NGGP: NextGen Garbage Pile (aka: No one reads the topics or stays on topic) *spawn*

Status
Not open for further replies.
I would expect that approach from publishers to backfire seriously.

I expect most games to aim for 1xxx*1080 resolution with vertical scaling, so it will be relatively easy to just decrease the vertical resolution if performance isn't adequate. Say, run the PS4 version at 1600 and the Xbox at 1280 and you have immediately compensated for a 25% performance advantage, with a very little to modest visual difference, especially considering the HW overlay for GUI text on Durango.

Then there's the question of virtual texturing and titles relying on it. If UE and Cryengine both implement it with efficient platform specific code, the DMA hardware can go a long way to compensate Durango's weaker raw performance. Same could be the case for deferred rendering, and so on.

Last but not least, I'd expect the Xbox dev tools to be better once again, which would mean that maintaining stable performance would require less resources compared to doing the same on the PS3.

Some people assume that Orbis uses a derived 7970/7850. If so, Orbis should have 2 DMAs tied to the ACEs for such purposes. Not sure if Sony removed them. A few folks are asking if the rumored 4 compute units are in fact the ACEs.

I expect MS's entire workflow toolset to be a lot more advanced in any case.

EDIT: I also expect MS to gamify their non-games.
 
There will be at least some gameplay difference between the different console versions. The MS TCRs will guarantee that.

Not sure how it's possible to guarantee when Move 2.0 looks to be replicating kinect 2.0 ...

With that being the case, Sony could/would/should pack in Move 2.0 and devs won't have to change a thing WRT gameplay interface.

Unless I missed a patent or two?
 
I would expect that approach from publishers to backfire seriously.

It isn't so much "an approach" as the reality of the situation. The compute, ROP and bandwidth advantages are such that coding to Durango will leave headroom for Orbis.

This "extra" will go to something. Framerate, stability, resolution, IQ etc.
 
Not sure how it's possible to guarantee when Move 2.0 looks to be replicating kinect 2.0 ...

With that being the case, Sony could/would/should pack in Move 2.0 and devs won't have to change a thing WRT gameplay interface.

Unless I missed a patent or two?

I think both will have some kind of 3D motion sensing technology.

EDIT: And Siri-like voice commands.
 

Nevermind - I've just seen ChefO's week-old posts in the Orbis vs. Durango thread where he expected Orbis to produce twice as many frames as Durango. So he's already recovering from whatever caused the problem... :)
 
I expect most games to aim for 1xxx*1080 resolution with vertical scaling, so it will be relatively easy to just decrease the vertical resolution if performance isn't adequate. Say, run the PS4 version at 1600 and the Xbox at 1280 and you have immediately compensated for a 25% performance advantage, with a very little to modest visual difference, especially considering the HW overlay for GUI text on Durango.

Which is, I think, exactly why this custom hardware exists. There are things that are obvious to consumers and things that are not. Will be interesting to see more details about its features if VGleaks spills the beans. Was there also some custom, high powered scaling hardware rumored in Durango as well?
 
...Thought for years (and still now) I think that lot of people think the ps3 is significantly more powerful so whereas it is mostly a wash...

Thing is, it took effort (a lot in some cases) to get that "power advantage" out of ps3. That isn't the case this time. The differences in architecture are minimal, both are easy to dev for, and the differences in "power" are substantial.

It will be readily apparent on day-one (multiplat games) which console is superior.
 
We only have rumors about the bandwidth, and nothing about compute and ROPs.

There must be some balance between ROPs and CUs? Cape Verde has 10 CUs and 16 ROPs, so I think we can assume Durango will have at least 16 ROPs if it has 12 CUs. And it should have 4 TMUs per CU, so 48 total. I think one of the diagrams showed the ROPs, not labelled as rops, but colour and depth backends or something.
 
This is pure speculation about ROPs and the rest and thus no basis to make statements about system performance.
 
It isn't so much "an approach" as the reality of the situation. The compute, ROP and bandwidth advantages are such that coding to Durango will leave headroom for Orbis.

This "extra" will go to something. Framerate, stability, resolution, IQ etc.
Well it looks like indeed the SOny system is to pull ahead, but imo you missing part of the equation. From my pov, both Durango and orbis, looks to have the same kind of "balance".
As laa-yosh points, yes one could render at 1280x1080 the other 1600x1080 but that's it.
If you look at the bandwidth available, the ROPs, etc. (or the difference between those data), nothing point to the massive difference you had when comparing the ps3 and 360 for something like blending.

I don't expect to see the same kind of discrepancies you could have on some game like GTA4 were basically the 360 render at higher resolution with x2 MSAA, or games were transparency are dealt at half (or less) the resolution. The system are more "even", if not in power but in how it is balanced between the functional units. I don't expect the system to perform differently under heavy loads.
SO pretty much it is not a matter of having one system doing better in action scene, or when a lot of transparencies are involved, etc.
They will behave the same, so if Durango render say 33% pixels, it should be the same for all the render targets, the system should suffer in the same the scenes, etc.

The architecture are way more comparable than the ps360, they should mostly have the same strengths and weaknesses and behave the "same".
 
Last edited by a moderator:
Personally I wouldn't discuss performance after at least a year into the next hw cycle, with precise DF data on resolutions, frame rates, torn frames etc.
 
Thing is, it took effort (a lot in some cases) to get that "power advantage" out of ps3. That isn't the case this time. The differences in architecture are minimal, both are easy to dev for, and the differences in "power" are substantial.

It will be readily apparent on day-one (multiplat games) which console is superior.

I don't agree. There is still a lot of fixed function and specific architecture stuff that will make this hard to predict, and will make it very likely that we'll still see different strengths between the two platforms.
 
Well it looks like indeed the SOny system is to pull ahead, but imo you missing part of the equation. From my pov, both Durango and orbis, looks to have the same kind of "balance".
As laa-yosh points, yes one could render at 1280x1080 the other 1600x1080 but that's it.
If you look at the bandwidth available, the ROPs, etc. (or the difference between those data), nothing point to the massive difference you had when comparing the ps3 and 360 for something like blending.

I don't expect to see the same kind of discrepancies you could have on some game like GTA4 were basically the 360 render at higher resolution with x2 MSAA, or games were transparency are dealt at half (or less) the resolution. The system are more "even", if not in power but in how it is balanced between the functional units. I don't expect the system to perform differently under heavy loads.
SO pretty much it is not a matter of having one system doing better in action scene, or when a lot of transparencies are involved, etc.
They will behave the same, so if Durango render say 33% pixels, it should be the same for the render targets, the system should suffer in the same the scene, etc.

The architecture are way more comparable than the ps360, they should mostly have the same strengths and weaknesses and behave the "same".

I agree that for the most part they will "behave the same" in that one will not have an edge in "x" over the other. I'm saying that because they are so similar, the difference in power (bandwidth, compute, and ROPs) and the desire to have pretty screen shots will have most multiplat games running a bit choppy on Durango and/or less resolution and/or lower IQ.
 
I don't agree. There is still a lot of fixed function and specific architecture stuff that will make this hard to predict, and will make it very likely that we'll still see different strengths between the two platforms.

Unless there is some missing secret sauce, I think it will be pretty clear cut and dry.

The only thing the Durango Architecture *might* be able to do is lessen the gap in performance. The thing about Orbis is it's not only more powerful, it's EASY to tap that power and turn it into a real world difference on the screen.

ps360 there was an effort to be made to tap into this power reserve so it was a trade-off. This time, there isn't. In fact, it could be argued that the extra effort will have to be made by Durango devs just to match (or come close to) Orbis. Given the DME's etc.

On Orbis, one nice fat pool of fast ram, same cpu, same gpu, just more gpu resources to play with.

Pretty straight forward to me ...
 
I agree that for the most part they will "behave the same" in that one will not have an edge in "x" over the other. I'm saying that because they are so similar, the difference in power (bandwidth, compute, and ROPs) and the desire to have pretty screen shots will have most multiplat games running a bit choppy on Durango and/or less resolution and/or lower IQ.
Well I just had a loook to that screenshot supposedly from PGR5, look at the DoF effect...
Imho with more of that type of effects involved (be it DoF, motion blur, etc.), how sharp is screenshot become less and less relevant.
I do not agree, the main factor in comparison made between the ps3 and the 360 is frame rate, devs do their best to have that right foremost. Then on the ps3 say you have only every once in a while a scenery with high amount of overdraw, etc. (/ unpredictable), you cut the res of transparent layers already by 2, what do you do? You go with even less for the corner cases scenario?
Thing is neither Durango or Orbis should have free blending like the 360 had for example.

Imo that is close to angst, I think that people could not tell the system apart by watching by them-selves and it is already the case in most mp games this gen. DF and early comparisons between the ps360 did they work but now I think that it became a bit ridiculous as people have an opinion based on those analysis but could not tell by them-selves. I think I would laugh my ass off is DF (or other media as I'm not that ok with their attitude and policies of late) were do blind tests (even including PC with tweaked setting in some cases as a pushover).

Anyway, perceptions are to be set most likely by TFLOPS figures (and MSFT can't do much in this regard), and others things (price, overall offering, etc.). I don't think that how games look for "real" (no magnifier x50, tearing detector that catch stuff in the upper or lower parts of the screen you are highly unlikely to notice in the heat of action (not speaking of tearing happening in overlay). Like in plenty other matter people opinion can and more than often is based on BS.
 
A wild SuperMTW appears!

jTy0RjIxFlMWe.png
 
Not so clear cut. Sony's tools may hamper progress (e.g., the FXAA guy said he doesn't expect low level GPU access in first year). Lead platform, budget, art, schedule may also affect production.

I think both will have some kind of 3D motion sensing technology.

EDIT: And Siri-like voice commands.

Perhaps.

When oh when will we acknowledge that marketing and developers/content will make most of the differences.

PCs and Macs share pretty much the same feature set and even SKU parts (ok, different build quality), but they behave differently and are perceived differently. TCRs aside, games can look the same or wildly different on both platforms just because the devs can.
 
Status
Not open for further replies.
Back
Top