Substance Engine shows PS4 CPU performing better than Xbox One's

Status
Not open for further replies.

DriedMangoes

We The North 🦖🍁
Sep 12, 2013
26,982
9,219
3,930
First heard this from consolewarz.

At first when I saw this news, I kind of dismissed it and said "okay, only 2 extra MB/seconds", nothing much to report here. But after reading up on this some more, could this actually mean the the PS4's CPU could yield better performance than the Xbox One's?

Searching around, it seems the most plausible reasons were either the PS4 is clocked at the same speed as the Xbox One (1.75 ghz) and reserves only 1 instead of 2 cores for the OS resulting in the performance boost or the PS4 is clocked even higher than the Xbox One.

Over at GAF, a fellow named "Matt", apparently a developer/insider who is presently working with both next-gen consoles, gave the comment "Yes, you can get more out of the PS4's CPU than you can the Xbox's."

http://www.neogaf.com/forum/showpost.php?p=94264594&postcount=50

If this is true, this is news to me since I've always thought the Xbox CPU was clocked higher therefore performing better than the PS4's. If this is the case, basically the RAM, CPU and GPU are all in favour of the PS4?


Substance-Engine_Texture-Generation.jpg


http://gamingbolt.com/substance-eng...eneration-speed-to-14-mbs-12-mbs-respectively

In this era of next gen consoles, we’re seeing the emergence of engine developers at a much faster rate than before, compared to the days where Unreal Engine and CryEngine would hog the headlines. Allegorithmic’s Substance Engine which is used to drastically reduce the file size and provides unique visual effects by using dynamic effects in real time, is another example, and supports just about every platform available including the PS4 and Xbox One.
Interestingly while demonstrating texture generation speed, DXT compressed, using Substance Engine, for one CPU, it was found that the PlayStation 4 is able to generate 14MB/s of textures compared to 12MB/s for the Xbox One. Obviously, the Intel Core i7 trounces both of them rather easily at 26 MB/s. Check out the table above.
[/quote
 
Yes there are several things about PS4 not officially confirmed. These are minor issues, though. One CPU being 10-15% faster than the other isn't going to have a major impact on games performance.

1. CPU clock speed. Most people assume vgleaks 1.6ghz but new evidence is contradicting that.
2. CPU reservation. Most people assume 2 cores from killzone shadowfall tech demo.
3. GPU reservation. I'm assuming less than 5% as PS4 has no snap, hdmi input, kinect, app multitasking, etc. to process.
4. Audio hardware. All we know is that it's based on AMD Trueaudio technology, and some comments by Cerny.
5. GPU GCN version. Could be 1.0 or 1.1.
6. RAM latency/memory controller/general bandwidth benchmarks. We don't have official numbers on any of these things.

Almost all the low level info we have about system details is from leaked vgleaks slides that may be outdated. Sony is less than forthcoming about PS4's specs than MS and their hotchips conference, Eurogamer interview, and Penello/Hyrb's specs PR campaign on social media and neogaf.
 
Last edited:
Interesting....I also thought the Xb1 had a slightly higher clock speed.
Wouldn't surprise me though....I heard/read articles just after the PS4 launch saying that the CPU was getting a clock increase in the future.
This would align with that.
 
Also, I think many people, myself included, had always thought the Xbox One's CPU was better than the PS4's. This article shows that this may actually not be the case?
It's possible. I wouldn't be surprised if we get more surprises like PS4's audio hardware being very good, tiny GPU reserve, great scaling on the 6 extra CUs, GCN 1.1, lower latency RAM/memory controller than was assumed, etc. But again those are mostly minor things, we've known the big picture for months.
 
Looks like mobile will catch up to the X1 and PS4 a few years.

PS4/XBO CPU are both basically over-clocked and widened (more cores) mobile CPU's.

So yeah, at least per core, won't be long before mobile is matching the XBO/PS4, but it's unlikely any mobile device is going to have 8 cores or the memory bandwidth of a PS4.

I'd imagine this company is taking advantage of the on-board hardware as much as they can as well. MS's "more memory bandwith" comments aren't a total lie, but what this seems to show is that at least with this companies code (and they are widely used apparently) real world conditions don't have the same performance as MS's "well if you use all of the ESRAM + DDR3 at all the same time you have more bandwidth!" comments.

I'm not some hardcore game developer, but they always seemed suspect. Just logically, you'd have to be using 100% of the ESRAM at all times to get that magical "more bandwidth." Whereas with the PS4, it's all just one memory pool.. so you will get the maximum bandwidth for the assets being used.

We COULD still see this company improve their code, and you don't know where the performance numbers will go in the future, but I imagine if anyone is taking advantage of specific architectures it's a middleware company like that.

There is absolutely nothing to suggest that PS4 has some amazing audio hardware that can even come close to the XBO though consolewarz, you are just kind of pulling things out of thin air with that. Cerny went over the PS4 audio hardware, not sure why he'd neglect to describe some hidden featureset.
 
The first indication PS4 had dedicated audio hardware came from this interview with Cerny. Later it was revealed PS4's audio hardware is based on AMD trueaudio technology, which seems to have quite a few extra features. I guess that's "absolutely nothing pulled out of thin air", though.

It depends on what's actually in PS4's audio hardware and how much of XB1's audio hardware is dedicated to kinect (I'm guessing a decent amount of it). I'm guessing they'll be comparable at offloading audio from the CPU, not that it has much to do with graphics performance anyway.

The "offloading upscale bottleneck custom balance upclock latency" claims are looking increasingly thin.

There are a lot of things Sony hasn't voluntarily revealed about PS4's hardware, for whatever reason. We only know a most low level details due to vgleaks, which might be outdated.
 
Last edited:
lol Core i7 trounces both of them with 1/2 the amount of cores.
 
There is absolutely nothing to suggest that PS4 has some amazing audio hardware that can even come close to the XBO though consolewarz, you are just kind of pulling things out of thin air with that. Cerny went over the PS4 audio hardware, not sure why he'd neglect to describe some hidden featureset.

PS4 outputs in 7.1 channel LPCM and has AMD's latest TrueAudio specification. It even supports 5.1 headphone output out of the controller. Seriously as a gamer what more could you want.
 
Why would the lead designer of PS4's audio hardware say this:

“There’s dedicated audio hardware,” he revealed. “The principal thing that it does is that it compresses and decompresses audio streams, various formats. So some of that is for the games – you’ll have many, many audio streams in MP3 or another format and the hardware will take care of that for you.
Read more at http://www.gamechup.com/mark-cerny-...ed-audio-processing-chip/#YPXMJEv6HHSImg4q.99

If it had some advanced audio processing chip?

That's fine that AMD tweeted that, but Cerny seems oblivious to it's capabilities.

He even mentions in that interview using the GPU for audio processing.

I think that as you go forward we will see a hybrid approach in a couple of years where certain aspects of the audio are being done on GPU.
 
I'm not saying PS4 has more audio hardware than what Cerny indicated, but it seems with the AMD reveal that it's possible (not "pulled out of thin air"), and I wouldn't be surprised if we later learned more.

But again this sort of discussion is mainly a response to straw grasping claims that "Xbox One has tons of offloading and a faster CPU and PS4's CPU is slower and has little offloading therefore PS4 has a CPU bottleneck and/or CUs are being forced to do compute to make up for it."
 
I'm not saying PS4 has more audio hardware than what Cerny indicated, but it seems with the AMD reveal that it's certainly possible, and I wouldn't be surprised if we later learned more.

Fair enough, I certainly have to take back my "pulled out of thin air" comment.

Reading AMD's info.. it's sort of hard to determine whether or not it mimics what SHAPE does anyways. But I also agree that it's hard to determine how much of SHAPE is reserved for Kinect anyways.
 
Not for nothing but isnt this just typical Sony? I mean, we've heard this type of talk EVERY SINGLE Playstation generation. "Most Powerful" GPU or CPU etc, has never decided a clear victor for any of the big 3, ever. Why is it so important now? This gen it seems even the method of HOW we actually get good looking games in called into question.

This has been one strange generation. At least for me mate. Just sayin....:cool:
 
Not for nothing but isnt this just typical Sony? I mean, we've heard this type of talk EVERY SINGLE Playstation generation. "Most Powerful" GPU or CPU etc, has never decided a clear victor for any of the big 3, ever. Why is it so important now? This gen it seems even the method of HOW we actually get good looking games in called into question.

This has been one strange generation. At least for me mate. Just sayin....:cool:

Typical Sony?

That is a benchmark provided by a 3rd party middle-ware developer.

For texture decompression.. a very real world constantly used feature for games.. not some random hidden performance that will never be used.

But we don't know how much of an actual CPU benchmark this is, or if it's down to their codebase differences across PS4/XBO. This definitely isn't "typical Sony" though, and it is more interesting than comparing megahurtz and other details.
 
Typical Sony?

That is a benchmark provided by a 3rd party middle-ware developer.

For texture decompression.. a very real world constantly used feature for games.. not some random hidden performance that will never be used.
Either way. I'm no techie, but its all "potential". Are we divulging all this information because some are expecting the Ps4 to lay waste to the Xbox One in the future or something? How big of a difference will this make in the games? Will we see a considerable jump in graphical fidelity in multi platform and first party games between both consoles? Will there be a more prevalent visual quality (aside from resolutions mind you) in PS4 games? How does this translate into something that we will actually see? Or have we seen it? I genuinely want to know so dont get up in arms about my post.

What can we expect here? Because I am whole-heartedly interested in what this new information means mate. Thats it. Keep it civil.
 
Either way. I'm no techie, but its all "potential". Are we divulging all this information because some are expecting the Ps4 to lay waste to the Xbox One in the future or something? How big of a difference will this make in the games? Will we see a considerable jump in graphical fidelity in multi platform and first party games between both consoles? Will there be a more prevalent visual quality (aside from resolutions mind you) in PS4 games? How does this translate into something that we will actually see? Or have we seen it? I genuinely want to know so dont get up in arms about my post.

What can we expect here? Because I am whole-heartedly interested in what this new information means mate. Thats it. Keep it civil.

Well, noone knows the answers to all your questions because the consoles have only been out for a month and games will continue to improve in the future. But what we do know is that the PS4's CPU may have more performance available for games as evidenced by this test result. How that translates to real games is still up for discussion.
 
Well, noone knows the answers to all your questions because the consoles have only been out for a month and games will continue to improve in the future. But what we do know is that the PS4's CPU may have more performance available for games as evidenced by this test result. How that translates to real games is still up for discussion.

Understandable. However, how prevalent will the differences be, if any? Should we see effects, frame rates and resolutions make a considerable jump or only menial differences? I'm sure you can understand my point. If you are looking for differences you almost certainly will find them in games for all 3 platforms. And judging by how powerful the Ps4 is allegedly, I'm thinking that we should see differences that wont have to be sought after, yet irrefutably speak for themselves. Are these the type of differences we're talking about here?

But again, if there are "enough" smaller differences I'm almost certain that those would stand out---IF there are enough of them. Does this new information translate into the type of "real world evidence" that the Ps4 is indeed as powerful as many believe? This is my question mate...

However, allow me to be frank: if this new information (which go hand-in hand with everything else we know about the PS4) wont bring about irrefutable evidence that the PS4 is clearly as powerful as we've heard--then it all amounts to simply "potential" and a continuation of theories. (e.g. paper specs)
 
Either way. I'm no techie, but its all "potential".

No, it is actually a real engine. No "potential" here.

Are we divulging all this information because some are expecting the Ps4 to lay waste to the Xbox One in the future or something? How big of a difference will this make in the games? Will we see a considerable jump in graphical fidelity in multi platform and first party games between both consoles? Will there be a more prevalent visual quality (aside from resolutions mind you) in PS4 games? How does this translate into something that we will actually see? Or have we seen it? I genuinely want to know so dont get up in arms about my post.

You will probably see a difference that is representative of the actual, literal 30-40% difference in power between the two GPU's alone. If the above substance engine proves true, then the gap is wider since the PS4 also has a more powerful CPU, so together more like a 50% difference in power. Whether that is better resolution or a better frame rate or a combination thereof.

BTW are you sincerely interested in this? You already stated you believe the Xbox One is as or more powerful than the PS4 despite the actual specifications and actual cross platform games.
 
Last edited:
It's a false equivalence to compare Sony in PS2/PS3 specs talk to now. Yes Sony made a bunch of BS claims about PS3, MS did some BS math with the 360 specs as well. With PS4 so far everything official has panned out, and we haven't heard anything outlandish. If anything the roles are reversed.

The chart was released by a middleware developer to sell their product, that's why "it's being divulged". A CPU difference of 10-15% either way won't make much if any difference to graphics performance unless the game is very CPU heavy. Threads like these are mainly a response to mostly nonsense "offloading upscale bottleneck custom balance upclock latency" claims.

I expect to most major multiplatform games to run at 720-900 or 900-1080. There might be slight differences in framerate or IQ, but mainly resolution as it's the easiest thing for devs to change. In exclusive games the difference will be greater as first party devs utilize GPGPU and unified memory features more effectively.

Whether someone can notice the performance difference, whether they think it's a large difference, or whether they care is all subjective. The specs are what they are and the hardware is being proven by dev statements, benchmarks, and DF analysis.
 
Last edited:
Typical Sony?

That is a benchmark provided by a 3rd party middle-ware developer.

For texture decompression.. a very real world constantly used feature for games.. not some random hidden performance that will never be used.

But we don't know how much of an actual CPU benchmark this is, or if it's down to their codebase differences across PS4/XBO. This definitely isn't "typical Sony" though, and it is more interesting than comparing megahurtz and other details.

According to the tech geeks on B3D the numbers fit nicely in a 2GHz clock for the PS4 CPU and a 1.75GHz clock for the Xbone. Sony has not announced their final CPU clock rate and probably never will as I don't see why they would need to given there is already a perceived power difference, but most of the credible rumors point to 1.8GHz-2.0GHz.
 
Understandable. However, how prevalent will the differences be, if any? Should we see effects, frame rates and resolutions make a considerable jump or only menial differences? I'm sure you can understand my point. If you are looking for differences you almost certainly will find them in games for all 3 platforms. And judging by how powerful the Ps4 is allegedly, I'm thinking that we should see differences that wont have to be sought after, yet irrefutably speak for themselves. Are these the type of differences we're talking about here?

But again, if there are "enough" smaller differences I'm almost certain that those would stand out---IF there are enough of them. Does this new information translate into the type of "real world evidence" that the Ps4 is indeed as powerful as many believe? This is my question mate...

However, allow me to be frank: if this new information (which go hand-in hand with everything else we know about the PS4) wont bring about irrefutable evidence that the PS4 is clearly as powerful as we've heard--then it all amounts to simply "potential" and a continuation of theories. (e.g. paper specs)

Some real world, irrefutable evidence is in some of the multiplatform games already:

COD:G - 720p on Xbox One, 1080p on PS4
AC4 - 900p on Xbox One, 1080p on PS4
BF4 - 720p on Xbox One, 900p on PS4

The PS4 has the graphical advantage through resolution, higher FPS average and better AA in all cases. If the developers could make both versions equal, they would. Whether you discern the difference or not depends on a number of variables such as your perception and how far you sit from your TV/monitor etc. but factually speaking, the differences are there already.

Whether the gap will narrow or widen, we won't know until the future titles actually come out.
 
Sony never said they had the most powerful CPU. Just the tests from substance engine says that they do.

None of this is really important as evidenced by the PS2 and Wii which were #1 in sales yet were not as powerful as their competition. But if you want the best version of the games the choice is pretty easy this generation.


How? based on "potential"? Based on Launch games? Dont you think thats pretty premature at this point? Sure. If we were 2-3 yrs into this generation I'd be inclined to agree with you. But....that isnt the case. The consoles have only been out a month. Most agree the the XO had the best launch line up of both consoles (this is NOT a bash on the PS4 so dont take it there mate.....) RYSE (yes I'm using this game as an example) is almost universally praised as the best looking launch game on both consoles (and this is on the alleged weaker system).

But, if games will eventually look better on the PS4 in the future, that sure is easier to believe than stating that they do right now. You know?
 
Whether you discern the difference or not depends on a number of variables such as your perception and how far you sit from your TV/monitor etc. but factually speaking, the differences are there already.
True, perception is subjective. The problem is, people hear things like "56% more teraflops" and think it should translate to the difference between a PS2 and a GTX Titan. No, it's not nearly that much. It's about enough to go from 720 to 900, or 900 to 1080, with maybe a bit of extra framerate and IQ thrown in, and extra compute physics if the devs take advantage of GPGPU. That's all. And it's close to what we're seeing in multiplatform games right now.

Based on the known hardware specs, if Ryse were ported to PS4 devs could up the resolution to 1080p with no penalty at a minimum. Probably improve the framerate and add some GPGPU stuff as well. Regardless if you think "Ryse looks the best", it would look better on PS4 if devs took advantage of the extra hardware.
 
Last edited:
Resolution, frame rate and color bit depth all scale linearly with processing power, and they always will. I say this as someone who works on and develops radiology imaging software. Image quality is everything in my field and you can't get something for nothing.

The reason I bring this up is I have seen arguments on this board confusing a 40% difference in power with a 40% qualitative visual difference, for instance, which is the wrong way to think about image quality. As an example it takes approximately 56% more power (including processing, bandwidth and random access storage) to run something at 900P than it does at 720P, if frame rate and color bit depth are equal. If it doesn't look "56% better" to you that doesn't mean that it isn't actually requiring 56% more power to do it.
 
How? based on "potential"? Based on Launch games? Dont you think thats pretty premature at this point? Sure. If we were 2-3 yrs into this generation I'd be inclined to agree with you. But....that isnt the case. The consoles have only been out a month. Most agree the the XO had the best launch line up of both consoles (this is NOT a bash on the PS4 so dont take it there mate.....) RYSE (yes I'm using this game as an example) is almost universally praised as the best looking launch game on both consoles (and this is on the alleged weaker system).

But, if games will eventually look better on the PS4 in the future, that sure is easier to believe than stating that they do right now. You know?

The games already are technically superior on PS4. There isn't a single example of a cross platform game that is technically better on XO, and as long as the specs stay the same, which they will, this should remain the case. These boxes aren't exotic they are a Radeon 7870 and a 7790 with nearly identical CPU's.
 
Status
Not open for further replies.