Discussing the power differences of the PS5 vs Xbox Series X, and which console will you get?

This is too much fun to watch so, enjoy your day off. :)

Straight up. SW is boring lately, and he adds some levity.

@Pedro: Coolio😌

@davillain-: I wouldn't dare!

Hey you let casualbanndodger4ever stay then you should let cash angry stay, at least he's entertaining.

@i_p_daily: Oh I'm not going to do anything. Looking at the thread at hand, it seems the situation has been handle and by that, BoxRekt dug himself deep he'll never recover :P

Still PS5, just not at launch. I can still enjoy Xbox content on PC. I can pass on PS5 hardware should Sony decide to go the same route.

I still don't know which one i'll get, but those who think PS5 is inferior to Series X because it has less teraflops, are making the same mistake as those who thought PS3 would destroy the 360 because of the Cell. The whole scenario is way more complex, and interesting, than that. Check this out:

https://m.youtube.com/watch?v=l1aHy8uBg2o

Think I'll get both. Series X looks really appealing with better supported BC (speaking as someone who ignored Xbox One entirely), Gamepass, a first party lineup and unlimited powah.

They've really got their shit together after a generation of meh.

Sony, well they have a nice SSD I suppose. But there's plenty of time for everything to swing back in their favour.

The difference between the Cell and the 360 debacle is that these consoles are using the same tech and the same manufacturer. So, its not remotely the same situation.

@Pedro: Not in the hardware, but in how people took the infornation (and that was Sony's fault, by showing the world a conference that should have been aimed exclusively to devs).

Just like what happened with the cell, people took the teraflop numbers as a guarantee of superiority, and ignored the fact that a console is the sum of all its components and how they interact with each other.

Series X is factually a ridiculously powerful console, but PS5 does seem to have its advantages.

PS5 no doubt for me.....so exciting to see what it can do and developers all over Twitter wetting themselves in excitement over it and what that SSD will allow them do.

One next generation machine has got fan boys excited about 2 tflops and one has developers excited about the new possibilities.

That should be even easier to calculate.

In fact i posted an example of the 5700XT and 5700 which have 1.8TF difference in power and the gap was pretty small.

I think the xbox series X will always always be ahead more power on the same architecture means the one with more power will always win.

But the dude in the video has a point so does Cerny,AMD has problem with GPU with a great number of CU,and is clear when you Over clock Vega 56 which can beat Vega 64.

Vega 64 is 64 CU 12,65TF

Vega 56 is 56 CU 10.54TF

Not saying this will be the case here,but there is more than 2TF gap between this 2 GPU.

But i firmly believe the xbox should come on top always fallowing the same like of the PS4 and Pro examples.

In certain situations PS5 will be faster.... Higher GPU clock increases fill rate, texel rate, internal GPU bandwidth, geometry rates, reduces latency............. Increasing shader count doesn't do any of that so if your ALU performance is limited by any of the back end functions such as the ones about then they won't do shit.

Vega 64 has 14% more shader cores then Vega 56.........but run both of them at the same clock speed and Vega 64 isn't 14% faster......... it's less then 1% faster.

If you do that for the 290 vs 290x......Fury vs Fury X..etc...etc... you'll see that AMD GPU's have shit scaling with CU count as they're limited by other areas of the GPU which is why they respond better to frequency increases.

Cerny even said this in the presentation.

AMD RDNA uses WGP which is not the same as the traditional CU. So, using the older architecture to leverage understanding of the newer is not accurate. People need to stop damage controlling the reality between the two systems and move on. All of this dancing around and speculation to make up for the blatant differences in performance is just silly.

As I have stated many times before and now, the difference is most likely going to be negligible. I could be wrong (I didn't think the Series X would be 12 TFLOPS ¯\_(ツ)_/¯ ). We will all see at the end of the year.

Who damage controlling? And CU has nothing to do with it, the example was to show that having extra shaders doesn't translate to better performance if you don't have the back end throughput to support it.

Recently bought PC gaming rig, so PS5 easily.

Does anybody else remember how just a few days ago, Cows were saying "All Mark Cerny has to do is come on stage and.......etc." and bragging about the Instagram impressions of the Playstation 5 logo?

Pepperidge farm remembers.

Ah the good ole days. When we were all the tender young age of........however old you were Tuesday lmao.

You crawled out of your hole for that damage control LOL.

Go back in and come back with something better.

😉

@tormentos: Still trying to claim the PS5 is faster I see. Poor thing. Have fun with the power draw and temps.😉

@tormentos:

NAVI 10 doesn't have variable-rate shading, mesh shaders, sample feedback, and hardware raytracing.

• Sample Feedback, the shading for a 3D surface is recycled for the next frame, hence conserving shading resource power. It's better than the checkerboard rendering.

• Variable-Rate Shading, change shading resolution while preserving geometry resolution, hence conserving shading resource power. It's better than the checkerboard rendering.

• Mesh Shaders, enables efficient geometry amplification, hence conserving geometry related resource power. It's unknown if AMD will enable Mesh shaders for NAVI 10.

• RDNA 2 supports DXR Tier 1.1.

Existing games don't use RDNA 2's mentioned features.

For TU104, RTX 2080 class would be the boundary between render back end and 11.66-12 FLOPS compute balance.

MSI RTX 2080 Super Gaming X Trio has 12.15 TFLOPS average with 496 GB/s memory bandwidth.

For TU102, NVIDIA switches to 88 ROPS to 96 ROPS with 14 to 17 TFLOPS compute range.

Your argument would be valid but 2-week raw Gears 5 benchmark port with PC Ultra settings shows XSX's RDNA v2 12 TFLOPS scaling into RTX 2080 class results

Don't assume RDNA v2 is like RDNA v1 or Vega GCN.

@jahnee:

From https://www.techpowerup.com/gpu-specs/xbox-series-x-gpu.c3482

XSX GPU is rumored to have 80 ROPS.

PS5 GPU would have 64 ROPS based on NAVI 10's 40 CU layout.

@scatteh316: I'm excited about sending pictures of my ears in....

https://www.eurogamer.net/articles/digitalfoundry-2020-inside-xbox-series-x-full-specs

"Our second component is a high-speed hardware decompression block that can deliver over 6GB/s," reveals Andrew Goossen. "This is a dedicated silicon block that offloads decompression work from the CPU and is matched to the SSD so that decompression is never a bottleneck. The decompression hardware supports Zlib for general data and a new compression [system] called BCPack that is tailored to the GPU textures that typically comprise the vast majority of a game's package size."

Microsoft increased decompressed storage I/O into "over 6GB/s" PR.

@hrt_rulz01: He's not wrong though.

Well no one knows what the games on either system will look like. So, at the moment its just an assumption. (Well if Hellblade 2 is legit demonstration then his point doesn't hold much value)

@vfighter:

He is

HAH

It’s easy this seeing into the future stuff

@vfighter: If you say so.

@tormentos: I am saving this quote from you.

You heard it from Tormentos folks: “The difference in power will only result in 5 frames”

You are in such denial.

I think eventually I’ll get the xbox mainly for its xbox backwards compatibility through all consoles. Xbox have been the only systems I can’t emulate well on pc. I have pc and Switch for all other gaming needs. If the price comes down to about $300 eventually I’ll get one.

Read https://gpuopen.com/gaming-product/true-audio-next/

AMD's "True Audio Next" which includes audio raytracing via RadeonRays which runs on Polaris/Vega/RDNA v1.

Sony asked AMD to slice at least 1 CU without raster hardware for DSP audio work.

RDNA v2 GPUs have RT cores to accelerate RadeonRays.

PS5 technically has 36 RDNA v2 CU for graphics + at least 1 CU for AMD's True Audio Next. Total CU count: 37.

Sony asked CU based DSP to be equal to 8 Jaguar CPU cores from PS4 which is about 100 GFLOPS.

For more information.

XSX's raytracing audio from https://www.windowscentral.com/xbox-series-x-project-acoustics-audio-ray-tracing-revealed

Both Sony and MS follow AMD's True-Audio Next.

@ronvalencia: But will this mean that the XsX will need to use GPU CU for audio? While the PS5 has a dedicated CU.

The PS4 and X1 had a AMD TrueAudio processor for sound, but it could only do around 200 audio channels, which is less than the Cell on the PS3. So if the XsX only has a TrueAudio v2, it will be inferior to the PS5 I would guess.

Well if you are a blind lemmings who can't read yes that is what i am doing,if you can see past that and actually see that i already claim that the xbox series X would always be ahead then no.

So take it how it serve you best.

I just showed you that even on RDNA that still the case,just not as bad but again the series X has more power more bandwidth alto mix,so yeah i know it will have always an advantage,now i think it will be ultimately small and smaller than this gen.

And?

Aren't you the guy who defended the xbox one 16ROP at broken back this gen over the PS4 doubling that with 32ROP?

So now more is better but this gen less was ok for the xbox one.😂

Watching you flip flop is quite something.

The xbox series has more powerful so is ok it has more ROPS.

I just counting until you come and say 64ROPS is not enough for 4k,after all you did defending 16ROP been enough for 1080p.

Well it would not be the first shit you quote out of context don't worry we knot how severely limited your brain is.

@tormentos: I understand that you are stressed and you are in super damage control mode. I am sorry Sony opted for a weaker GPU,CPU and Memory. You can always file a complaint to Sony explaining how you have been slighted and now have to work overtime to defend the PS5. Maybe they will up the frequency even more, just for you. 😂

@tormentos: iirc the ROP count for both consoles remains the same as there really aren't indicator (ROP amount is decoupled from the rest of the other hardware blocks in AMD GPUs), the only reason for ROP to be worth it over 64 would be to push 8K gaming assuming other aspects of GPU is also expanded upon (such increased memory, geometry, etc blah blah) proportionally.

I already did complain with sony.🤷‍♂️

I guess they don't care. :(

1. RDNA generation has ROPS with L2 cache links which are decoupled from the memory controllers e.g. RX 5600 XT has 36 CU with 192 bit bus and still has 64 ROPS. LOL.

X1X GPU's ROPS with 2MB render cache enables decoupled 384-bit memory bus i.e. it's not wedged like Polaris 10's ROPS with 256bit bus.

ROPS with multi-MB-cache can run to its near potential via tile cache render tricks.

Unlike NVIDIA's ROPS linked with fast L2 cache, XBO's 32 ESRAM bandwidth is close to 256bit GDDR5-5500 bandwidth, hence it's pointless.

2. Do the math on 16 ROPS at 853 Mhz read/write memory bandwidth consumption and get back to me.

PS; MS's argument is correct on 16 ROPS at 853 Mhz but it wouldn't solve XBO's CU bound issues.

@-DaVillain- seems like your have to mod this thread after all. 😋