mean by default that my Q6600 is particularly powerful.

 

[hernaldo]

Not that I disagree...interesting reading But definately an attempt by all of them to sell the cell to the potential buyer.

 

[Perfect_Chaos]

Xenos is more powerful just so u know

not a fanboy i have both consoles.

to answer the question though, yes more powerful GPU's are around.. however, why games still run poorly on PC is the big question that are ported from consoles with old technology.

So.. lower res textures on consoles compared to pc such as Mass Effect.

of course, there is also the fact that some games are sloppily ported to PC.. so they are more optimised on the consoles.

 

[drunkenmaster]

lol, none of those quotes actually say the SPU's do anything, at all.

THe main theme there is that SPU's are used to get as precise data as possible so the GPU doesn't draw extra stuff that won't be shown. Thats common practice, thats how games work its still not doing any classical GPU work as would be done on a PC on the GPU.
For years both game engines and graphical drivers have been improving this. Way back when the gpu in general terms got everything that was on screen, drew it all then overlayed anything that was supposed to be seen over anything that wasn't supposed to be seen. Ever since engines are designed to be as efficient as possible in trying to work out exactly what will be on screen at any given time and trying to send only the render commands for those and excluding things that will be blocked at that particular frame. This is an ongoing process of algorithms in the gpu trying to exclude things and game engines trying to send the right data in the first place. All i get from all those quotes is thats what the SPU is doing, which put another way is "the game engine is run on the SPU's" which is hardly unexpected.

AS for animation, its NOT saying its drawing the animation, its calculation where they are, and if theres enough power there sure you can make the animation smoother. In say Lotro you have different levels of animation smoothness. This might be in the graphical options but its the CPU simply adding more "inbetween" frames for the model to give it a smoother appearance. Same thing here with the SPU's. Insomniac as i said, they are very much trying to give the appearance of doing graphical work because at the end of the day it makes the whole platform sound graphically more powerful. But read insomniac's own presentations. They fully explain they don't do graphical rendering on the SPU's, they are experienting with vertex shaders which as I said isn't being done yet, is problematic in many ways at best, and will most likely just be some more efficiency calculations. AT the end of the day if theres spare power, rather than leave the GPU to do the vertex calculations and work out whats shown where, the SPU's are most likely to try to do the work first and exlude anything that wouldn't be shown from being sent so the chances of the GPU drawing stuff that won't be shown is reduced.

"Historically, CPUs have provided course grain scene culling using view frustums, occlusion planes, portal visibility and BSP-trees with GPUs left to perform fine grain rejection using guard band clipping, occlusion and backface culling. While such features improve fragment performance, they don't reduce vertex processing overhead.

The leap in performance provided by Cell gives us the bandwidth to significantly reduce RSX time spent processing vertices that don't contribute to the final scene. The favoured approach is to use SPUs to generate minimal scene/instance specific index and vertex buffers from compressed data."

This says it all, its essentially redoubling efforts to do what the CPU ALREADY does. But even then its to reduce the amount of time the GPu spends doing calculations on things that will be hidden and not shown, it does not, in any way remotely render anything that ends up on the final screen.

While the SPU's might be great for this, CPU's are too. Crysis put my quad under less than 50% load, so there was 50% more cpu power there to do more calculations, the same things the SPU is doing. But theres a limit to how much you can get done to be honest, a basic algorithm using 10% cpu might reduce 90% of the excess gpu work that need not be done. but the next 5% excess might take 30% cpu time. While gpu's might do a better job of the final 10% themselves.

But at the end of the day no one seems to be saying the SPU is fully loaded, because even with all that power theres a limit on how much you have to do.