Support NeoGAF

[mckmas8808]

Something else to think about. Yes the PS5 is 18% less powerful than the XSX. But if Lockhart is real and really is 4TFs (because it's designed with 1080p TVs in mind), then you are dealing with a situation where the XSX has 300% more power than the baby Lockhart. Yes, we have a dev(s) that are claiming that due to "engine scaling" those games will run just fine and look just as good (but just at a 1080p resolution, which means blurrier). And a dev specifically here said that it's as easy as hitting a toggle switch on the game code to switch it from 4K to 1080p to run on Lockhart. He said raytracing and all would still work great on Lockhart. I believe that dev is Gavin Stevens

So with that being said, why couldn't 3rd parties just take a few weeks time to make their games that have RT and decrease the resolution by 18% [the game would basically be 2000p(7.0 million pixels instead of 8.29 million pixels)] and then use checkboard rendering to achieve 4K (like) resolution? I doubt literally any regular person would be able to notice the difference with their own eyes.

That's alot of assumptions made off of generlizations here. Me doing the same, assumes that the compute advantage will win out on rasterization and rt and noticeably so. All of the vrr, mesh shading, SSD benefits, etc.. will be on the XSX as well.

Again, while I absolutely could be proven wrong, it sounds like the same rationalizing that xbox fans made back in 2013. This causes me to be very skeptical

Give me one assumption that I made in that post. Those numbers are numbers that MS and Sony have given us about their next-gen and current-gen consoles. We already know what's possible, given this current gen's performance differences with the PS4/PS4 Pro and the Xbox One/Xbox One X.

 

[BattleScar]

The Series X is more powerful than the PS5. There is no debating this. There is no way you get around a 44% shader deficit. The question is how big is the difference.

 

[ThaMonkeyClaw]

That's insane! So there's a whole Zen2 CPU worth of power in the PS5 just to decompress data to take the load off of the CPU and GPU? Geez! And a PS4 CPU worth of power to crunch the sound? Talk about smart engineering. But it also sounds expensive too. With a cooling solution that has to cost more than the XSX, this ready does sound like a $500 console to me.

But that's cool with me, because it sounds like the CPU and GPU will be free to actually crunch videogame data, instead of all the boring stuff that it used to do in the past.

Well the BOM from what we have heard is right around $450, so if they decide to just sell the system at cost we are looking at a $450 console, but even if they do go to $500 I am fine with that. People keep looking over all these co-processors and other customizations, which I am not saying it is creating extra TF or anything, but it frees up the main CPU/GPU for just game graphics where as other systems the CPU/GPU have to jump in for some of this stuff and I think that in the end it is going to be a more even playing field than people are thinking.

 

[pasterpl]

Microsoft Just Made It a Lot Easier to Port Cool PC Games to Xbox

Today, Microsoft announced a new DirectX standard that it claims will not only sync the graphics improvement process in games across both PC and console, but will also be backwards compatible with older PC hardware. If it works as explained it’s going to be a lot easier to port games from PC to...

gizmodo.com

 

[Md Ray]

Microsoft Just Made It a Lot Easier to Port Cool PC Games to Xbox

Today, Microsoft announced a new DirectX standard that it claims will not only sync the graphics improvement process in games across both PC and console, but will also be backwards compatible with older PC hardware. If it works as explained it’s going to be a lot easier to port games from PC to...

gizmodo.com

UPDATE: In a statement provided to Gizmodo from AMD, the RX 5000-series and older GPUs will not support DirectX 12 Ultimate. However, “next-generation GPUs based on AMD RDNA 2 gaming architecture will support DirectX 12 Ultimate,” says AMD.

Wow, didn't know RDNA 1 GPUs would get outdated this quickly.

 

[Rolling_Start]

I'm particularly interested in the Machine Learning capabilities of the XSX. Quad rate int8 and Octo rate int4.

I've no idea how the magic happens on an algorithmic level, but Nvidia DLSS is crazy impressive wizardry when it's working right. The idea of console being able to enhance images and assets on the fly using ML is really next gen stuff for me. Hopefully MS throw their weight behind an research and platform tools for this.

Fingers crossed.

 

[SSfox]

PS5 To Support AMD SmartShift For Improved Optimisation On PS5 Games - PlayStation Universe

PS5 is set to support AMD's SmartShift technology - read on to discover why that will be a game changer for PlayStation 5 games.

www.psu.com

 

[Bo_Hazem]

Fanboys are a hilarious bunch.

I love those parodies/banter even against Sony Wonderfully made.

 

[kyliethicc]

I find that unrealistic.

Because it’s not true. There is an EXPANSION bay. Expansion is addition. Not replacement.

 

[kyliethicc]

wow so one Nvme Slot ?

Nope this is incorrect. The SSD is not removable. You can add to it.

That Eurogamer article is wrong because it starts off with the “PS5’s hard drive” which is does not have. The PS5 has no hard drive. So it literally leads with a huge error. Don’t trust it.

 

[Bo_Hazem]

Sony’s first party studios FTW!

Nah, mediocre, this:

Spoiler: try not to laugh challenge

 

[tryDEATH]

PS5 To Support AMD SmartShift For Improved Optimisation On PS5 Games - PlayStation Universe

PS5 is set to support AMD's SmartShift technology - read on to discover why that will be a game changer for PlayStation 5 games.

www.psu.com

Since SmartShift is a laptop technology, what was it primarily designed for? Was it solve the lack of cooling in laptops or a bottleneck of overall peak performance so they came up with this solution that Sony ended up adopting?

 

[VanSiriusGabriel]

1: games
2: ecosystem
3: power

 

[Bo_Hazem]

Is he Scottish?

 

[BattleScar]

Since SmartShift is a laptop technology, what was it primarily designed for? Was it solve the lack of cooling in laptops or a bottleneck of overall peak performance so they came up with this solution that Sony ended up adopting?

When gaming, the CPU will rarely if ever be constantly at 100% utilisation. There are points where it has to wait for the GPU to finish rendering a frame or perform calculations or whatever. During those periods, the CPU doesn't really need to be at maximum power draw. SmartShift is supposed to take power away from the CPU during those periods and pump it to the GPU, so it can work a little bit faster.
Its very clever, and is a great tech to have in consoles and Laptops, where you have a strict power constraint.

Is he Scottish?

Yes he is.

 

[Ascend]

Okay think about it like this....

- PS4 had 40% more power than the Xbox One. The result? Games were in 1080p on the PS4 and 900p on the Xbox One. That difference in resolution is also about 40%. Plus the PS4 had about 130% more memory bandwidth than the Xbox One. There was a difference is the same games this gen but it wasn't a huge gap that truly mattered.

- The Xbox One X is 45% more powerful than the PS4 Pro. The difference in most games is that some games are 4K on the One X, whereas they are 1650 to 1800p on the Pro. Checkerboard Rendering on the PS4 Pro gets the game up to 4K or "faux K" and most people can only tell a litter difference.

- The XSX is 18% more powerful than the PS5. And the RAM bandwidth is only 20% faster (remember the PS4's bandwidth was like 130% faster than the Xbox One). This time the PS5's SSD is about 130% faster and it has more hardware in the system to help eliminate more bottlenecks than the XSX. So yes, the average 3rd party game will technically have more resolution on the XSX, but due to the SSD speeds and "faster" GPU some things will run faster on the PS5. So, in the end, most 3rd party games will be within 10% of each other. And at 4K resolutions, DLSS for quality edges, and VRR to help with performance almost nobody will be able to tell the difference in that 10%. Digital Foundry will because they'll zoom into the same character with a 500% zoom to show you the difference. But if you played both games on the same TV, you'd probably wouldn't notice much of a difference.

I'm not sure if % is the best way to go about it though. Percentage wise, the XSX is closer to the PS5 than the previous generation consoles are to each other. That is definitely true. However....

In terms of pure processing power, the difference is as big as the difference between the PS4 Pro and the Xbox One X.
Or to put it more dramatically, the difference between the XSX and the PS5 is pretty much 4 times as big as the difference between the PS4 and the original Xbox One.
Or putting it yet another way, the PS5 is about 5.5 times more powerful than the original PS4, while the XSX is over 9 times more powerful than the original Xbox One.
Or putting it yet another way again... The XSX is like having the PS5, and still having a full PS4 available for additional graphical processing power.

 

[ph33rknot]

Imagine the meltdown... when a PS5 exclusive looks better than Halo Infinite and any other MS exclusive at launch and throughout the gen?

Isn't Halo infinite crossgen if it don't that would be sad

 

[farmerboy]

Sony’s first party studios FTW!

The REAL secret sauce. Sony's first party is easily worth a couple of teraflops , what say you GAF?

Imagine the meltdown... when a PS5 exclusive looks better than Halo Infinite and any other MS exclusive at launch and throughout the gen?

It'll happen.

 

[ph33rknot]

I know it came out now, but Doom Eternal makes Halo looks like my ass.

You only saw a cutscene

 

[Joho79]

Okay think about it like this....

- PS4 had 40% more power than the Xbox One. The result? Games were in 1080p on the PS4 and 900p on the Xbox One. That difference in resolution is also about 40%. Plus the PS4 had about 130% more memory bandwidth than the Xbox One. There was a difference is the same games this gen but it wasn't a huge gap that truly mattered.

- The Xbox One X is 45% more powerful than the PS4 Pro. The difference in most games is that some games are 4K on the One X, whereas they are 1650 to 1800p on the Pro. Checkerboard Rendering on the PS4 Pro gets the game up to 4K or "faux K" and most people can only tell a litter difference.

- The XSX is 18% more powerful than the PS5. And the RAM bandwidth is only 20% faster (remember the PS4's bandwidth was like 130% faster than the Xbox One). This time the PS5's SSD is about 130% faster and it has more hardware in the system to help eliminate more bottlenecks than the XSX. So yes, the average 3rd party game will technically have more resolution on the XSX, but due to the SSD speeds and "faster" GPU some things will run faster on the PS5. So, in the end, most 3rd party games will be within 10% of each other. And at 4K resolutions, DLSS for quality edges, and VRR to help with performance almost nobody will be able to tell the difference in that 10%. Digital Foundry will because they'll zoom into the same character with a 500% zoom to show you the difference. But if you played both games on the same TV, you'd probably wouldn't notice much of a difference.

So how do we know Sony has more hardware to eliminate bottlenecks? Things get said so often that it turns in to some weird truths.

 

[tryDEATH]

When gaming, the CPU will rarely if ever be constantly at 100% utilisation. There are points where it has to wait for the GPU to finish rendering a frame or perform calculations or whatever. During those periods, the CPU doesn't really need to be at maximum power draw. SmartShift is supposed to take power away from the CPU during those periods and pump it to the GPU, so it can work a little bit faster.
Its very clever, and is a great tech to have in consoles and Laptops, where you have a strict power constraint.

I know what it does, but since this is specifically created for laptops and isn't an issue for PC's what so ever, I am wondering what bottleneck does SmartShift alleviate in laptops because it could give us an insight into the PS5 as well.

 

[RagingTiger]

Yea its pretty crazy how everyone is in denial, if sony and microsoft flipped positions there would be no debate which one is stronger. Cerny is the guy who said the PlayStation 4 Pro Could Achieve 8.4 Teraflops Of Computational Power lol.
Everyone is talking about how the ps5 SSD is faster that it can run at 5.5 gigabytes per seconds raw and the 9 gigabytes per second compressed but that only matters for first party titles, do you think game developers who are going to have everything on a cross platform are going to specialize it so that one runs faster on playstaion than it does on xbox?! No you need raw compute power and the only way the ps5 is getting its compute power for its 10.3 teroflops is because they're overclocking the crap out of it, thats not going to be a sustainable boost clock. 2.25 gigahertz are you kidding me on 36 CU versus the 52 CU that are on the SeriesX.
The fact that the CPU runs slower and has no simultaneous multithreading at least according what Sony said in they're reveal. the 3.5 gigahertz versus the 3.66 base on the SeriesX and 3.8 gigahertz on the regular one with no multithreading.

Long story short at a raw comparison of the ps5 its just a flatout weaker console out of the two. The good news for Playstation fans is that sony will probably continue to have the better first party titles,

 

[VanSiriusGabriel]

You only saw a cutscene

Dude, the game is out, ya know?

 

[Hendrick's]

Nothing like a good smoke after xSeX.

 

[Self High Five]

I'm not sure if % is the best way to go about it though. Percentage wise, the XSX is closer to the PS5 than the previous generation consoles are to each other. That is definitely true. However....

In terms of pure processing power, the difference is as big as the difference between the PS4 Pro and the Xbox One X.
Or to put it more dramatically, the difference between the XSX and the PS5 is pretty much 4 times as big as the difference between the PS4 and the original Xbox One.
Or putting it yet another way, the PS5 is about 5.5 times more powerful than the original PS4, while the XSX is over 9 times more powerful than the original Xbox One.
Or putting it yet another way again... The XSX is like having the PS5, and still having a full PS4 available for additional graphical processing power.

That's some funky Scott Steiner type of math:

 

[ph33rknot]

Dude, the game is out, ya know?

Halo infinite

 

[ph33rknot]

You only saw a cutscene

And if you mean Halo 5 yes it dose

 

[Grinchy]

Anyone in here who's a dev who can answer my question? I was wondering since Microsoft put out all the games coming out will have to work on X1 does that cripple game design for third-party games? It seems like that would be extremely crippling in areas like AI development or even just stuff that could impact gameplay like environmental destruction. The reason being is the X1 has a horrible CPU/HDD. It's not the same thing scaling say a hypothetical Lockhart that had the same CPU/SSD just underclocked and weaker. Scaling down to X1 would cripple the XseX in game design for third-party would it not? I would also think this would also cripple PS5 third-party games as they're not going to give an advantage to a Sony console even if it's a possibility?

I don't know if scaling works differently now but I'm curious to know the answer to this. I would hope not.

I wouldn't worry about it. They didn't say games had to work with X1. They said their own first party games would, for 12ish months.

There is no design philosophy of making every Xbox play the newest games. They were just putting a spin on having only enhanced current gen games for a year while they catch up and get their real next-gen software put out. No 3rd parties are going to be forced to make last-gen versions of games. They may do it by choice, but none will be forced IMO.

 

[ethomaz]

I'm really trying to be open minded here, but all I hear are the same arguments Xbox fans made in 2013 "esram will make up bandwidth limitations, the 4 dedicated move engines will get the data to where it needs to be faster, CPU upclock, etc..". In the end Sony's compute advantage was the most important factor when comparing game performance.

P.S. While I have consistently kept up with console development since 1999 with the development of Dreamcast and ps2, I don't know what I'm talking about at a technical level. This is just my skeptical take.

But bandwidth are similar em both consoles... I dont believe any will have advantage in that point.

 

[mckmas8808]

I'm not sure if % is the best way to go about it though. Percentage wise, the XSX is closer to the PS5 than the previous generation consoles are to each other. That is definitely true. However....

In terms of pure processing power, the difference is as big as the difference between the PS4 Pro and the Xbox One X.
Or to put it more dramatically, the difference between the XSX and the PS5 is pretty much 4 times as big as the difference between the PS4 and the original Xbox One.
Or putting it yet another way, the PS5 is about 5.5 times more powerful than the original PS4, while the XSX is over 9 times more powerful than the original Xbox One.
Or putting it yet another way again... The XSX is like having the PS5, and still having a full PS4 available for additional graphical processing power.

But that's not how any of this works. You just made all of that up. lol

Yea its pretty crazy how everyone is in denial, if sony and microsoft flipped positions there would be no debate which one is stronger. Cerny is the guy who said the PlayStation 4 Pro Could Achieve 8.4 Teraflops Of Computational Power lol.
Everyone is talking about how the ps5 SSD is faster that it can run at 5.5 gigabytes per seconds raw and the 9 gigabytes per second compressed but that only matters for first party titles, do you think game developers who are going to have everything on a cross platform are going to specialize it so that one runs faster on playstaion than it does on xbox?! No you need raw compute power and the only way the ps5 is getting its compute power for its 10.3 teroflops is because they're overclocking the crap out of it, thats not going to be a sustainable boost clock. 2.25 gigahertz are you kidding me on 36 CU versus the 52 CU that are on the SeriesX.
The fact that the CPU runs slower and has no simultaneous multithreading at least according what Sony said in they're reveal. the 3.5 gigahertz versus the 3.66 base on the SeriesX and 3.8 gigahertz on the regular one with no multithreading.

Long story short at a raw comparison of the ps5 its just a flatout weaker console out of the two. The good news for Playstation fans is that sony will probably continue to have the better first party titles,

Give us one person that said the PS5 is more powerful than XSX. And Cerny never said the PS4 Pro could acheive 8.4 TFs of power. You don't know how to read properly. The PS5 can do SMT. Why do you think it can't?

 

[VanSiriusGabriel]

And if you mean Halo 5 yes it dose

Yea its pretty crazy how everyone is in denial, if sony and microsoft flipped positions there would be no debate which one is stronger. Cerny is the guy who said the PlayStation 4 Pro Could Achieve 8.4 Teraflops Of Computational Power lol.
Everyone is talking about how the ps5 SSD is faster that it can run at 5.5 gigabytes per seconds raw and the 9 gigabytes per second compressed but that only matters for first party titles, do you think game developers who are going to have everything on a cross platform are going to specialize it so that one runs faster on playstaion than it does on xbox?! No you need raw compute power and the only way the ps5 is getting its compute power for its 10.3 teroflops is because they're overclocking the crap out of it, thats not going to be a sustainable boost clock. 2.25 gigahertz are you kidding me on 36 CU versus the 52 CU that are on the SeriesX.
The fact that the CPU runs slower and has no simultaneous multithreading at least according what Sony said in they're reveal. the 3.5 gigahertz versus the 3.66 base on the SeriesX and 3.8 gigahertz on the regular one with no multithreading.

Long story short at a raw comparison of the ps5 its just a flatout weaker console out of the two. The good news for Playstation fans is that sony will probably continue to have the better first party titles,

PS5 has SMT, look at the psblog official specs.
Also Cerny didn't just say that Pro is 8TF, he stright up confirmed that 4 TF (in Pro context) aren't enough to reach 4K, which is true, and that PS4 Pro is a 4 TF system. He just specified that the system is 8 TF with different bit instructions, also true because it applies to every system as well.

And if you mean Halo 5 yes it dose

Yes, I meant Halo 5.

 

[ph33rknot]

PS5 has SMT, look at the psblog official specs.
Also Cerny didn't just say that Pro is 8TF, he stright up confirmed that 4 TF (in Pro context) aren't enough to reach 4K, which is true, and that PS4 Pro is a 4 TF system. He just specified that the system is 8 TF with different bit instructions, also true because it applies to every system as well.

Yes, I meant Halo 5.

It has been five years though

 

[Ascend]

But that's not how any of this works. You just made all of that up. lol

I expected a deeper answer than that.

 

[Moses85]

The REAL secret sauce. Sony's first party is easily worth a couple of teraflops , what say you GAF?



It'll happen.

Nintendo is thinking the same way

 

[mckmas8808]

I expected a deeper answer than that.

Deeper answer than what? You just made up the stuff that you said.

So how do we know Sony has more hardware to eliminate bottlenecks? Things get said so often that it turns in to some weird truths.

It'll be weird for MS to have the same amount of hardware processors in XSX and them NOT talk about it by now right? Especially considering they are talking more about the console than Sony has. With their strategy why would they hide this information?

 

[Ascend]

Deeper answer than what? You just made up the stuff that you said.

Made up? That's what the numbers say. The same numbers you use to calculate that 18%. I used the flops of all of them. I didn't make anything up.

I put the different perspectives there, because percentages can often be deceiving. Technically, Microsoft can say that the XSX is over 900% more powerful than the Xbox One. That obviously makes it sound a lot larger than it really is. My point is that the 18% (or whatever it really is) makes the XSX advantage sound smaller than it really is. Having an extra PS4 as additional graphical processing capability is not really a small thing, imo.

 

[thicc_girls_are_teh_best]

I'm not sure if % is the best way to go about it though. Percentage wise, the XSX is closer to the PS5 than the previous generation consoles are to each other. That is definitely true. However....

In terms of pure processing power, the difference is as big as the difference between the PS4 Pro and the Xbox One X.
Or to put it more dramatically, the difference between the XSX and the PS5 is pretty much 4 times as big as the difference between the PS4 and the original Xbox One.
Or putting it yet another way, the PS5 is about 5.5 times more powerful than the original PS4, while the XSX is over 9 times more powerful than the original Xbox One.
Or putting it yet another way again... The XSX is like having the PS5, and still having a full PS4 available for additional graphical processing power.

That's actually selling it short in a way. These are RDNA2 systems, not GCN, so even taking the lowest possible IPC gain of RDNA2 over RDNA1 (10%), it's closer to a PS4 if the PS4 were 2.525 TF GCN. IMO IPC gains over RDNA1 are probably about 15% - 20% at best (would hope so, that bodes well for Sony, MS and AMD); if you translate that it's closer to a 2.618 TF - 2.711 TF difference in GCN terms.

Combine that with improved understanding of GPGPU techniques (especially with 1st party titles like many of Sony's on PS4), and better engine scalability of various non-graphical elements like physics, AI, logic, etc., and IMO GPGPU is at least as much of a potential influence on game design for next get as the SSDs are, especially if Sony and MS can build tools to streamline utilization of it for 3rd-parties (might also depend on middleware tool support for leveraging such things too, such as Unreal Engine 4, Unity etc).

It's not as new to the console space as standardized SSDs are about to become, but it's going to offer some massive gains in game design next gen especially since the CPUs this time around are actually good!

Made up? That's what the numbers say. The same numbers you use to calculate that 18%. I used the flops of all of them. I didn't make anything up.

I put the different perspectives there, because percentages can often be deceiving. Technically, Microsoft can say that the XSX is over 900% more powerful than the Xbox One. That obviously makes it sound a lot larger than it really is. My point is that the 18% (or whatever it really is) makes it sound smaller than it really is. Having an extra PS4 as additional graphical processing capability is not really a small thing, imo.

Not graphical processing; think bigger. That extra GPU power has so many other uses besides just graphics, and this time it wouldn't be to "make up" for weak CPUs like it was on PS4 and XBO. It's use-cases outside of just prettier graphics are more interesting IMHO, especially if Sony and MS can help significantly streamline usage of it for 3rd-parties.

It's just that it's an area XSX stands to have a fixed hardware lead in, similar to how PS5 will always have a fixed hardware lead in the SSD area.

 

[VanSiriusGabriel]

It has been five years though

Woah, time goes fast wtf.

I expected a deeper answer than that.

To be fair, you're premise wasn't that deep either. Honestly asking: how is the difference 4x those of PS4 and One? Do you refer to the 40% ish difference in CUs? Xbox One has 12, PS4 20, which is almost double even if is just 8 CUs, proportion matters. Also One didn't have much higher clocks or much more faster and efficient SSD, it had 3x slower RAM also, including diminishing return the situation will ne much closer this time, al least until we really see all the systems secrets and how RT is used.
PS: also the difference is 5x vs 9x because One was shit.

 

[mckmas8808]

Made up? That's what the numbers say. The same numbers you use to calculate that 18%. I used the flops of all of them. I didn't make anything up.

I put the different perspectives there, because percentages can often be deceiving. Technically, Microsoft can say that the XSX is over 900% more powerful than the Xbox One. That obviously makes it sound a lot larger than it really is. My point is that the 18% (or whatever it really is) makes it sound smaller than it really is. Having an extra PS4 as additional graphical processing capability is not really a small thing, imo.

No you are disregarding the speed difference to make your point. The truth is you can't do that and it's being dishonest. Sony has decided to make a "quicker" GPU that's less powerful. The percentages aren't deceiving at all. An extra PS4 GPU in power is a "relatively" small thing when you consider the overall power of the XSX.

The best way to describe the difference in something is to your percentages. You ALWAYS should use percentages, otherwise, nominal numbers can be confusing if people don't understand the context. It's a lie to say the XSX is a PS4's worth of power better than the PS5, because the PS4 had a CPU, RAM, and a HDD in it to produce videogames. The XSX is a PS4's GPU worth of power stronger and that's it.

 

[IllegalLemon]

PS5 has SMT, look at the psblog official specs.
Also Cerny didn't just say that Pro is 8TF, he stright up confirmed that 4 TF (in Pro context) aren't enough to reach 4K, which is true, and that PS4 Pro is a 4 TF system. He just specified that the system is 8 TF with different bit instructions, also true because it applies to every system as well.

Yes, I meant Halo 5.

Cerny specified that it could reach that 8.4 TFLOP metric provided that devs utilised Amd's Rpm tech built into their own vega gpus which also had certain of its features built into the ps4 pro. Rpm or Rapid Packed Math in full, is a precision technique where the gpu can do two 16 bit floating point instructions (fp16) in the same time it would take to do one 32 bit floating point instruction (fp32) theoretically doubling float performance, but it appears very few people were paying attention to the ps4 pro stream when he said that and have fabricated their own assumptions. But the problem was that you had to code that float logic into your game to allow the gpu to run it as such and the games that did that were few and far between so, it was never fully utilised although it was a very versatile feature of the vega gpu architecture that Cerny had incorporated into the gpu silicon of the polaris based gpu present in the ps4 pro.

 

[Ascend]

Woah, time goes fast wtf.

To be fair, you're premise wasn't that deep either. Honestly asking: how is the difference 4x those of PS4 and One? Do you refer to the 40% ish difference in CUs? Xbox One has 12, PS4 20, which is about the same difference in proportion but One didn't have higher clocks or much more faster and efficient SSD, it had 3x slower RAM also, including diminishing return the situation will ne much closer this time, al least until we really see all the systems secrets and how RT is used.

Xbox One: 1.31TF
PS4: 1.84TF
PS4 Pro: 4.2 TFLOPS
Xbox One X: 6 TFLOPS
PS5: 10.3TF
XSX: 12.2TF

Difference between XO and PS4: 0.53TF
Difference between PS4 Pro and Xbox One X = 4x the difference between XO and PS4, rounded up
Difference between PS4 Pro and Xbox One X: 1.8TF -> Same difference between XSX and PS5, which is 1.9TF
PS5 10.3TF / PS4 1.84TF = 5.6x
XSX 12.2 TF / XO 1.31 TF = 9.3x


Didn't think I would need to explain, but, there you go.

No you are disregarding the speed difference to make your point. The truth is you can't do that and it's being dishonest. Sony has decided to make a "quicker" GPU that's less powerful. The percentages aren't deceiving at all. An extra PS4 in power is a "relatively" small thing when you consider the overall power of the XSX.

The best way to describe the difference in something is to your percentages. You ALWAYS should use percentages, otherwise, nominal numbers can be confusing if people don't understand the context.

The speed difference is accounted for in the TFLOPS. Effectively it will be a bit lower for the XSX, but not THAT much lower.

 

[VanSiriusGabriel]

Cerny specified that it could reach that 8.4 TFLOP metric provided that devs utilised Amd's Rpm tech built into their own vega gpus which also had certain of its features built into the ps4 pro. Rpm or Rapid Packed Math in full, is a precision technique where the gpu can do two 16 bit floating point instructions (fp16) in the same time it would take to do one 32 bit floating point instruction (fp32) theoretically doubling float performance, but it appears very few people were paying attention to the ps4 pro stream when he said that and have fabricated their own assumptions. But the problem was that you had to code that float logic into your game to allow the gpu to run it as such and the games that did that were few and far between so, it was never fully utilised although it was a very versatile feature of the vega gpu architecture that Cerny had incorporated into the gpu silicon of the polaris based gpu present in the ps4 pro.

This is what I remember too more or less, point is: Cerny wasn't wrong in saying that and it had the right to say it.
Thanks for the dive anyway.

 

[Kusarigama]

Made up? That's what the numbers say. The same numbers you use to calculate that 18%. I used the flops of all of them. I didn't make anything up.

I put the different perspectives there, because percentages can often be deceiving. Technically, Microsoft can say that the XSX is over 900% more powerful than the Xbox One. That obviously makes it sound a lot larger than it really is. My point is that the 18% (or whatever it really is) makes the XSX advantage sound smaller than it really is. Having an extra PS4 as additional graphical processing capability is not really a small thing, imo.

Having the flops performance score more by 1.8tflops doesnt equate to having "an extra PS4" of additional graphical prowess. Because it is only 1 metric that is more, the flops performance. The CPU, RAM, various other GPU component that don't contribute to the flops performance aren't there actually. So yeah, you are making stuff up.

 

[Evilms]

Nvidia unleashed the kraken, now RTXGI for everyone

 

[Ascend]

Having the flops performance score more by 1.8tflops doesnt equate to having "an extra PS4" of additional graphical prowess. Because it is only 1 metric that is more, the flops performance. The CPU, RAM, various other GPU component that don't contribute to the flops performance aren't there actually. So yeah, you are making stuff up.

I am quite aware of that. But those will be bottlenecks that prevent the GPU from reaching its max potential. They will not be boosting any GPU power. So it is still perfectly viable.

 

[VanSiriusGabriel]

Xbox One: 1.31TF
PS4: 1.84TF
PS4 Pro: 4.2 TFLOPS
Xbox One X: 6 TFLOPS
PS5: 10.3TF
XSX: 12.2TF

Difference between XO and PS4: 0.53TF
Difference between PS4 Pro and Xbox One X = 4x the difference between XO and PS4, rounded up
Difference between PS4 Pro and Xbox One X: 1.8TF -> Same difference between XSX and PS5, which is 1.9TF
PS5 10.3TF / PS4 1.84TF = 5.6x
XSX 12.2 TF / XO 1.31 TF = 9.3x


Didn't think I would need to explain, but, there you go.

Ok, and as you said this is just to be dramatic, right? Because as the other dude specified, % is always better. It's 4x only with linear math, proportionally the difference is closer because numbers are bigger. Meaning, yes, difference will be less because even if is 4x, PS5 is more powerful, meaning higher fidelity too, meaning it catches up with SeX, meaning it's 18%.
Having an extra PS4 means nothing, I could say it's like having one PS5 and 8 PS3 and like daaamn. Again: percentage.
Plus diminishing returns.

 

[Kusarigama]

I am quite aware of that. But those will be bottlenecks that prevent the GPU from reaching its max potential. They will not be boosting any GPU power. So it is still perfectly viable.

And just how many more time does it need to be stated that GPU power is not flops performance. It's a multi-dimensional entity. Flops performance is one part contributing to it and not the entire thing itself.

 

[ethomaz]

The Series X is more powerful than the PS5. There is no debating this. There is no way you get around a 44% shader deficit. The question is how big is the difference.

18% you meant...

40% was last gen.

 

[Ascend]

And just how many more time does it need to be stated that GPU power is not flops performance. It's a multi-dimensional entity. Flops performance is one part contributing to it and not the dimension itself.

GPU power is not flops performance, but the amount of flops is a direct indication of GPU power, especially within the same architecture. If anything, the performance differences are even larger now compared to previous gen, considering the instruction efficiency increase in RDNA over GCN.

I find it funny how FLOPS being expressed in percentages is perfectly fine. Express those same FLOPS directly as FLOPS instead, and suddenly it's not viable anymore. Boggles the mind...

 

[VanSiriusGabriel]

18% you meant...

40% was last gen.

He's likely talking about CUs difference.