Windom Earle
Member
Congratulations for only noticing this now.
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
Virtual testing of PS4 and XBO GPUs prove PS4 has bigger grafix numbers
- Thread starter GoofsterStud
- Start date
- Status
Congratulations for only noticing this now.
freshVeggie
Member
I don't understand. How did they make the benchmark? Do they have root access or a dev kit?
How else can a somewhat scientific and in-depth benchmark be made.
That said, the gpus from both consoles have been leaked to great detail but I'd like to know how the cpus differ from each other when a very-well known insider on gaf said that the ps4 cpu is more capable/powerful.
It's really interesting. Usually between rivaling consoles pros and cons are split but this time sony's console seems to have every pro hardware-wise; gpu, cpu, ram. Wonder what went wrong with the xone dev team.
How else can a somewhat scientific and in-depth benchmark be made.
That said, the gpus from both consoles have been leaked to great detail but I'd like to know how the cpus differ from each other when a very-well known insider on gaf said that the ps4 cpu is more capable/powerful.
It's really interesting. Usually between rivaling consoles pros and cons are split but this time sony's console seems to have every pro hardware-wise; gpu, cpu, ram. Wonder what went wrong with the xone dev team.
GoofsterStud
Member
PS4 has many advantages over a standard 7850. It has the 8 ACE / 64 Queues setup only get on a R9 290x and has access to much more DDR5 than most GPUs. That being said, a high-end PC's will smoke these machines. Which this test also didn't take into account, which should increase the PS4 numbers.
Premium eSRAM real state on the CPU/GPU die.
It also didn't take into account that PS4 GPU has 8 ACE's vs Xbox One's 2 ACE's.
They're both already dated IMO, one is just more dated than the other.
Hopefully this helps people visualize what's happening. (This is napkin math and is only used to give us an idea of what is going on, I think I am being lenient on how fast it drops to be safe)
Assumptions
1.) Only eSRAM is used if under 32 MBs
2.) The numbers given are accurate
3.) We are using the unreachable theoretical max of eSRAM
When they equal out the chart goes up, once the max threshold is met the max goes down on a curve and it will eventually drop to DDR3 numbers if we had infinite DDR3 ram. As you can see around 500MB of total ram used and your back to the eSRAM line. This is all theoretically. The question now lies in how much ram do the developers need for the visuals alone....
Ha now you made a graph anyways - and thanks for that!
I assume the X axis is MB / frame read (so any textures, meshes and whatnot that are used to render the image) and the Y axis is the relative points (same as in the graph from the article).
Well on one hand the peak would be lower (since we can safely assume that at least part of the framebuffer resides in the eSRAM, taking away from the available pool (assuming we are not doing postprocessing or other things that read from the framebuffer)). On the other hand, in an ideal case there is no need to use (a lot) more RAM to read from than the framebuffer's size (e.g. when using Mipmaps etc., not sure if tiled resources would also come into play here).
So the question of how much RAM the developers need for rendering a given frame might not be that obvious - even if they certainly use as much as possible of the 8 (well 5 or so) GB of available DDR3 RAM (edit) as a buffer for the slow HD / Disc reads. If you have an extreme case like in KZ:SF where they use hundreds of megs (or even more?) for framebuffers alone, the answer is clear. But if you are having a heavily optimized case, it might get quite a bit closer to the theoretical peak..
Thanks again for your effort, I do think I understand the situation a lot better now ^^
GoofsterStud
Member
NP! No matter how much they optimize though, they are still going to use at least .5 GB probably a GB, so that's why I keep pointing to the eSRAM line. If anything, you probably helped me demonstrate a point I wasn't achieving through my broken English.
Also I don't know how to label the axis on that software. :-( It's not the one I use at work and don't care to learn. lol
Basically the only ones that could probably use the full benefit of the ram are games without textures and effects...
hydrophilic attack
Member
Long have PS4 fans claimed the PS4 to be so much more powerful, citing numbers like 50 % more powerful. But these numbers and some simple math show that the difference is much smaller than that.
The PS4 bar is at 42 points and the first XBO bar at 28 points.
42 - 28 = 14
This would make the PS4 14 % more powerful, which by itself is almost negligible. Add to this the fact that the article states that optimal memory management would give better results than the top bar, and the performance improvements of DirectX 12 and beyond and the power of the clowd, and we should arrive at parity or maybe even make XBO slightly more powerful.
So worst case scenario: PS4 is 14 % more powerful. Best case scenario: XBO is slightly more powerful. So I don't see how this article causes so much celebration among Sony Gaf.
The PS4 bar is at 42 points and the first XBO bar at 28 points.
42 - 28 = 14
This would make the PS4 14 % more powerful, which by itself is almost negligible. Add to this the fact that the article states that optimal memory management would give better results than the top bar, and the performance improvements of DirectX 12 and beyond and the power of the clowd, and we should arrive at parity or maybe even make XBO slightly more powerful.
So worst case scenario: PS4 is 14 % more powerful. Best case scenario: XBO is slightly more powerful. So I don't see how this article causes so much celebration among Sony Gaf.
I thought GAF had wizened up to this shit site by now but I guess it will take a little more time.
Here's one of their previous works:
oh god misterxmedia. You have to be a special person to source a mentally deficient nutcase like misterxmedia or one of this half dozen personas. the whole logic on that site is pure stupidity. if you have amazing hardware you tell people even if the software side isn't ready to take full advantage of it. letting the competition gain market share and fanboys feed you shit while you keep your hardware power a secret makes zero business sense, even a fool can see that.
GoofsterStud
Member
Math fail or funny guy?
licktheenvelope
Member
Yes, yes I do.
Dreamgazer
Member
Joke post.
And another joke post.
fanboy maybe
Nothing whatsoever. But please don't spoil the fun.
We already have SenjutsuSage on board, just need TheKayle and LittleJohnny to join in and we're possibly onto another winner. And if things take off enough and we're really lucky, we might even get Pennello back in.
Surely you wouldn't want to miss that?
I don't think we'll ever see Penello in a tech specs / perfs comparison thread again.
Your tag betrays you, friend.
GhostTrick
Banned
There's no need of such benchmarks. Concrete facts such as clockspeed and numbers of compute units on a GCN GPU is enough to tell the difference.
And it's the case for any AMD GCN GPU.
And it's the case for any AMD GCN GPU.
DOWN
Banned
My heart goes out to this person ^
Maximus P
Member
Me too. It even has the old 'have you seen Titanfall' joke in the first few posts. I assumed this was a bump.
This graph clearly doesn't accurately represent the situation. If Xbone were only this much less powerful the gap wouldn't be anywhere near as large as it is in actual games. I think that either the bottleneck is far worse than we've been led to believe or that the Xbone's API is very slow and cumbersome.
scoobs
Member
Well ya that, and the idea that developers aren't able to "find" the hidden hardware inside the machine... I mean, come on. It's so wildly idiotic, its hard to believe anyone with half a brain could even begin to believe it
It has already been stated numerous times the Xbone is currently more difficult to develop for. The difference of power will fluctuate dependent on how well developers will continue to utilize the hardware.
Fruitless argument really.
The 30%+ "power" difference surely isn't.
It's hilarous how painfully obvious the power difference is between these two machines yet people fail to grasp that difference. its absolutely insane. this isnt a 360 / ps3 situation where they used wildly different hardware. both boxes have the same type of gpu and the same cpu its never really been easier to quantify the performance difference in console hardware.
Her dad is also John Dykstra aka pioneering Star Wars and general FX legend, so she has all the nerd bullet points (including the bonus ones) covered.
scoobs
Member
Well that's not the whole story though. The two systems have pretty substantially different memory solutions, which can be hard to quantify the real-world performance differences. ie... eSRAM + ddr3 vs gddr5 and how they effect performance. But ya, GPU wise its pretty cut and dry the PS4 is just better.
But the xbox one's memory setup only serves as a hindrance really, There is no way It can put itself above the PS4's memory setup.
And yet we may never know because it will be locked to a platform. Is there any game thus far where a PS4 outperforms the 7850 though?
JordanN
Banned
Is there any reason Sony just doesn't announce PS4 is more powerful? It's kind of a pet peeve of my mine when information like this comes from secondary sources instead of Sony just making it widely known (just like how Nintendo keeps hiding their specs even though it still gets leaked).
If it's because MS or Nintendo would get pissed, that's their problem.
If it's because MS or Nintendo would get pissed, that's their problem.
scoobs
Member
They have, on numerous occasions, called the PS4 the most powerful console in the world.
aspiegamer
Member
It'd look dickish so there's really no need to push it constantly. They have a hype, momentum, value perception and pop culture advantage right now without even having to talk about what's in the box. Besides, they have officially put out numbers (Tflops, etc) if anyone asks for statistics.
People who know/understand/care all know it already.
People who don't know/understand see PS4 with the superior third-party versions. Same impact.
JordanN
Banned
I keep looking for the quotes but they show up so infrequently (last time it was march, before that it was last year?).
It was nothing like back in day when say Sony, wanted to convince you PS2 was the matrix that would launch missiles for Saddam Hussein while rendering Toy Story and Star Wars.
They have already said several times it's the "most powerful console ever made".
And we have the specs.
Just look at the spec sheets and you can see that the PS4 is clearly more poweful.
I mean, look at this (pic on the site, which the OP mentioned):
Xbone: 1.31 TF GPU (12 CUs) for games
Xbone: 768 Shaders
Xbone: 48 Texture units
Xbone: 16 ROPS
Xbone: 2 ACE/ 16 queues
PS4: 1.84TF GPU ( 18 CUs) for games
PS4: 1152 Shaders +50%
PS4: 72 Texture units +50%
PS4: 32 ROPS + 100%
PS4: 8 ACE/64 queues +300%
We are surely not talking about a 14% difference here
And the PS4's better memory/RAM solution is another big advantage.
But yes.... we already know all that (for months) now.
Laughing Banana
Weeping Pickle
Wait, hasn't it always been obvious that PS4 is more powerful than XboxOne............?
We must go deeper to discover the truth.
Yes, they have different memory solutions. One is far better. That is the difference.
Hah, indeed.
scoobs
Member
I agree the ps4's memory solution is better in general, but it provides an area where you could argue in certain scenarios the esram solution could outperform the ps4's (although I'm struggling to think of any) just due to pure speed. But ya, I'm definitely not saying the xb1's solution is better in any way, just that at least debatable that there can be cases where its close.
- Status
Similar threads
- Poll
- Poll