SUBSTANCE ENGINE

[Shortbread]

http://www.allegorithmic.com/technology/substance-engine

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

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.

i7 laying the smack down... lol

Anyhow, it seems the PS4 Jaguar CPU has slightly more grunt than its XB1 counterpart.

 

[Shifty Geezer]

i7 laying the smack down... lol

Considering the i7 is clocked at least twice as fast probably, it's not that bad. Also, is one CPU all the cores of that CPU? Or 1 core?

This would point to either PS4 being clocked a little higher, or having less reservation than XB1.

 

[Bigus Dickus]

Considering the i7 is clocked at least twice as fast probably, it's not that bad. Also, is one CPU all the cores of that CPU? Or 1 core?

This would point to either PS4 being clocked a little higher, or having less reservation than XB1.

I would think this reflects the current reservation differences in xb1 and ps4. Bandwidth may play a role but I think less likely.

 

[Shifty Geezer]

I would have thought so to, although it points to a substantial difference that doesn't tally with other info. We have heard two reserved cores on PS4, and two on XB1. Actually, perhaps this is the best evidence that PS4 is only one reserved core? That'd tie in perfectly with the above numbers. 12 MB/s from 6 cores on XB1, and 14 MB/s from 7 cores on PS4, at 2 MB/s from each Jaguar core. At which point, they'd have to be clocked the same.

Current theory had XB1's CPU clocked faster though, 1.75 GHz vs. 1.6. Well, the figures could be well rounded, but my calculator can't get 14 at 7*1.6 GHz and 12 from 6*1.75 GHz.

 

[picosec]

The slide is labeled ".. - 1 CPU", which I would assume means 1 CPU core. The difference could just be down to how well the PS4 and XONE compilers optimize their code.

 

[Davros]

provides unique visual effects by using dynamic effects in real time

what engine doesnt ?

 

[Shifty Geezer]

The slide is labeled ".. - 1 CPU", which I would assume means 1 CPU core. The difference could just be down to how well the PS4 and XONE compilers optimize their code.

That's the after interpretation, which would place Jaguar very well against i7, although it's a poor of phrase. With 2x the number of cores it'll have comparable performance at half the clock rate. Factoring in small size, Jaguar would the 'down-layerer-of-the-smack'. Is there any reason for MS and Sony compilers to be that different seeing as it's x86? I assume a lot of the compiler will come from AMD.

 

[Shifty Geezer]

what engine doesnt ?

They mean texture effects. The substance demos are very impressive. The shaders support parameters like 'age' and you can easily adjust age of a material and have it computed in realtime, adding decay.

From 1:10

He mentions scalability. 3x improvement across 4 cores versus 1 core.

 

[Shortbread]

I would have thought so to, although it points to a substantial difference that doesn't tally with other info. We have heard two reserved cores on PS4, and two on XB1. Actually, perhaps this is the best evidence that PS4 is only one reserved core? That'd tie in perfectly with the above numbers. 12 MB/s from 6 cores on XB1, and 14 MB/s from 7 cores on PS4, at 2 MB/s from each Jaguar core. At which point, they'd have to be clocked the same.

Current theory had XB1's CPU clocked faster though, 1.75 GHz vs. 1.6. Well, the figures could be well rounded, but my calculator can't get 14 at 7*1.6 GHz and 12 from 6*1.75 GHz.

Shifty, I think you may have figured out the PS4/XB1 CPU core reservations... :smile:

 

[pjbliverpool]

The slide is labeled ".. - 1 CPU", which I would assume means 1 CPU core. The difference could just be down to how well the PS4 and XONE compilers optimize their code.

That would put the IPC of an i7 at the same level as a Jaguar core which we know from other benchmarks (and common sense) is not the case. If the benchmark really is just running on one CPU core then its a useless comparison across architectures as its completely misrepresentitive of actual performance.

 

[onQ]

I would think this reflects the current reservation differences in xb1 and ps4. Bandwidth may play a role but I think less likely.

Shouldn't Xbox One have the Bandwidth advantage on the CPU end?

 

[Globalisateur]

Shouldn't Xbox One have the Bandwidth advantage on the CPU end?

That's what the Microsoft bandwidth numbers told us, right?

 

[Graham]

Considering the i7 is clocked at least twice as fast probably, it's not that bad. Also, is one CPU all the cores of that CPU? Or 1 core?

This would point to either PS4 being clocked a little higher, or having less reservation than XB1.

I would have assumed it meant single core - are dual i7 motherboards even available? Otherwise why mention it?

 

[AlNom]

Current theory had XB1's CPU clocked faster though, 1.75 GHz vs. 1.6. Well, the figures could be well rounded, but my calculator can't get 14 at 7*1.6 GHz and 12 from 6*1.75 GHz.

Clearly it's 14/12 ~= 2.0/1.75

Clearly.

 

[DieH@rd]

Does bandwidth speeds to RAM pools have effect on this CPU benchmark? Xbone is limited by that. Also does this benchmark needs a lot of caching? 32MB of ESRAM is also possible limitation.

 

[DuckThor Evil]

Clearly it's 14/12 ~= 2.0/1.75

Clearly.

I noticed that too, but wouldn't Sony have mentioned something like that?

 

[Shortbread]

I noticed that too, but wouldn't Sony have mentioned something like that?

Why would they?

It's not in Sony's best interest to release every detail of the PS4 architecture... plus, Sony probably just doesn't give a damn about going tit-for-tat with MS. The PS4 hardware "as of now" is proving to be more robust with most 3rd party games.

 

[Shifty Geezer]

That would put the IPC of an i7 at the same level as a Jaguar core which we know from other benchmarks (and common sense) is not the case. If the benchmark really is just running on one CPU core then its a useless comparison across architectures as its completely misrepresentitive of actual performance.

Not useless. The chart is intended for devs using Substance. It gives an indication of how much Substance they can use on each of 5 common platforms - 3 mobile devices, 2 consoles, and a decent PC. It wasn't intended as a performance benchmark for processors.

Shouldn't Xbox One have the Bandwidth advantage on the CPU end?

That shouldn't come to much. It's an algorithmic process, so very little read, and the output is MB/s, so very little write. It should be a pure computation test with no benefit from BW.

 

[KKRT]

Love the tags in this thread

 

[pjbliverpool]

Not useless. The chart is intended for devs using Substance. It gives an indication of how much Substance they can use on each of 5 common platforms - 3 mobile devices, 2 consoles, and a decent PC. It wasn't intended as a performance benchmark for processors.

Fair enough. But you would have to question what's going wrong with their engine if a 3.4Ghz+ i7 core is less than twice as fast as a 1.6Ghz Jaguar core in this operation. I suppose memory bandwidth constraints could come into play.