DF: Xbone Specs/Tech Analysis: GPU 33% less powerful than PS4

The PS4 has 50% more Compute Units, this equates to 33% more processing power, that along with much greater memory bandwidth from one, unified pool means that we will see some major differences .. especially from First Party devs.


The margin between 360 and PS3 was far less than the difference between PS4 and Xbox One ......... there will be differences.
 
Shit yes you are right, comparing the FLOPS of XBO it is 50%, I was working backwards from PS4's number.. then there's the rest of the hardware.

What do we know about the CPU and Audio processing for both machines? Are they the same?

The CPU is identical but more cores are locked down for OS usage for Xbox One. The RAM in PS4 is a generation ahead of the Xbox One. The Xbox One is reserving 3 GB for the OS, and the PS4 is rumored to be using 2 or less (some reports say 512 mb some say 1 GB).

tldr;
Xbox One has ZERO advantage in terms of RAW numbers or specs. If a game runs at 1920X1080p, 30 FPS, Ultra on PS4; the Xbox One version will run at 1600X900p, 30 FPS, High.
 
GAF isn't very good at math unfortunately

Here's my attempt.

PS4 = 1.8 TF
XBO = 1.2 TF

Difference = 1.8 - 1.2 = 0.6 TF difference .. So by what percentage is PS4 (GPU) better?

0.6/1.8 * 100 = 33% MOAR FLOPS

If you compare both GPU's directly you get 1.2:1.8 = 2:3 POWER RATIO

TLDR; PS4 has a GPU that can output 33% more than XBO .. Is that a huge difference? Only the games will tell us. But there are many other factors like CU, RAM etc

....

1.8/1.2 = 1.5

So 1.2 goes into 1.8, 1.5 times, thus 1.8 is 50% increase over 1.2.

1.2/1.8 = 0.66*

Thus 1.2 consitutes 66*% of 1.8, thus 1.2 is 33% less than 1.8.
 
Funny how the most important takeaway from all those PS4 / Xbox speculation and then reveal threads is that US education system really needs improvement :(
 
GAF isn't very good at math unfortunately

Here's my attempt.

PS4 = 1.8 TF
XBO = 1.2 TF

Difference = 1.8 - 1.2 = 0.6 TF difference .. So by what percentage is PS4 (GPU) better?

0.6/1.8 * 100 = 33% MOAR FLOPS

If you compare both GPU's directly you get 1.2:1.8 = 2:3 POWER RATIO

TLDR; PS4 has a GPU that can output 33% more than XBO .. Is that a huge difference? Only the games will tell us. But there are many other factors like CU, RAM etc

This is so sad.

Edit: Mods were right about putting forth a few simple maths test as a part of the registration process.
 
Holy crap is this debate still going?

In the simplest of terms..

PS4 is 50% more powerful than the XO.
XO is 33% less powerful than the PS4.

Traditionally people don't go with the less powerful descriptor so I have no idea why they are now.

PS4 is 50% more powerful. There. Done.



EDIT:

To add further layman's explanation. The power difference between the two consoles is roughly 600 Gflops. Eg 1.8 Tflops is 600 Gflops more than 1.2 Tflops.

- 600 Glops is 50% of the XO's total power (1.2 Tflops)
- 600 Glops is only 33% of the PS3's total power (1.8 Tflops).

For the purposes of sanity, let's just stick to the PS4 is 50% more powerful. No idea when "is less powerful" became fashionable. You'd only really use the latter if you're trying to downplay the power difference of the weaker device.

Apologies if my post comes off as condescending.
 
You know how there are sometimes experiences that happen in your life that restore your faith in humanity?

This isn't one of them.
 
GAF isn't very good at math unfortunately

Here's my attempt.

PS4 = 1.8 TF
XBO = 1.2 TF

Difference = 1.8 - 1.2 = 0.6 TF difference .. So by what percentage is PS4 (GPU) better?

0.6/1.8 * 100 = 33% MOAR FLOPS

If you compare both GPU's directly you get 1.2:1.8 = 2:3 POWER RATIO

TLDR; PS4 has a GPU that can output 33% more than XBO .. Is that a huge difference? Only the games will tell us. But there are many other factors like CU, RAM etc


Oh for god's sake, this is getting painful.
 
GAF isn't very good at math unfortunately

Here's my attempt.

PS4 = 1.8 TF
XBO = 1.2 TF

Difference = 1.8 - 1.2 = 0.6 TF difference .. So by what percentage is PS4 (GPU) better?

0.6/1.8 * 100 = 33% MOAR FLOPS

If you compare both GPU's directly you get 1.2:1.8 = 2:3 POWER RATIO

TLDR; PS4 has a GPU that can output 33% more than XBO .. Is that a huge difference? Only the games will tell us. But there are many other factors like CU, RAM etc

Maths hert brainz.
 
The CPU is identical but more cores are locked down for OS usage for Xbox One. The RAM in PS4 is a generation ahead of the Xbox One. The Xbox One is reserving 3 GB for the OS, and the PS4 is rumored to be using 2 or less (some reports say 512 mb some say 1 GB).

tldr;
Xbox One has ZERO advantage in terms of RAW numbers or specs. If a game runs at 1920X1080p, 30 FPS, Ultra on PS4; the Xbox One version will run at 1600X900p, 30 FPS, High.

How recent is your info about the core reserve on PS4? I had heard 1 core a while back but the tech slides for the Shadow Fall demo said that they were only using six cores, although that may have been an early hardware/time issue. I hope the OS reserve on PS4 is no more than 1 core/1GB of RAM. If they can't beat MS in the OS game, at least don't make the OS bloated and let the games use the power.
 
How recent is your info about the core reserve on PS4? I had heard 1 core a while back but the tech slides for the Shadow Fall demo said that they were only using six cores, although that may have been an early hardware/time issue. I hope the OS reserve on PS4 is no more than 1 core/1GB of RAM. If they can't beat MS in the OS game, at least don't make the OS bloated and let the games use the power.
I have a feeling the RAM will be scalable. Background operations while ingame (e.g. chats, friends list, recording etc) will take 1GB but once you go out of the game, the OS will be much more robust. I think they use this sort of system on Vita. The speed of the RAM should hopefully make this transition as seamless as possible
 
I have a feeling the RAM will be scalable. Background operations while ingame (e.g. chats, friends list, recording etc) will take 1GB but once you go out of the game, the OS will be much more robust. I think they use this sort of system on Vita. The speed of the RAM should hopefully make this transition as seamless as possible

This is what I'm expecting too, which is why I think there's a chance that the old 512mb number may still hold up. I'm not worried about the RAM all that much though, more about the CPU since that's far less capable than the GDDR5 and needs all the resources it can get.
 
How recent is your info about the core reserve on PS4? I had heard 1 core a while back but the tech slides for the Shadow Fall demo said that they were only using six cores, although that may have been an early hardware/time issue. I hope the OS reserve on PS4 is no more than 1 core/1GB of RAM. If they can't beat MS in the OS game, at least don't make the OS bloated and let the games use the power.

For the ps3 they disabled one core for higher yields and set aside one core for security purposes. This might be how they are proceeding here as well.
 
For the ps3 they disabled one core for higher yields and set aside one core for security purposes. This might be how they are proceeding here as well.

I don't think they'll really need to worry too much about yields with a Jaguar CPU. MS isn't disabling any of the cores for the CPU, both of them are disabling GPU CUs to aid with yields.
 
Holy crap is this debate still going?

In the simplest of terms..

PS4 is 50% more powerful than the XO.
XO is 33% less powerful than the PS4.

Traditionally people don't go with the less powerful descriptor so I have no idea why they are now.

PS4 is 50% more powerful. There. Done.



EDIT:

To add further layman's explanation. The power difference between the two consoles is roughly 600 Gflops. Eg 1.8 Tflops is 600 Gflops more than 1.2 Tflops.

- 600 Glops is 50% of the XO's total power (1.2 Tflops)
- 600 Glops is only 33% of the PS3's total power (1.8 Tflops).

For the purposes of sanity, let's just stick to the PS4 is 50% more powerful. No idea when "is less powerful" became fashionable. You'd only really use the latter if you're trying to downplay the power difference of the weaker device.

Apologies if my post comes off as condescending.
word
 
The CPU is identical but more cores are locked down for OS usage for Xbox One. The RAM in PS4 is a generation ahead of the Xbox One. The Xbox One is reserving 3 GB for the OS, and the PS4 is rumored to be using 2 or less (some reports say 512 mb some say 1 GB).

tldr;
Xbox One has ZERO advantage in terms of RAW numbers or specs. If a game runs at 1920X1080p, 30 FPS, Ultra on PS4; the Xbox One version will run at 1600X900p, 30 FPS, High.

5 Billion transistors!
 
How recent is your info about the core reserve on PS4? I had heard 1 core a while back but the tech slides for the Shadow Fall demo said that they were only using six cores, although that may have been an early hardware/time issue. I hope the OS reserve on PS4 is no more than 1 core/1GB of RAM. If they can't beat MS in the OS game, at least don't make the OS bloated and let the games use the power.

GopherD posted this a while back and you don't dispute the Gopher.

http://www.neogaf.com/forum/showpost.php?p=47871129&postcount=1620

GoperD said:
codeblew said:
Well, each core only has 1 hardware thread so maybe Durango needs 2 hardware threads for whatever. My point is that Orbis will most likely use at least 1 as well. Possibly 2.
Nope all evidence is pointing to less than 1


For the ps3 they disabled one core for higher yields and set aside one core for security purposes. This might be how they are proceeding here as well.

I think a bunch of that is going to be taken care of by the ARM cores this time around.
 
The CPU is identical but more cores are locked down for OS usage for Xbox One. The RAM in PS4 is a generation ahead of the Xbox One. The Xbox One is reserving 3 GB for the OS, and the PS4 is rumored to be using 2 or less (some reports say 512 mb some say 1 GB).

tldr;
Xbox One has ZERO advantage in terms of RAW numbers or specs. If a game runs at 1920X1080p, 30 FPS, Ultra on PS4; the Xbox One version will run at 1600X900p, 30 FPS, High.

Kinect camera is 1080p vs. 720p of the Sony camera :)
 
The reason why the PS3 and 360 have marginal graphics different between games from third party is because the Cpu on the PS3 is superior than the 360 but the Gpu of the 360 is better so it kinda cancel each other out. The situation with the ps4 n Xbox one is that the cpu are the same but the ps4 gpu has 50% more power and the 8gb gddr5 ram. Not only will ps4 first party game will be a huge different than the first party games from xbox one but third party games will run better on the ps4.
 
so with the news that the PS4 is going to be more powerful than the Xbone are we going to see higher frame rates and better visuals on the PS4 or will Microsoft somehow convince third parties to push for parity?
 
The PS4 has 50% more Compute Units, this equates to 33% more processing power...

Haha, what? No, that's not where the 33% number comes from. It comes from extremely basic and simple math, but unfortunately people here seem to have skipped out on that particular subject all throughout their many years of education.

Ok, here...

This is the XBO's GPU computing power: øø (let's call these "power units" - the XBO has two of them)

This is the PS4's GPU computing power: øøø (three "power units")

Ok, so the PS4 has one more ø of GPU power than the XBO. How much more power is that, given as a percentage of the XBO's power? Well, ø (the difference) is half of øø (the XBO's total GPU power), and half of something equals 50%. So the PS4 has 50% more GPU power than the XBO. Or: the PS4 has 150% of the XBO's GPU power.

Now, let's look at it from the other point of view. The XBO has one less ø of GPU power than the PS4. How much less power is that, this time given as a percentage of the PS4's power? ø is a third of øøø (the PS4's total GPU power), and a third of something is about 33%. So the XBO has 33% less GPU power than the PS4. Or: the XBO has 67% of the PS4's GPU power.

You arrive at 50% more power or 33% less power depending on what you're using as the base value to compare against. You could never arrive at 33% more power or 50% less power in this case.
 
What is the breakdown of the XO transistor budget, i.e., how many for for CPU, GPU, Kinect, etc? How many transistors does the PS4 have?

Looong explanation.
Someone needs to bring in the hand puppets, this is getting abso-fuckin-lutely ridiculous.
 
Who cares about the maths. At this point it's distracting us from the real news of PS4 being a considerable chunk more powerful than Xbone. If such a hoo ha was made out of multiplatform differences this gen on two machines that were essentially in the same power ballpark, imagine the incoming shitstorm next-gen. Will be glorious.
 
If gflops was an accurate description of performance we wouldn't need to do game benchmarks
But PS4 and XO share the same basic GPU architecture. In which case, yeah you can do those comparisons. You can scale the numbers almost linearly.
 
But PS4 and XO share the same basic GPU architecture. In which case, yeah you can do those comparisons. You can scale the numbers almost linearly.
Well, GPU's aren't only made out of ALU's. But the PS4 has even more of an advantage in other GPU components, such as ROPS count or bandwidth...
 
What is the breakdown of the XO transistor budget, i.e., how many for for CPU, GPU, Kinect, etc? How many transistors does the PS4 have?

Console launches are both great and horrible.. It's only this time when you'll have people comparing Flops and how many transistors each console has.

Hint, transistors don't mean a damn thing. What does matter is the GPU, the CPU, the RAM, the bottlenecks between each of the above, and the tools that developers have to program the above.

maxcer0081 said:
so with the news that the PS4 is going to be more powerful than the Xbone are we going to see higher frame rates and better visuals on the PS4 or will Microsoft somehow convince third parties to push for parity?

Thankfully, console makers don't have that kind of authority over 3rd parties. I remember these same questions being asked last time around of Sony, it doesn't happen, folks. Developers are going to do the most that they can with each console. They'll have a baseline for next gen, but whichever one is more powerful (combined with being easy to program for) is likely going to have the better version.
 
So if Jody has four apples, and Billy has six apples, by what percentage does Billy have more apples than Jody?

Show you working out for full marks.
 
Well, GPU's aren't only made out of ALU's. But the PS4 has even more of an advantage in other GPU components, such as ROPS count or bandwidth...
Right, was only referring to floating-point calcs. So the PS4's GPU is at least 50% faster, more in some cases. And that doesn't necessarily mean--all else being equal--that it's always gonna run any application 50% faster. But that's not because the GPU isn't actually 50% faster -- it's because of the inefficiencies of the program itself.
 
As it seems a high percentage of NeoGAF has trouble with percentages, I suggest just using ratios which are easier for people to grasp.

The PS4 is 1.5X the GPU FLOPS of the XBone GPU. (1.8/1.2 = 1.5X)

Just use the 1.5X number everywhere and it becomes easier to follow that 50% greater or 33% less.
 
People will be pretty dissapointed when they find that the PS4 advantage over One is no greater than 360 difference over PS3 at gaming, once the likes of Digital Foundry start with their Faceoffs.

Better AA, steadier frame rate and better subtle effectes here an there such a SSAO. Don't expect more.
 
Console launches are both great and horrible.. It's only this time when you'll have people comparing Flops and how many transistors each console has.

Hint, transistors don't mean a damn thing. What does matter is the GPU, the CPU, the RAM, the bottlenecks between each of the above, and the tools that developers have to program the above.
Is it totally incomprehensible to you that the reason I wanted to know that was not to start some transistor dick-waving contest, but merely academic curiosity?

People will be pretty dissapointed when they find that the PS4 advantage over One is no greater than 360 difference over PS3 at gaming, once the likes of Digital Foundry start with their Faceoffs.

Better AA, steadier frame rate and better subtle effectes here an there such a SSAO. Don't expect more.
You have to remember that people's expectations based on the perceived discrepancy in numbers are totally subjective. Those who do not understand what the numbers mean will inevitably set themselves up for disappointment.
 
People will be pretty dissapointed when they find that the PS4 advantage over One is no greater than 360 difference over PS3 at gaming, once the likes of Digital Foundry start with their Faceoffs.

Better AA, steadier frame rate and better subtle effectes here an there such a SSAO. Don't expect more.

Technically you could have rendering techniques which have a floored ceiling of performance (in terms of bandwidth or shading power)... that could lead to radically different looking games.
 
Is it totally incomprehensible to you that the reason I wanted to know that was not to start some transistor dick-waving contest, but merely academic curiosity?

No, that didn't cross my mind.

I don't think Sony released the information on the amount of transistors their APU has. It's a largely meaningless statistic, but I assume Microsoft shared it because they couldn't share other details about the CPU and GPU because they couldn't 1 up Sony.
 
People will be pretty dissapointed when they find that the PS4 advantage over One is no greater than 360 difference over PS3 at gaming, once the likes of Digital Foundry start with their Faceoffs.

Better AA, steadier frame rate and better subtle effectes here an there such a SSAO. Don't expect more.

Higher rendering resolution.
 
If gflops was an accurate description of performance we wouldn't need to do game benchmarks
ai9kxm.png
 
Top Bottom