ChiefDada
Gold Member
Sarcasm?
-
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.
PS5 Pro Specs Leak are Real, Releasing Holiday 2024(Insider Gaming)
- Thread starter jm89
- Start date
Sarcasm?
Gaiff
SBI’s Resident Gaslighter
Nope. RDNA2 was a dud. It should have been quite a bit better than it turned out.
It’s using TSMC instead of the shitty Samsung node used by Ampere but still didn’t manage to beat GA equivalent parts. The performance at higher resolution also just dies and the power efficiency isn’t where it was anticipated to be despite the more advanced node.
BattleScar
Member
I mean this is pretty silly. All the RDNA2 dies were narrower than the 'equivalent' Ampere dies, and we're pushed a bit further beyond their sweet spot to be competitive. And they still were more power efficient than Ampere.
They could have gone with wider designs if they wanted to, but they decided that wasn't cost effective enough for them, so they didn't.
The main area where RDNA2 disappointed was in ray tracing.
Gaiff
SBI’s Resident Gaslighter
Of course, they were more power efficient than Ampere since they were on a smaller and more advanced node. This isn’t an accomplishment. The performance per watt isn’t even that much better than Ampere either.
Look at how RDNA3 is getting demolished in power efficiency when NVIDIA got up to speed with TSMC.
Last edited:
RDNA 2 improved performance per watt by 50% on the same process as RDNA 1. It's probably one of the biggest architectural improvements in AMD history and wasn't matched by RDNA 3, which only slightly exceeds the listed boost for moving to N5.
The performance loss when moving from 1440p to 4K is in line with prior AMD and Nvidia architectures. It's just that Ampere does comparatively better, perhaps due to its increased compute.
Gaiff
SBI’s Resident Gaslighter
The performance isn’t fine when moving up in resolution and isn’t in line with other architectures. RDNA3 has no such problem and I don’t even think RDNA1 did. The 5700 XT vs the 2070’s performance differential remained consistent across different resolutions. The 6800 XT however loses more and more performance as you go up. This didn’t happen with prior generations of AMD GPUs nor is it happening now.
Polygonal_Sprite
Gold Member
Switching to Nvidia.
Alex's facial expressions whenever a PS5 Pro is brought up or that consoles might in the future use the magic of DLSS are absolutely delicious haha! He's a total PC fanboy which he struggles to control on that direct show. His posts from back in the day on Beyond 3D were pure "master race" tripe. He cannot handle that the gap between $500 consoles are $3000 PC's is going to be negligible to the average person in the coming years.
Gaiff
SBI’s Resident Gaslighter
Just because Alex says something stupid doesn't mean you have to try and one-up him.
Polygonal_Sprite
Gold Member
How so? PS6 is going to run all AAA games at their own reconstructed 4k/120fp or 8k/60fps with full RTGI, RTAO and RT Reflections. What will a more expensive PC do better outside of exclusives of course? No one cares about "real" resolution or framerate anymore because reconstruction along with A.I. has gotten so good so fast.
Gaiff
SBI’s Resident Gaslighter
And you took it even further. Bravo lmao.
Polygonal_Sprite
Gold Member
Explain why what I just said is wrong?...
Gaiff
SBI’s Resident Gaslighter
Dude, we're in a thread about a yet-to-be-released PS5 Pro and you're going off about a PS6. Literally everything you said is wrong unless you have a PS6 in handy and can show me what it can do.
Schmendrick
Member
The notion that somehow tech stands still outside the console realm is mindbogglingly stupid, especially with something like CP2077`s PT already available......
Last edited:
winjer
Gold Member
A console will never match the high end hardware released in the same generation.
Just look at the PS5, it has a cutdown Zen2 CPU, and a GPU with a mix of RDNA1 and RDNA2 features.
At the same time, PCs already had Zen3 CPUs and full RDNA2 GPUs doing more than twice the compute power. And Nvidia's Ampere was doing even better, regarding RT and AI.
Consoles do have many advantages, but peak performance is not one of them.
By the time the PS6 releases, there will be products from AMD, Nvidia and Intel, from the same generation, performing a lot better.
Polygonal_Sprite
Gold Member
Totally agree but already the gap is small between a PS5 and a 4080 equipped PC when it comes to most games to most people. The gap that is there is mostly down to more RT which is leveraged by specific hardware for reconstructed resolution and reconstructed RT effects. Sony will have similar tech inside PS5 Pro for reconstruction of resolution, RT and framerates nevermind PS6...
It's alarming to suggest that gap won't get even less noticeable to most people next gen once you add more RT specific hardware for reconstruction of resolution, RT, framerate and another huge leap in CPU power with more and faster RAM + storage.
winjer
Gold Member
Small gap between the PS56 and a 4080?
The 4080 is more than 3 times faster than a PS5 in rasterization. And in ray-tracing it's a massacre.
Gaiff
SBI’s Resident Gaslighter
The gap between a PS4 and PS5 is negligible to most people. What's your point here? That the average gamer doesn't care about graphics, performance, or resolution? Of course, they don't, which is why the most popular games run on PS2-era hardware.
With that said, the context of this thread and DF are enthusiasts or people who appreciate technology, not those who game on 1050 Ti-powered laptops or Switches.
Last edited:
Schmendrick
Member
That´s something only someone who doesn`t have the direct comparison can say...well, or someone like certain people around here who pretend to not even see the difference between 30 and 60fps....
The gulf of quality improvement for more money is wide and will stay wide. The amount of money needed to have a wide gulf however has increased drastically due to diminishing returns...and well Nvidia`s greed.
Last edited:
Gaiff
SBI’s Resident Gaslighter
The craziest part is it's the same guy who said this:
PS5 Pro has a huge increase in RT features compared to the PS5...but the gap between the RTX 4080 and the PS5 is small.
Polygonal_Sprite
Gold Member
The gap on a general level as in what you see on screen from a normal viewing distance to the average person is very small between PS5 and a 4080. I have a PS5 and have just bought a top of the line £2500 PC with a 4080. They provide a very similar experience in general terms outside of the two or three AAA games that don't have 60fps modes on console or ones that are a mess like Alan Wake 2 (the game I bought the PC for!).The craziest part is it's the same guy who said this:
PS5 Pro has a huge increase in RT features compared to the PS5...but the gap between the RTX 4080 and the PS5 is small.
Gaiff
SBI’s Resident Gaslighter
Then you literally won't be seeing a difference between the PS5 and PS5 Pro. Why would you buy a Pro?
Schmendrick
Member
You need glasses if the vasiline smeared atrocity that is the 60 fps mode IQ of most modern AAA games on console is "similar" to a high end PC´s image quality for you....
Last edited:
Bojji
Member
Maybe you are using 1080p screen.
Gap between 720p-1080p console games in performance modes and 4k (even with dlss) on pc is massive on high resolution screen.
Gaiff
SBI’s Resident Gaslighter
This was on April 5th. So you went ahead and bought an RTX 4080-powered PC anyway?
Last edited:
Here are the figures for the percentage performance loss when going from 1440 to 4K in the TPU 6900 XT review.
6900 XT 36.8%
6800 XT 37.9%
2080 Ti 40.5%
5700 XT 43.6%
3080 37.2%
3090 35.4%
So at the time of launch, the RDNA 2 parts were ahead of the prior generation and only beaten by Ampere.
AMD Radeon RX 6900 XT Review - The Biggest Big Navi
AMD's Radeon RX 6900 XT offers convincing 4K gaming performance, yet stays below 300 W. Thanks to this impressive efficiency, the card is almost whisper-quiet, quieter than any RTX 3090 we've ever tested. In our review, we not only benchmark the RX 6900 XT on Intel, but also on Zen 3, with fast...
Now a few years later in the TPU 4080 Super review, we have these results:
6800 XT 43%
6900 XT 42.7%
2080 Ti 41%
5700 XT 43.8%
7900 XTX 39.7%
7900 XT 42.1%
7800 XT 43%
3080 40%
3090 39.1%
4090 36.7%
4080 39.9%
4070 43%
NVIDIA GeForce RTX 4080 Super Founders Edition Review - Savings of $200
NVIDIA's new GeForce RTX 4080 Super introduces a noteworthy $200 price reduction compared to the non-Super 4080, placing significant pricing pressure on AMD's RX 7900 XTX. Despite this, the performance gains vs RTX 4080 non-Super are only marginal, we expected more.
So RDNA 2 has slipped behind Turing by a couple of percentage points. The 7900 XTX competes with the non-4090 Nvidia cards. However the 7900 XT is behind and the 7800 XT is exactly where the 6800 XT is! So if for you losing 2-3% extra percentage points means the performance is "dying", then OK. I guess we just disagree on language. But we should be clear that the mid-range RDNA 3 parts like the 7800 XT do not see great performance when moving to higher resolutions.
Last edited:
Gaiff
SBI’s Resident Gaslighter
Not an apples to apples comparison because the 7900 XTX is overall faster than the 4080 at pretty much any given resolution and the gap actually widens the higher you go.Here are the figures for the percentage performance loss when going from 1440 to 4K in the TPU 6900 XT review.
6900 XT 36.8%
6800 XT 37.9%
2080 Ti 40.5%
5700 XT 43.6%
3080 37.2%
3090 35.4%
So at the time of launch, the RDNA 2 parts were ahead of the prior generation and only beaten by Ampere.
AMD Radeon RX 6900 XT Review - The Biggest Big Navi
AMD's Radeon RX 6900 XT offers convincing 4K gaming performance, yet stays below 300 W. Thanks to this impressive efficiency, the card is almost whisper-quiet, quieter than any RTX 3090 we've ever tested. In our review, we not only benchmark the RX 6900 XT on Intel, but also on Zen 3, with fast...www.techpowerup.com
Now a few years later in the TPU 4080 Super review, we have these results:
6800 XT 43%
6900 XT 42.7%
2080 Ti 41%
5700 XT 43.8%
7900 XTX 39.7%
7900 XT 42.1%
7800 XT 43%
3080 40%
3090 39.1%
4090 36.7%
4080 39.9%
4070 43%
NVIDIA GeForce RTX 4080 Super Founders Edition Review - Savings of $200
NVIDIA's new GeForce RTX 4080 Super introduces a noteworthy $200 price reduction compared to the non-Super 4080, placing significant pricing pressure on AMD's RX 7900 XTX. Despite this, the performance gains vs RTX 4080 non-Super are only marginal, we expected more.
So RDNA 2 has slipped behind Turing by a couple of percentage points. The 7900 XTX competes with the non-4090 Nvidia cards. However the 7900 XT is behind and the 7800 XT is exactly where the 6800 XT is! So if for you losing 2-3% extra percentage points means the performance is "dying", then OK. I guess we just disagree on language. But we should be clear that the mid-range RDNA 3 parts like the 7800 XT do not see great performance when moving to higher resolutions.
Look at the 6800 XT vs 3080 for instance.
At 1080p, the 3080 loses by 3%.
At 1440p, it leads by 1%, effectively a tie.
But then at 4K, the 3080 is 7% faster.
I'd say going from losing by 3% to winning by 7% is a signifcant difference. The 6800 XT loses 10% going from 1080p to 4K compared to its GeForce counterpart. That's a huge difference and we've never seen this kind of scaling with any architecture besides with RDNA2. The performance differential cannot be chalked up to obvious advantages like bandwidth either. Now with RDNA3, the performance across resolutions tends to remain consistent vs Lovelace. But then again, you could argue that it's Ampere that scales poorly at lower resolutions due to games having difficulties keeping its cores fed but we see the same kind of scaling vs any other architecture. RDNA2 is the odd one out.
1.) 1080p is going to be heavily CPU limited on high end cards, so performance differences may simply be due to how well each platform handles this limitation.
2.) According to the TPU numbers, the gap between the 7900 XTX and 4080 is almost the same when moving from 1440p to 4K. But if the 7900 XTX has less of hit at 4K due to having more headroom at 1440p (like the 4090) then that indicates that RDNA 3 is worse at scaling then the figures indicate. And that would be borne out by the 7900 XT result.
3.) We know Ampere does comparatively better at higher resolutions. That point was established in the 3080/3090 reviews before RDNA 2 arrived. The comparison should include Turing and RDNA 1 to establish a baseline, since your claim was that RDNA 2 did not perform like prior architectures. And if you're claiming RDNA 3 is "better", RDNA 3 cards need to be compared too. (In my figures above the performance loss for the 7800 XT is the same as for the 6800 XT, and the 7900 XT only loses 1% less).
Last edited:
sigrad
Member
33 Tflops?
PS5 Pro Specs Leak are Real, Releasing Holiday 2024
The leaked PS5 Pro specs leaked earlier today are real and the PlayStation 5 is still tentatively targeting a 2025 holiday release.insider-gaming.com
Update: More Specs
EXCLUSIVE - More PlayStation 5 Pro Specs Detailed
Following this week's leaks on the PlayStation 5 Pro GPU specs and performance targets, Insider Gaming has learned more PS5 Pro specs.
Get your wallet ready to be emptied out. No fucking way this thing releases for anywhere under $600-$700.
jm89
Member
Gaiff
SBI’s Resident Gaslighter
Techpowerup is good for a general ballpark figure. If you wanna get into the nitty-gritty, they're generally inaccurate. There isn't that much difference in the scaling between the 7900 XTX and 4080.
Goes from 3% to 7% in favor of the 7900 XTX. That's not -3% to +7%. This is normal scaling.
Rocco Schiavone
Member
Don't get too caught up on the TFLOPS number on the normal rendering side it's around 1.5X better than PS5 . The compute is going to be a monster but straight forward rendering it will be a small jump.
On the bright side Ray-Tracing & ML will be well above average.
Price should be $499 - $649 and that's only because of the price hikes in chips , memory & storage.
Polygonal_Sprite
Gold Member
Yes I’ve been on the fence about waiting for a Pro or just jumping on a powerful PC. I cracked and got an Intel Core i7 13700k / 32GB RAM / 4080 PC. It’s in a lovely case too. If there’s something I can print off in my bios or something to prove this to you then I’ll gladly do it. I’ll probably end up getting a Pro too as the FFVII Rebirth performance mode is pretty blurry.
So far to get a locked 60fps on modern AAA games like Cyberpunk on my 4k OLED TV I have to use reconstruction techniques on resolution, framerate and rt. This is what PS5 Pro and PS6 will do to run their games at 4k/60fps with advanced RT. I don’t understand your (or anyone else’s) issue with this inevitability… The industry as a whole is all going in this direction.
The next step imo with be full procedurally made content created by A.I. in real time using cloud compute (PS7).
I believe PS5 Pro will show one of the biggest visual differences that we have seen in a long time in the console space.
Not because of the power but more to do with design philosophy.
Sony could have made PS5 Pro just a PS5 with more of the same & just enjoyed better resolution & performance but their approach will see a small boost to the normal rendering pipeline while having a big focus on Ray-Tracing & ML.
Instead of telling devs to get higher resolutions with the visuals we already have they're saying here is a lot of compute change the visuals with it.
Not because of the power but more to do with design philosophy.
Sony could have made PS5 Pro just a PS5 with more of the same & just enjoyed better resolution & performance but their approach will see a small boost to the normal rendering pipeline while having a big focus on Ray-Tracing & ML.
Instead of telling devs to get higher resolutions with the visuals we already have they're saying here is a lot of compute change the visuals with it.
The TPU numbers already indicate that the 4080 scales similarly to the 7900 XTX. And we already know that Ampere scales better than RDNA 2! So these numbers don't tell us anything new.
The question to be addressed is whether RDNA 2 scaled worse than prior generations, and how other RDNA 3 parts compare. So if you don't like TPU, we can use the TechSpot 6900 XT review (Steve from HUB). We have these figures for the performance loss when going from 1440p to 4K
6900 XT 39.6%
6800 XT 40.8%
3080 35.6%
3090 34.5%
2080 Ti 40.8%
5700 XT 46.8%
AMD Radeon RX 6900 XT Review
The RX 6900 XT is AMD's most expensive Radeon product ever to feature a single GPU at $1,000. On paper it should be only slightly faster than...
www.techspot.com
So compared to the TPU numbers, the 3080 and 2080 Ti do a little better comparatively and the 5700 XT does a little worse. But RDNA 2 still matches Turing and exceeds RDNA 1. If you now think the TechSpot numbers aren't representative, lets go to the 6900 XT Meta Review:
Here we see that RDNA 2 increases its advantage over both Turing and Pascal at every resolution.
Now, lets go to the TechSpot 4080 Super review for the most up to date benches. Here Turing and RDNA 1 have been dropped. We get these results:
7900 XTX 38%
7900 XT 40.9%
6950 XT 42.7%
6800 XT 41.5%
7800 XT 42.4%
4090 34%
4080 41.8%
3080 41.7%
Nvidia GeForce RTX 4080 Super Review
One would assume that Nvidia's "new" GeForce RTX 4080 Super is a special version of the original. But in reality, it's essentially the RTX 4080 with a...
www.techspot.com
Now maybe these benchmarks aren't representative and AMD does do much better here than the TechSpot 4080 benchmarks from over a year prior (also the TPU benches). But it is kind of funny to see the 6800 XT almost matching the 3080 at 4K. I guess it could be due a memory limitation. Even so though, RDNA 2 is doing fine when compared to the mid range RDNA 3 parts.
Last edited:
PUNKem733
Member
Why do they know more then us? Also there are hundreds of active members, they know more then ALL of US?! You do know that there are devs on here, right? You know people who work to make games on these systems. LOL But they know more.
Leonidas
AMD's Dogma: ARyzen (No Intel inside)
HeisenbergFX4
Member
I know for a fact there are people who visit these forums (would never post especially with their true identity or they would be hounded mercilessly) that know WAY more than those guys at DF
In fact a couple of people that visit these forums have actually fed DF some of their info leading into this current generation
Gaiff
SBI’s Resident Gaslighter
Nah, a couple of people on this very board know more than DF lol. Those who don’t are at least more logical.
Rocco Schiavone
Member
Better bigger than smaller. Especially for heat dissipation and noise levels. The slimmer PS5 is already shit because It's the same amount of heat packed in a smaller case with a fan that needs to spin faster.
Last edited:
Gaiff
SBI’s Resident Gaslighter
What do those numbers even refer to?
AMD Radeon RX 6900 XT Review
The RX 6900 XT is AMD's most expensive Radeon product ever to feature a single GPU at $1,000. On paper it should be only slightly faster than...
So compared to the TPU numbers, the 3080 and 2080 Ti do a little better comparatively and the 5700 XT does a little worse. But RDNA 2 still matches Turing and exceeds RDNA 1. If you now think the TechSpot numbers aren't representative, lets go to the 6900 XT Meta Review:
2080 Ti and 5700 XT are two different classes of GPUs. You should compare the 5700 XT to the 2070S, its direct competitor. And the Techspot article has the 6900 XT leading by 5% at 1080p, 2.4% at 1440p, and then losing by 6% at 4K. It lost 11% going from 1080p to 4K vs the 3090.
Here we see that RDNA 2 increases its advantage over both Turing and Pascal at every resolution.
Now, lets go to the TechSpot 4080 Super review for the most up to date benches. Here Turing and RDNA 1 have been dropped. We get these results:
7900 XTX 38%
7900 XT 40.9%
6950 XT 42.7%
6800 XT 41.5%
7800 XT 42.4%
4090 34%
4080 41.8%
3080 41.7%
Nvidia GeForce RTX 4080 Super Review
One would assume that Nvidia's "new" GeForce RTX 4080 Super is a special version of the original. But in reality, it's essentially the RTX 4080 with a...
Now maybe these benchmarks aren't representative and AMD does do much better here than the TechSpot 4080 benchmarks from over a year prior (also the TPU benches). But it is kind of funny to see the 6800 XT almost matching the 3080 at 4K. I guess it could be due a memory limitation. Even so though, RDNA 2 is doing fine when compared to the mid range RDNA 3 parts.
Once again, what are these numbers supposed to be? I'm not sure what they're supposed to represent.
Last edited:
HeisenbergFX4
Member
It's the percentage performance loss moving from 1440p to 4K. So if it's 0% then you lose no performance, while if it's 33.3%, you lose a third of the performance etc.
You originally said that the performance "dies" at higher resolutions with RDNA 2 and that this "isn’t in line with other architectures". My point is that RDNA 2 scales similarly to RDNA 1 and Turing as per the benchmark data I presented. So it's not as if RDNA 2 went backwards in any way. Rather Ampere improved Nvidia's performance scaling vs Turing and this put AMD in a worse competitive situation. So I am not trying to deny that AMD scaling vs. Ampere was worse than RDNA 1's scaling vs. Turing. I am saying this was not a failure of RDNA 2, unless you think every advancement your competitor makes that you don't is a "failure".
And with RDNA 3, based on the benchmark data, we seem to have a situation where the 7900 XTX achieves scaling parity with Ada/Ampere, the 7900 XT is between Ada/Ampere and RDNA 2, and the 7800 XT is in line with RDNA 2. (Check out the below meta review where we see the 6800 XT hold its ground against the 7800 XT)
)
The comparison is not between the 2080 Ti and the 5700 XT but between RDNA 2 and RDNA 1/Turing. Now it's fair to object that RDNA 1 was only used in the 5700 XT which is midrange card, so not designed for 4K. But at 1080p you will be CPU limited on RDNA 2, so that's not ideal either. If we go back to the Radeon VII which has a lot more bandwidth, we still see it behind RDNA 2 in terms of scaling.
Yes, because Ampere scales better than RDNA 2 and Turing.
Last edited:
ChorizoPicozo
Member
i mean....the meme is going to become real
PS5Pro leaked image No CLICKBAIT!!!! share and subscribe
