• 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.

Dusk Golem reiterates that the Xbox will be more powerful than the PS5 (Admitted to starting console wars, demodded)

Mister Wolf

Gold Member
Is this a riddle? The XSX isn't a 2080ti equivalent and the PS5 isn't a 2070 Max Q equivalent so what the hell are you saying?



Sure, it's cone tracing. Producing similarish results as RT on a PS4 game. Is that impressive to you?

Just say it produces inferior results. If the results were close to being equal then it wouldn't have been a push to do real time triangle raytracing in the first place.
 

JonnyMP3

Member
guys, they both are using amd rdna2.
and they both are going to be running the same games from the entire 3rd party industry
but... whatever floats your boat I guess
It's an analogy, go ahead get exact if you like though.
It's not like the actual hardware is physically different at all is it?
And yeah it's the same manufacturer... But not like Mercedes, Renault, Ford etc etc don't supply their engines to other car makers and then the car makers customise to their needs.... Oh wait, both Sony and Microsoft have customised stuff.
And as for 3rd party... That's kinda like the public roads. Access for everyone instead of their respective races.
 

Thirty7ven

Banned
Just say it produces inferior results. If the results were close to being equal then it wouldn't have been a push to do real time triangle raytracing in the first place.

Point being nobody is impressed by what goes behind the image, they only see what's in front of them.

And I know you're a hater, but I thought you were above spreading big time bullshit, but with this "2080ti(XSX) vs 2070 Max Q(PS5)" stunt? Add another to the pile I guess.

Seems like the Halo fiasco didn't humble you at all. You should've listened to me before, and you should listen to me now.
 
Last edited:

JonnyMP3

Member
This applies to both systems though, that includes PS5.

I think people are mistaking variable frequency to mean the system is running at peak clocks by default and takes dips when power budget is exceeded. But that makes no sense. Both systems will only provide the clocks and power that a scene requires. If that's peak, then peak (or just about; 100% peak won't be hit by either system) is provided. If it's 6 TF, then 6 TF is provided. If it's even just 1 TF, then 1 TF is provided.

The variable frequency kicks in when peak is required and needs to be required for a period of time that may exceed the normally allotted power budget. In that case, power is diverted from the CPU (as an example) to give the GPU more of a kick. If no additional power from other processor components can be routed, then the GPU will downclock. The rate of all this depends on the severity of power certain operations/instructions would require, the consistency of those operations (in overall cycle time for any given stretch of time), and requirement of power by other processor components like the CPU and Tempest engine. I'm particularly interested on what effect transient results of persisting simultaneous instructions/calculations (like a somewhat power-intensive instruction loop running alongside intensive non-looping instructions over a measure of cycle time) will have on power draw and could trigger a GPU downclock through reducing the power load, and just how large that would be in practice (i.e if Cerny's 2% frequency reduction holds true most of the time or if that was a theoretical expected range).

So PS5 monitors the power state to determine shifting and reduction of power which in turn will adjust frequency, once peak power budgets are being stressed. MS's systems OTOH aren't monitoring for power; if a sequence of operations need a certain amount of power, the system will always provide that to them and simply increase the cooling to compensate. The main advantage with Sony's approach is that the power load adjustment is very quick, a couple of ms supposedly. MS's approach's advantage is that devs don't need to be as mindful of power management when it comes to their code, since the systems will just ramp up the cooling to compensate for excess if required.
You've just written a wall of text explaining why 12TF isn't the be all and end all. Thank you.
 

Mister Wolf

Gold Member
Point being nobody is impressed by what goes behind the image, they only see what's in front of them.

And I know you're a hater, but I thought you were above spreading big time bullshit, but with this "2080ti(XSX) vs 2070 Max Q(PS5)" stunt? Add another to the pile I guess.

Speak for yourself. Also whether that was through ignorance or just being disingenuous please stop calling everything you see "raytracing".
 

Bogroll

Likes moldy games
I wonder if these people here realise that Redlining your car engine constantly isn't healthy. The engine revs go up and and down depending on how much and hard you put your foot down. The principle is the same, you can't run everything at max all the time. Oh well... Physics hey!
Yeah stick a limiter at 5500rpm and cruise at that (XSX) The less powerful car(PS5) let it have a redline but don't let it stay there too all the time and its still down on power but it narrows the gap.
 

MastaKiiLA

Member
It's not a cop-out but I damn sure don't like people holding me responsible for shit that I hear when it doesn't come out as accurate. If I worked at these companies, that would be a different matter. I'd take all the blame right there.

My contacts are smart enough not to judge by theoretical numbers. You are talking programmers that run tests on realworld hardware. I did the same thing. And no, none of them will give details like what you guys are looking for so forget it.
So don't share information if you don't know the veracity of it. You realize you ran into the same exact trap that Dusk Golem threw himself into, in a thread about Dusk Golem throwing himself into such a trap, right?

Like...the irony is...just...wow!

Also, I'd be curious to know more about your developer friends who run these tests, because I can tell you that I manage a team of developers myself, and at no time has there been any value in benchmarking. The spec of the box doesn't hold a whole lot of value when all you care about is the performance of your actual code. Given that processor architecture varies from console to console, why exactly are they spending resources (time) on with rewriting benchmarks for each new system, when the ultimate value is in execution time of their code?

They might find a difference in the speed of each system in processing certain processes that they need to execute, but how is that supposed to translate into FLOPS exactly? Or are they pacing for loops for intra-office dick measuring contests? I don't exactly see how coders would be able to extrapolate anything in relation to the TFLOPS pissing match that has made up the vast majority of the console warring. They would see a % time reduction (speed increase) in executing their code on one system over another. If you were provided that information, it would be worlds more valuable than trying some backhanded comment about the TFLOPS gap being bigger than the rated values.

You suggested that people might not be able to distinguish your reviews of PS5 games from your takes on the forum, in another thread, and this is why. You can't engage in console warring, and then pretend that you're somehow above the fray. You're very much part of the chaos that engulfs these console launches. You've espoused the same "leveling the playing field" type of argument that Dusk Golem has. There's no reason anyone should take your rumors as anything more than FUD.
 
Last edited:
It's an analogy, go ahead get exact if you like though.
It's not like the actual hardware is physically different at all is it?
And yeah it's the same manufacturer... But not like Mercedes, Renault, Ford etc etc don't supply their engines to other car makers and then the car makers customise to their needs.... Oh wait, both Sony and Microsoft have customised stuff.
And as for 3rd party... That's kinda like the public roads. Access for everyone instead of their respective races.
I am really trying hard to agree on something with you here...
can you at least agree that:

- when ps5 is at its best case scenario, when at highest possible gpu clock, the difference is 2 teraflops
- 2 rdna2 teraflops are bigger difference than the 2 GCN teraflops xbox x-ps4pro had?


cerny.jpg


Thus, the teraflop difference this time is more significant than xbox x - ps4pro?

I mean, its simple logic, normally not even up for discussion, yet I've spent a good portion of two hours while some go around in circles
 
Last edited:

Thirty7ven

Banned
Speak for yourself. Also whether that was through ignorance or just being disingenuous please stop calling everything you see "raytracing".

I don't label everything I see as ray tracing. And cone tracing has similarities to ray tracing. But while I don't have a problem saying that what you're saying is fair enough, Tomorrow Children doesn't have RT and I was wrong, I also don't have a problem pointing out you're distancing yourself from the bullshit comparison.

And trying to pass yourself as any sort of tech/rendering expert when you're out here comparing 2080ti to 2070 Max q to drive home a point about RT and XSX vs PS5 says it all. Just another one of those Xbox fans who defaults to PC master race when the going gets rough.

Yo remember when I told you the Discover Hope trailer was representative of the visuals you would see on Halo XSX? And you said no way, that I didn't know what I was talking about? That's how much you know about this. Still got the crow here for you to feast if you want.
 
Last edited:

Tripolygon

Banned
How about you use showcase Lumen(VXGI mixed with inferior less demanding techniques) in Minecraft or RTGI in that demo. You cant? Then what was the point of you making this comparison.
That is literally what that video demo shows. Lumen vs Path Tracing running "realtime". You're asking me to take the minecraft demo and replace the lighting with Lumen? What sort of stupid request is that?

Tracing rays for audio and convolution reverb is raytracing too. Voxel cone tracing is raytracing. There are all sorts of raytracing.

And what modern games do is tracing a few rays 1spp or 2 spp through a bvh which is a smaller less detailed representation of the scene, and using a denoiser to fill in holes and smoothen out the image with lots of sharpening to bring out more detail , done over a couple of frames, you know, sort of like voxel cone tracing. Before you call someone ignorant, try being less ignorant yourself. It is for the same reason a lot of games use temporal reconstruction rather than native resolution. Get more details out of less for smaller cost.
 
Last edited:

Mister Wolf

Gold Member
I don't label everything I see as ray tracing. And cone tracing has similarities to cone tracing. But while I don't have a problem saying that what you're saying is fair enough, I also don't have a problem pointing out you're distancing yourself from the bullshit comparison.

And trying to pass yourself as any sort of tech/rendering expert when you're out here comparing 2080ti to 2070 Max q to drive home a point about RT and XSX vs PS5 says it all. Just another one of those Xbox fans who defaults to PC master race when the going gets rough.

Yo remember when I told you the Discover Hope trailer was representative of the visuals you would see on the XSX? And you said no way, that I didn't know what I was talking about? That's how much you know about this.

The entire point of the benchmark was to show that the Series X RT performance is right there with the Nvidia cards. So if it can do that on a fully path traced game then why the hell couldn't it do the reflections showcased in R&C and GT7 that's culling/omitting objects that can be reflected to save performance. The RT reflections shown in those games aren't even on the level of what was in Battlefield
 
12.15 TF will NEVER be fully utilized with locked clocks. 12.15 is theoretical max. And some power will be left on the table. 10.28 TF is also a theoretical max on PS5. But it is a much more achievable with variable clocks

I wouldnt say this for sure .... I think you can see it fully utilized but once you DO that you must begin to adjust resolution etc if you need to improve performance or remove dips.
This is what they did for RE3 on One X, because they were up against the GPUs limit- which was higher than the PS4 Pro but not enough to make 4k smooth enough
in all scenes. I do think there are going to be times the Series X has some kind of measurable advantage- AND times its better optimized for. I dont think thats off the table at all.
 
Expected outcome.

We knew he was full of nonsense when he repeatedly failed to name other games -- outside of an unoptimized early build of Re8 -- that were supposedly struggling to run at a native resolution/60FPS on PS5. Demodding him was the right move because his need to defend Philip Spencer's upcoming toy is a reflection of his ability to make mature decisions.

I hope he understands that the reason people were and are disappointed with Halo Infinite and the Xbox Series X has nothing to do with the PS5 and instead, has everything to do with Xbox leadership overpromising and underdelivering again. It's really that simple.
 

Thirty7ven

Banned
The entire point of the benchmark was to show that the Series X RT performance is right there with the Nvidia cards. So if it can do that on a fully path traced game then why the hell couldn't it do the reflections showcased in R&C and GT7 that's culling/omitting objects that can be reflected to save performance. The RT reflections shown in those games aren't even on the level of what was in Battlefield

"Nvidia cards" what is this? That's a big range. The 2080ti was doing 60 average at 1080p, 60 4k with DLSS 2.

But yeah it did flew right over my head that you were comparing XSX RT performance to a 2070 Max Q. That's weird, I didn't get that at all. Odd comparison.
 

JonnyMP3

Member
I am really trying hard to agree on something with you here...
can you at least agree that:

- when ps5 is at its best case scenario, when at highest possible gpu clock, the difference is 2 teraflops
- 2 rdna2 teraflops are bigger difference than the 2 GCN teraflops xbox x-ps4pro had?


cerny.jpg


Thus, the teraflop difference this time is more significant than xbox x - ps4pro?

I mean, its simple logic, normally not even up for discussion, yet I've spent a good portion of two hours while some go around in circles
I never disagreed with any of that.
Factually the XSX is 2TF stronger than the PS5. No one's been denying that. At all! But with the racing analogy of lots and lots of customised tech, that RAW power is just top end that the engine can do. But as mentioned by others, there's also other stuff on the hardware that makes the 2 machines completely different.
What I'm trying to get at is that despite the PS5 being the weaker of the machine, it does not matter because of all the clever custom tech.
F1 has KERS, DRS, CFD Aerodynamics etc etc.
But when it comes to pure GPU processing power and ability, yes the Xbox wins hands down because it has more CUs which is why I've compared it to V8 rather than a V6.
What's the issue mate?

FYI: I would have no problem owning a V8 car. Actually I'd prefer a V8 over a V6. 🤷🏻‍♂️
 
Last edited:

Clarissa

Banned
I am so tired of the Sony fans always crying about their ridiculous conspiracy theories; Digital Foundry, Dusk Golem, Xbox fans all paid shills, GitHub, FUD, FUD everywhere, Jeff Grub, Xbox Discord, Tom Warren and on and on and on. The whole world is plotting against poor Mother Sony.

Here is a website I think Sony fans would enjoy:
th

Typical Sony fan
Are you on your period now? You sound cranky. :messenger_squinting_tongue: :messenger_squinting_tongue: :messenger_squinting_tongue:

I hope this helps.

 

VFXVeteran

Banned
So don't share information if you don't know the veracity of it. You realize you ran into the same exact trap that Dusk Golem threw himself into, in a thread about Dusk Golem throwing himself into such a trap, right?

You are right. And I've already toned that down significantly.

Also, I'd be curious to know more about your developer friends who run these tests, because I can tell you that I manage a team of developers myself, and at no time has there been any value in benchmarking. The spec of the box doesn't hold a whole lot of value when all you care about is the performance of your actual code. Given that processor architecture varies from console to console, why exactly are they spending resources (time) on with rewriting benchmarks for each new system, when the ultimate value is in execution time of their code?

I never said they ran benchmarks in the way you are thinking about them. When we get a new piece of hardware and or software (i.e. an OpenGL version update for example), we would import real world assets into our renderer and see how the FPS deals with it and many options in our shaders that are expensive. Not measuring FLOPS at all. It's the real world performance that matters. That's why when I asked a Sony dev what was the performance like on the PS5, he didn't give a FLOP number, he compared it to a graphics card (i.e. 1080-1080Ti). You can gleen a LOT more information from that than a TFLOP number.

You suggested that people might not be able to distinguish your reviews of PS5 games from your takes on the forum, in another thread, and this is why. You can't engage in console warring, and then pretend that you're somehow above the fray. You're very much part of the chaos that engulfs these console launches. You've espoused the same "leveling the playing field" type of argument that Dusk Golem has. There's no reason anyone should take your rumors as anything more than FUD.

I've repeatedly stated that I'm not a console warrior. And you are right, they can and have taken a lot of my rumors as FUD - despite them eventually being true (i.e. Sony exclusives coming to PC, PS5 = 1080Ti performance, and some other things). I will most definitely keep my mouth shut in the future about anything I may know. I have no stake in this race like many of you.
 
Last edited:

Bogroll

Likes moldy games
I never disagreed with any of that.
Factually the XSX is 2TF stronger than the PS5. No one's been denying that. At all! But with the racing analogy of lots and lots of customised tech, that RAW power is just top end that the engine can do. But as mentioned by others, there's also other stuff on the hardware that makes the 2 machines completely different.
What I'm trying to get at is that despite the PS5 being the weaker of the machine, it does not matter because of all the clever custom tech.
F1 has KERS, DRS, CFD Aerodynamics etc etc.
But when it comes to pure GPU processing power and ability, yes the Xbox wins hands down because it has more CUs which is why I've compared it to V8 rather than a V6.
What's the issue mate?
Other than faster SSD what other custom tech do we know is better than the custom tech that XSX has ?
I've kept out of most next gen etc thread for a while now so I'm not up to date.
 

Bogroll

Likes moldy games
So apart from the custom SSD so far we know it also has a Custom I/O unit, custom Geometry Engine in GPU and a separate Tempest Audio processor.
Are those better or worse than XSX's custom tech ? I thought XSX has its own dedicated audio processor ?
 

JonnyMP3

Member
Are those better or worse than XSX's custom tech ? I thought XSX has its own dedicated audio processor ?
We have no idea what the XSX sound chip does atm this is why we're all waiting for the possible hotchips teardown.
But the PS5 audio chip is designed to work with HRTF.
 

Bogroll

Likes moldy games
We have no idea what the XSX sound chip does atm this is why we're all waiting for the possible hotchips teardown.
But the PS5 audio chip is designed to work with HRTF.
I'm not to fussed about that, yes i appreciate sound but most of us just use tv for sound.
 

Mister Wolf

Gold Member
"Nvidia cards" what is this? That's a big range. The 2080ti was doing 60 average at 1080p, 60 4k with DLSS 2.

But yeah it did flew right over my head that you were comparing XSX RT performance to a 2070 Max Q. That's weird, I didn't get that at all. Odd comparison.

Ok I will clear it up:
DLSS2_0_Minecraft_2.png


In 4 weeks of work from one engineer, he ported the official pathtraced version of Minecraft to the Series X and its 2060 Super/2070 level in performance. Also I don't have to be an expert to learn through reading and listening to explanations from more knowledgeable people.
 
Last edited:

jimbojim

Banned
I wouldnt say this for sure .... I think you can see it fully utilized but once you DO that you must begin to adjust resolution etc if you need to improve performance or remove dips.
This is what they did for RE3 on One X, because they were up against the GPUs limit- which was higher than the PS4 Pro but not enough to make 4k smooth enough
in all scenes. I do think there are going to be times the Series X has some kind of measurable advantage- AND times its better optimized for. I dont think thats off the table at all.

That's why PC GPU clocks aren't locked. Anyway, this very good post why variable clocks are great. Worth reading :

 

Bogroll

Likes moldy games
It's designed for normal TVs as well as theatre set ups and headphones.
I've bought enough Sony tv's in my time with whatever the latest sound at the time was they were pushing True Force Surround etc. So I'm a bit sceptical about that. I've bought a Sony Atoms soundbar claims 7.2 surround, while its great its nowhere near as good as having proper surround speakers. So i think people should be a bit realistic about what they are going to get out of their tv speakers regarding PS5 and i bet most folk won't even know the the difference on tv's.
 

JonnyMP3

Member
I've bought enough Sony tv's in my time with whatever the latest sound at the time was they were pushing True Force Surround etc. So I'm a bit sceptical about that. I've bought a Sony Atoms soundbar claims 7.2 surround, while its great its nowhere near as good as having proper surround speakers. So i think people should be a bit realistic about what they are going to get out of their tv speakers regarding PS5 and i bet most folk won't even know the the difference on tv's.
That's fine. Plenty of gamers have surround sound set ups and headsets are one of the biggest selling game accessories.
So I'm sure plenty of gamers will benefit, even if the average schmuck doesn't.
 

geordiemp

Member
Are those better or worse than XSX's custom tech ? I thought XSX has its own dedicated audio processor ?

In terms of the Geometry engine, Cerny and Naughty dog patented compression of vertices and shaders, have a patent on mesh type VRS for VR (both patents are not hard to find) and have chosen their own GE design.

If the GE design with what Sony first party desire is worse than stock RDNA2, then why did they choose their own version ?

Nobody has commented on the recent Sony GE patents, other than Matt Hargatt boasting about GE. on twitter



And thats all we know so far
 
Last edited:

Bogroll

Likes moldy games
That's fine. Plenty of gamers have surround sound set ups and headsets are one of the biggest selling game accessories.
So I'm sure plenty of gamers will benefit, even if the average schmuck doesn't.
Wow! That's a bit harsh cause the average prick doesn't appreciate something you do. You should see it as a fine wine you like instead.
 

Entroyp

Member
This is a dumb argument. Since when you have to be a game developer or have technical knowledge to be an insider?
Do you believe Tom Warren know the tech detail of every thing he leaks? He only transcribe what he is being told

Tom Warren doesn’t understand shit. That’s for sure.
 
Last edited:

Bogroll

Likes moldy games
In terms of the Geometry engine, Cerny and Naughty dog patented compression of vertices and shaders, have a patent on mesh type VRS for VR (both patents are not hard to find) and have chosen their own GE design.

If the GE design with what Sony first party desire is worse than stock RDNA2, then why did they choose their own version ?

Nobody has commented on the recent Sony GE patents, other than Matt Hargatt boasting about GE. on twitter



And thats all we know so far

So really when all the pluses and minuses for both machines are added to put in the pot, it boils down to the DF/NX Gamer face offs in the end. Despite what anyone says on here so far, the proof will be in the pudding.
 

JonnyMP3

Member
Wow! That's a bit harsh cause the average prick doesn't appreciate something you do. You should see it as a fine wine you like instead.
😂 🤣 😂 🤣 😂
That's because I knew you were taking it personally, fanboy!
I was just filling you in with what the custom tech was when you asked.
See... The thing is you're talking about 'My likes' which was NEVER MENTIONED AT ALL! I just knew what game you were playing.

But the fact that you're baited to being labelled a 'schmuck' because you chose to align yourself as one. Yeah. Grow up son! 🤣
 

93xfan

Banned
Xbox fans are much more louder than PS fans. Especially in FUD spreading. That's a fact, a big one. Just look what are they doing on twitter. It's hilarious to see.
Regarding Github, it was right about PS5 only regarding CU and half-right about bandwidth speed. Regarding GPU speed ( 2 GHz ) architecture ( RDNA 1 ), TFLOPS 9.2 it was wrong.

That shot about Berkanstel was fake. And it was wrong way to do it. But Berkanstel was a member of Xbox ERA Discord, now is in the Discord club with "main" devs like Louise Kirby. And also, some Discord shots are real. And i will keep them like they are gold. And they are gold.
Regarding that PS5 engineer. Yes. He said that PS5 is mixture of RDNA 1 and 2. With missing one feature from 2. Maybe Sony didn't needed that one feature from RDNA 2 because they developed something better. Also logic says that XSX also is a mixture of RDNA 1 and 2 because surely XSX has features from RDNA 1. Isn't it? It's like firmware updates on consoles. You download the newest one, but it also contains features from previous ones.
But regarding SSD, yes, it's way better on PS5 than on XSX. That's not an overplay when difference between the two is over 100%

you don’t own a PS4 pro and Uncharted 4, do you?
 

Bogroll

Likes moldy games
😂 🤣 😂 🤣 😂
That's because I knew you were taking it personally, fanboy!
I was just filling you in with what the custom tech was when you asked.
See... The thing is you're talking about 'My likes' which was NEVER MENTIONED AT ALL! I just knew what game you were playing.

But the fact that you're baited to being labelled a 'schmuck' because you chose to align yourself as one. Yeah. Grow up son! 🤣
ERRH!
 

Bogroll

Likes moldy games
Indeed!
See the moment you started talking about 'Your Sony TV' and 'Your Sony Sound bars' yeah, I could see you weren't actually bothered about the custom tech 😂
Yeah i bought one for the tv in my kitchen as the tv doesn't cut in with the noise from the Wife cooking.
Do you want me to post pics of my tv's or something lol, just cause you my love everything you buy doesn't mean i have to. The tv's are great but not the sound.
 
Last edited:
Top Bottom