A primer on the X360 shader ALU's as I understand it
- Thread starter superguy
- Start date
Lead SKU's are based solely on projected sales (or revenue is some cases). It's rarely if ever a technical decision.
A lot of people still think it's Sony's race to loose.
Having said that I know of several multiplatform games that have PS3 as lead SKU and the majority of development is done elsewhere because the devkits are so damn noisy :O
A lot of people still think it's Sony's race to loose.
Having said that I know of several multiplatform games that have PS3 as lead SKU and the majority of development is done elsewhere because the devkits are so damn noisy :O
So can we say the bottom line is that the X360 is better off with Xenos than say if MS had included a X1900XTX or 7900GTX instead (from a graphics capability standpoint). And this would be because of its superior technical features and edram?
Last edited by a moderator:
For developers more or less in control of their choice of platform... do they tend to choose PS3 as the lead SKU. I ask because it seems that the PS3 friendly ones are the ones most prone to discuss the technical benefits of each chip (xenos v RSX)... Of course i just have the impression that nAo, ERP and Faf are PS3 devs with only Mint master and Mr. Wibble maybe being x360 devs... so I may be wrong.
I cant tell if nAo was defending xenos' capabilities... or just enjoyed telling guilty bystander the he was wrong.
Do most PS3 devs also have MS kits????
I cant tell if nAo was defending xenos' capabilities... or just enjoyed telling guilty bystander the he was wrong.
Do most PS3 devs also have MS kits????
Mr Wibble is a PS3 dev also, and I don't believe Mintmaster is a game dev at all (correct me if I'm wrong Mint!).
But the reason why you'll see nAo or Deano or someone else 'defend' Xenos isn't because they want to paint someone else in a bad light or because they have 360 projects on the side; it's simply because technology is technology, and if someone is saying something untrue or fallacious about a certain part - and one knows better - how can one not speak up?
But the reason why you'll see nAo or Deano or someone else 'defend' Xenos isn't because they want to paint someone else in a bad light or because they have 360 projects on the side; it's simply because technology is technology, and if someone is saying something untrue or fallacious about a certain part - and one knows better - how can one not speak up?
Guilty Bystander
Newcomer
nAo said:It really does not matter if it can change its configuration on the fly every 4 cycles, you should have all the granularity you need.
What?! Last time I checked it only can't handle blending in that space.
WHAT?!
ERR!?!?!?
Are you kidding!?
meaningless number
meaningless number
this is the only correct thing you quoted..but you know, it's statistical..it is bound to happen
LOL, maybe you want to reconsider your statement about edram here..
other meaningless number
same old story..
Technically the Xenos can do 16bit FP blend HDR but because of the high performance penalty it doesn't.
On the eDRAM it only does 10bit FP HDR otherwise it wouldn't fit into the eDRAM with Z-testing, Z-buffering and FSAA.
All those numbers are VERY MEANINGFULL as they're all theoretical numbers and just shown that in terms of RAW power the Xenos is only on par with the X850XT PE in Shader ops (48B Xenos vs 44B X850XT PE) and Texel fillrate (8GTexel/s Xenos vs 8,64GTexel/s X850XT PE) while being worse at Pixel fillrate (4GPixel/s Xenos vs 8,64GPixel/s X850XT PE), has less bandwidth (22,4GB/s Xenos shared with Xenon vs 37,76GB/s X850XT PE) and does less Vertices (500Mverts/s Xenos vs 810Mverts/s X850XT PE).
So the only thing the Xenos puts to the table is that it can do more complex Shader which it can't push a lot of since it can only do 48B Shader ops and flexibilty.
Other then that I don't think the Xenos is that great and wouldn't be suprised if the RSX walks right over it and that with the Cell walking over the Xenon is all the PS3 needs to seriously kick Xbox 360's ***.
The Xbox 1,5 stories might be close to the truth after all.
On a sidenot all the numbers I state are correct and I'm not blinded by any bias or whatever as I don't have one but the fact just is that the Xenos doesn't have any RAW power compared to current high-end desktop GPU's.
The fact the Xbox 360 does games like Oblivion, Condemned and Call of Duty 2 better than most PC's is just because it's a closedbox architecture.
Microsoft would have been smarter by putting in a X1800/1900 @500-600MHz in the Xbox 360 with the eDRAM.
blakjedi said:
Well, some of them may indeed have 360 documentation and/or dev kits - who knows? But even beyond that, you have to understand that these devs talk to one another. And even beyond that, when you have a strong grasp of the technicalities of these systems, you could possibly derive more from even the public info than can the average individual. I mean in truth, that's why a lot of people start coming to B3D; because there *is* a strong technical community centered around a core of industry folk.
Keep in mind also, that in this thread a lot of the 'correction' I've seen has come more from the side of RSX capabilities being adjusted and/or SPE capabilities being defined. Since they *are* PS3 devs, this makes sense. Not to expound on Xenos per se, but to correct any fallacies they feel might be present in the comparissons between RSX and Xenos.
Remember also that we (the public) know a lot more about Xenos than we do RSX. Even without having a 360 dev kit or anything like that, it's much easier to hypothesize as to what Xenos is capable of achieving, because we are able to provide and discuss hard references.
Last edited by a moderator:
I'm confused as far as the RSX technical specs are concerned...I don't understand why they haven't been elaborated on more by Sony. Is it they don't want the competition (MS/Nintendo or ATI) to utilize the technology or know the RSX's weaknesses? NVidia, during conference calls with its shareholders, has said the final silicon was handed over to Sony for production. I assume that means RSX is done as it ever will be.
Somehow I have a feeling NVidia may want the technical details kept quiet so that its main competitor ATI can't leverage it in the PC market where traditional shader pipelines are still evolving. Does that sound feasible?
Somehow I have a feeling NVidia may want the technical details kept quiet so that its main competitor ATI can't leverage it in the PC market where traditional shader pipelines are still evolving. Does that sound feasible?
No it can't. Xenos does NOT have any fp16 blender, that's the same reason you can have MSAA on fp16 render targets on Xenos but you have to resolve your multisampled buffer to a non multisampled one on your own cause Xenos does not have any custom hw to do it (you can use a shader instead).Guilty Bystander said:
You forgot about tiling.
They're not and the reason is very simple: you're quoting some theoretical number but IF you could quote ALL the theoretical numbers you would quickly discover that those numbers are completely unsustainable, hence they're useless.
I won't go into every specific number you quoted but let just say that in any discipline you can't analyze a complex system by using one or two quantities.
a GPU can be bottleneck in some many ways that you really have no idea about..trust me
You're very far from the truth, whether truth is something real or not.
Even if you numbers can be strictly correct per se (but I haven't checked them all) theu're far from being sufficient in order to compare two or more GPUs.
If it was that easy we could write a stupid .net application in 30 minutes, plug your GPUs numbers, and in one second you would know which is the 'most' powerful, lol
I'm sure I can design so many contrived cases where RSX can spank Xenos and vice versa.
The underlying truth to everything here though is that you aren't going to visually see a difference cuz neither is going to be substantially faster than the other. Both companies went all out designing these things and both have probably done a superb job (PS3 is still not proven though).
It will a dumb question, can someone explian me what is SOA and AOS?Fafalada said:They are only really a big deal when you're not really computation bound though - most scalar quantizations are trivially short even without specialized instructions. I'd rather have something done on the issue of GPU understanding SoA, or free conversion between AoS and SoA (there's one console CPU that does this but it's not in a desktop machine).
Well nAo listed a lot of it - basically being able to access entire topology of your primitives opens up a whole new world that VS can't reach, or can only reach with excessive performance and memory overhead.
Deletion of primitives is an interesting area for optimizations as well - from avoiding entire batches to be sent over to GPU, to culling stuff on primitive level that only wastes GPU vertex setup (backfacing, degenerate etc. primitives)
And that's still just a few of the interesting things.
Anyway I see your point about lead platform in multiplat. titles, but we don't really know what the case is for MoH, maybe ERP could let us in on that secret
What nAo said about pixel work - and besides, what if the lead platform was PS3?
Could the VMX unit in xenon have enought muscle to do something close.
Does memexport can help in someways for some of fonctionnalities NAo and fafalada speak about.
