Link? There is no l2 like "infinite" cache in ps5
Last edited:
-
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.
[Digital Foundry] Watch Dogs Legion: PlayStation 5 vs Xbox Series X|S - Graphics, Performance, Ray Tracing!
- Thread starter BluRayHiDef
- Start date
Link? There is no l2 like "infinite" cache in ps5
Bernd Lauert
Banned
Because it's a gimmick and a lot of people are turning it off, especially in multiplayer games.
Concern
Member
Idk if ms will emulate if its not in from day one. I expect most devs would probably just ignore it unless its easy to develop for.
Valued Gamer
Banned
Digital Foundry have already debunked that in real world testing.
Md Ray
Member
No. It's part of the GPU. When a specific data in the GPU caches are no longer needed, the cache scrubbers evict those unnecessary data in a more fine-grained manner instead of flushing the whole cache which could hurt GPU perf, in other words, this could tank the frame-rate.
Hence the addition of cache scrubbers so that frame-rate don't tank when stale data from the caches are removed.
sircaw
Banned
It's not a gimmick, Sony has designed it with many game titles in mind, just because you can turn it's feature set off does not make it a needless function.
You want 30 fps modes with all the eye candy or 60 fps with the smooth performance, these you can also be changed or turned of in a lot of games, are those gimmicks, no there are not.
Sony put millions if dollars into their controller research to give what they believe is a better next gen performance. Just because you don't like it does not mean millions of casual players our there wont.
The controller is a win-win, no amount of trying to rubbish it will change that.
This is to keep the gpu fed from the IO. I don't believe that it has anything to do with fillrate
Last edited:
sircaw
Banned
there have been literally hundreds of reviews saying how remarkable this controller is, its a no brainer.
Md Ray
Member
Nowhere did I mention fillrate in that post. Cache scrubbers help GPU perf. It is a part of the GPU, not a part of IO like you keep saying.
ethomaz
Banned
Never lol
I don’t know where you guys take these false stories.
What DF tried is to simulate a PS5 using a RDNA card where the performance doesn’t scale after 2000Mhz and so shows no gain in performance.
But hey RDNA 2 does scale performance up to 2500Mhz or even more.
I even necro a thread to say I was right about they doing that misleading benchmarks lol
Last edited:
whole argument is that the ps5 has the fillrate advantage despite having less bandwidth
Last edited:
MasterCornholio
Member
I think there's some confusion here. I was talking about the GPU performance in general and not just rasterization. I'm just trying to better understand why the PS5 GPU appears to be an efficient design.
Md Ray
Member
DF didn't debunk anything. They concluded by saying they'd need a GPU with more than 40 CUs and more games need to be tested for an accurate picture.
Physiognomonics
Member
Everything is compressed in RDNA GPUs, from L1, L2 to GDDR6 (which wasn't the case with GCN).
Navi vs. Turing: An Architecture Comparison
You've followed the rumors, waited for the reviews and finally slapped down your dollars and walked away with one of the latest graphics cards from AMD or...
www.techspot.com
Md Ray
Member
Again, It's part of the GPU. They exist in the caches of the GPU. It's you who need to watch it again.
Valued Gamer
Banned
No they didn't. They ran real world tests having one GPU with higher frequency and lower CU count vs another with lower frequency and higher CU count. The GPU with the higher CU count performed better.
It is what it is. Your theory isn't matched in real world applications.
Md Ray
Member
And yeah, the PS5 has 142.72 Gpix/s fillrate vs SX's 116 Gpix/s.
sendit
Member
You're sounding really desperate right now. It's kind of sad. To put this in perspective, even PC parts use boost clocks as well as their theoretical advertised Teraflop performance calculated based on those boost clocks.
RTX 3090
Base: 1400mhz
Boost: 1700mhz
Radeon 6800XT
Base: 1825mhz
Boost: 2250mhz
Last edited:
Md Ray
Member
Nope. Watch that vid again and then come back. Rich tested just one game. He then concluded by saying he'd need more than 40 CU GPU and "more testing points" to fully conclude that higher CU is better.
Higher frequency is better and the games on PS5 vs SX clearly show PS5 outperforming SX. It is what it is.
Last edited:
Yeah, this is getting overlooked and dismissed. And its designed so that it can be built on the exact same production line. Remove the drive, change the cover and you have the DE.
Amazing job by Cerny, Sony.
Road to ps5 showed that it is on the ssd talk. My bad
No mention that helps with rasterizarion though
These are the theoritical max
l2 compression could help improve rasterizarion in both consoles though, so it is still limired to memory bandwidth.
No mention that helps with rasterizarion though
These are the theoritical max
l2 compression could help improve rasterizarion in both consoles though, so it is still limired to memory bandwidth.
Bernd Lauert
Banned
I actually like the controller, I had a chance to try it a couple days ago. You asked why people are not taking it more into consideration, I gave you an answer. It's a nice to have addition but ultimately it's a gimmick. 30 vs 60 fps is a much, much bigger improvement.
Valued Gamer
Banned
It had nothing to do with overclocking a GPU.
They wanted to see if you got more performance with a higher clock on lower CUs than with lower clocks and more CU.
They had two GPUs with identical tflops, one with higher frequency and lower CUs, and one with lower clocks and higher CUs.
Same architecture, and none were overclocked.
The results were that the GPU with lower frequency and higher CUs had better performance. It was open and shut.
No need for you to bring RDNA 2 into it, as that had nothing to do with it.
The results are the results.
But maybe DF were shilling for MS here. I forgot about that.
Md Ray
Member
I didn't say it helps with rasterization though. You seem confused.
You keep saying "limited to memory bandwidth" but so far nearly every game is performing better on PS5. Shouldn't it be limited by bandwidth since it has -20% less bandwidth than SX?
Forget it. I was replying to your first post about fillrate/rasterization, which both are affected by memory bandwidth
I have explained a few posts back of what I believe the bottleneck is
Last edited:
Valued Gamer
Banned
Lol. So the results showed that lower frequency and higher CU count had better performance, and your take is that the opposite is true.
You are plain wrong. The facts are the facts.
Md Ray
Member
PS5 is outperforming XSX though.
The facts are the facts.
Armorian
Banned
They both drop resolution in the same places (to the same res of course) so pretty much exact parity in terms of performance. How, why? I don't know.
kyliethicc
Member
So once again see 10 PS5 tittyflops = 12 xsx tflops
PS5 most powerful per titflop
PS5 most powerful per titflop
MasterCornholio
Member
But the terafloops are the terafloops and then there's Craig's special sauce. So smooth, creamy and green.
ChuckeRearmed
Member
It feels like with each generation, Ubisoft optimizes games less and less
DForce
NaughtyDog Defense Force
Lets not forget that Richard recently said that the PlayStation 5 is punching above its weight, which suggest it's performing higher than he anticipated.
You can quote Richard and leave out an important part.
RoadHazard
Gold Member
Identical except for each seemingly having their own minor visual bugs. And out of those the worse AF on XSX is probably more noticeable.
Anyway, PS5 doing well again. Clearly both could go higher than 30fps, since they never drop a frame, but we don't know how much higher each could go. A 60fps mode without RT definitely seems possible.
Anyway, PS5 doing well again. Clearly both could go higher than 30fps, since they never drop a frame, but we don't know how much higher each could go. A 60fps mode without RT definitely seems possible.
Shmunter
Member
Yeah I was going to say, and the compression looks god awful too.
MasterCornholio
Member
Demo Souls is a good start though.
sendit
Member
No one is thinking that. However, the difference between the two consoles graphically in multi-platform titles will be minimal at best. Before these comparison videos started showing up, the narrative was the XSX would have a clear advantage in multi platform titles.
Last edited:
Leyasu
Banned
Both will have advantages and disadvantages.
The next gen Battlefield in 2021 will be interesting
sendit
Member
They're too close to have any serious graphical advantages. You're setting yourself up for disappointment if you think a multi-platform game such as Battlefield 2021 will change that.
Digital Gex
Banned
For a second I though this was about those "gaming outlets" using Vgchartz to hype up PS5 vs. Xbox One. The fact that isn't a prominent thread looking through two pages of the forum shows me Gaf has improved well.
phil_t98
#SonyToo
It’s like people miss out what they don’t want to see, he definently said that and that it is likely the puddles missing in PS5 was a bug to.
the funny thing is that people are talking about these first gen games and if they utlilising the full power and maxing out the consoles, nobody knows how much power is being used. It’s could only be using 50% of what’s available, we know games will get massivly better next bY this time next year visuals when the devs have a better handles on how to use the systems better
Humdinger
Member
Just watched the video. So, no differences. He sounded a little mystified why the XSX wasn't performing better than PS5.
Even though it's a draw, I think PS5 is the winner, since it is exceeding expectations ("punching above its weight," as they say). Or, another way of seeing it is XSX is failing to live up to expectations. That 18% TF on-paper advantage is seemingly being nullified by ... something.
Even though it's a draw, I think PS5 is the winner, since it is exceeding expectations ("punching above its weight," as they say). Or, another way of seeing it is XSX is failing to live up to expectations. That 18% TF on-paper advantage is seemingly being nullified by ... something.
ReBurn
Gold Member
Valued Gamer
Banned
RDNA has nothing to do with the test. That's just him saying it would be good to get more info on different cards
That in no way changes the fact that real world tests showed the opposite of what some here are claiming.
And as for Richard's comments, if you are agreeing with him you are literally saying the PS5 shouldn't be able to run ACV as good as it does. I think PS5 can easily run that, so I don't think the PS5 is doing things here that it technically shouldn't be able to.
Last edited:
DForce
NaughtyDog Defense Force
It does.
You're ignoring key factors because it doesn't support your case at all.
If there were no benefits, then Mark Cerny wouldn't have gone for such high clocks.
JackMcGunns
Member
Missing items in reflections, whether puddles or anything else indicates less objects were considered for the ray tracing pipeline. Since ray tracing can be fine tuned aggressively or less aggressive, this would be considered a less aggressive approach