Support NeoGAF

[spyshagg]

The big two measurements me and most technical people were waiting for were the watts consumption VS performance on screen (perf per watt or FPS per watt).


Being two products almost identical to each other (navi2/zen2/gddr6), both built on the same silicon process together with excellent cooling solutions, it is acceptable to assume by default that the higher power consuming product will generally have the better performance on screen (FPS).


The first metric we got was a few weeks ago: The wattage consumption. Both consoles pull 200watts from the wall. Bizarrely Identical.
This suggested 2 weeks ago performance on screen would be pretty close. But there were doubts, specifically: The on-paper specs suggesting XSX was always going to be faster. The implication being that the high clock speed of the PS5 pushing it away from the ideal voltage/Mhz curve


The story so far after the first game face-offs:


- It would appear that the wattage metric is accurately predicting the low discrepancy between consoles.

- The different designs are negating the on-paper difference of ~20%

- There will be different bottlenecks in different areas of the game caused by the different designs.

- Operating silicon temperatures and Power Supply efficiencies are still relevant factors, and may give or take a few % to each other. Unknown quantity.



Cheers

 

[yurqqa]

Who is inerested in this ? Nvidia ? Greta ?

They are not handhelds after all and both have good and quiet cooling solutions.

 

[spyshagg]

Who is inerested in this ? Nvidia ? Greta ?

They are not handhelds after all and both have good and quiet cooling solutions.

Sorry, I don't understand your point. Watts is not just a measure of how much battery your device has or how loud a console can get. When paired with performance figures, you have an actual measurement of work-per-watt that transcends what the tflops figures suggest.


The point being is that people are surprised about PS5 performance, but when both PS5 and XSX were measured to consume ~200watts, it was the writing in the wall that both would perform close to each other in games.

People were comparing chip sizes and TFlops figures, but they both are doing the same work per watt in AC: Valhalla and close in DMC:5. The question is how is PS5 doing it exactly.

 

[yurqqa]

when both PS5 and XSX were measured to consume ~200watts, it was the writing in the wall that both would perform close to each other in games.

I'd say they are both equally good heaters.

But since they apparently have different ways for some new GPU features and SSD difference - we'll see.
I don't think they translate to wattage like this.

 

[Andodalf]

I'd say they are both equally good heaters.

But since they apparently have different ways for some new GPU features and SSD difference - we'll see.
I don't think they translate to wattage like this.

It's not exact, PS5 having higher clocks should be a touch less efficient, but in general the same Arch on the same Node will give very similar perf with the same wattage.

 

[jaysius]

Yea, these 2 are evenly matched. Only console warriors are left to argue over the scraps of performance differences.

Sane gamers see this as a POSITIVE because neither main platform will hold the other back.

Maybe this gen we'll see some great titles that take full advantage of these TWO machine's strengths.

The console wars can be over if you want it. The end results are more boring than you thought, right now it's coming up VERY CLOSE to a TIE.

We probably won't see an emergence until LATER in this cycle.

 

[FritzJ92]

Sorry, I don't understand your point. Watts is not just a measure of how much battery your device has or how loud a console can get. When paired with performance figures, you have an actual measurement of work-per-watt that transcends what the tflops figures suggest.


The point being is that people are surprised about PS5 performance, but when both PS5 and XSX were measured to consume ~200watts, it was the writing in the wall that both would perform close to each other in games.

People were comparing chip sizes and TFlops figures, but they both are doing the same work per watt in AC: Valhalla and close in DMC:5. The question is how is PS5 doing it exactly.

The only variable i can think of is inefficiencies... Dev engine compatibility with DX12u vs whatever PS5 uses (Vulkan?)...

API and devkits also affects this... and inefficient code can pull the same watt and give less performance on the same chip... so wattage just wont tell the whole story.