Lionel Richie
Member
Suck my balls man, Control is a great game. Please shitpost elsewhere.
DF: Control - PS5/Pro/Series X Update 1.30 - PSSR Upscaling and RT at 60FPS Tested on PS5 Pro!
- Thread starter Zuzu
- Start date
Suck my balls man, Control is a great game. Please shitpost elsewhere.
adamsapple
Or is it just one of Phil's balls in my throat?
There is no excuse for base but that's not what I'm saying or comparing. I'm saying 40fps with a highly consistent frametime is more "smooth" than a highly variable frametime that VRR (and especially LFC) compensates for. Take the previous photomode unlocked framerate, that with VRR would not look as smooth when you have 33ms frametime spikes between 16ms ones. It's presenting itself at 50fps+ but it is not a smooth experience.
In a nutshell read this:
Why VRR is not a magic bullet for fixing poor performance
Variable refresh rate is game-changer - but it has it limits and doesn't automatically fix performance problems. Digital Foundry reports.
www.eurogamer.net
VRR is a remedy for inconsistent frametimes with caveats but it isn't "smoother" than a highly consistent frametime.
Last edited:
Diego Brasil
Member
Guys, there are games that run better on the PS5 in terms of performance.
There are games that run better on the Xbox Series X.
This difference doesn't apply to all games.
In fact, for the vast majority of games, both have the same performance.
Here are some games that run better on the PS5... normal... (others run better on the Xbox... normal)
Call of Duty: Black Ops 6
Elden Ring: Shadow of the Erdtree
ghostrunner
Baldur's Gate 3
The First Descendant
Dead Rising Deluxe Remaster
Atomic heart
There are games that run better on the Xbox Series X.
This difference doesn't apply to all games.
In fact, for the vast majority of games, both have the same performance.
Here are some games that run better on the PS5... normal... (others run better on the Xbox... normal)
Call of Duty: Black Ops 6
Elden Ring: Shadow of the Erdtree
ghostrunner
Baldur's Gate 3
The First Descendant
Dead Rising Deluxe Remaster
Atomic heart
Last edited:
Diego Brasil
Member
The difference between the PSSR and the TAA in motion is huge.
Look at the outline of the table.
Both images are screenshots of a moving scene.
Look at the outline of the table.
Both images are screenshots of a moving scene.
Last edited:
adamsapple
Or is it just one of Phil's balls in my throat?
It's not smoother than something that's a flat consistent line, ofc, but it helps a lot keeping the games feel smooth and remove the judder that comes from fixed refresh displays.
Honestly, it's probably the best native feature this console gen. Pair it with 120hz and low frame-rate compensation and it's even more effective.
omegasc
Member
That, and SSDs. This game takes 6 or 7 seconds to load when you die or fast travel and I'm already thinking it's too much!
How long were they on PS4/XOne!?
adamsapple
Or is it just one of Phil's balls in my throat?
That, and SSDs. This game takes 6 or 7 seconds to load when you die or fast travel and I'm already thinking it's too much!
How long were they on PS4/XOne!?
Been ages but from what I remember playing, stock PS4 HDD was like 30+ seconds or so.
omegasc
Member
Just looked on YT: 55s to main menu, 33s to load "new game" on PS4 Pro. Down to 7s and 6s respectively on the base PS5.
SKYF@ll
Member
Thanks for your capture, AdamYep, can confirm first hand it's been fixed now. Contrast difference from my direct capture aside.
, you're always so fast.It looks almost identical to the reflection on the PS5 Pro.
I've also included a zoomed comparison shot to show the difference in image quality between PSSR and TAA, which everyone is curious about.
PSSR(PS5 Pro) vs. TAA(XSX)
PSSR(PS5 Pro) vs. TAA(XSX)
PS5 Pro Quality Mode(RT) unlocked frame rate / VRR 120Hz screen shot *The file size was too large so it was compressed into JPEG format.
Zathalus
Member
GPUs before Turing could do RT as well. There is a difference between doing something in software and having hardware support for it. Even if the Cell or PS2 Vector units were good at some forms of sparsity exploitation.
PS3 can do ML with sparsity...
PS2... can compute ML with sparsity.
Out of all the SIMD compute accelerators in consoles released for last 25 years, more of them do sparsity before NVidia even 'invented' ML than after.
Last edited:
You have the video to know what is said in it..... And yes, not a single frame below 40fps in areas where PS5 drops to low 30s and Pro to 40s.
You are saying that I have expressly said that the difference with Pro is 5% in that quality mode and I have said that given the references between PS5 base and XSX in that quality mode and the data given in performance mode, it is PROBABLY much closer to what is seen in performance mode than the 50% that you defend. It is a matter of reading that paragraph and deciding who is distorting it
Because what the fos data draws is that XSX works comfortably in performance in both modes. The performance mode says that XSX is far above PS4 and almost on par with PRo and the quality mode draws the same image between PS5 base versus XSX. 2 + 2. Enough to be able to conclude that there is no technical reason for that 60fps mode not to exist unlocked in XSX as in PRO.
Lol. Then explain why you have areas where XSX runs at a locked 40fps (which means it's above those 40fps) and PRo is at 40fps. It's funny that according to you, the only meaningful reference is the one you want, and you also ignore the results in performance mode.
Then you're the one who started arguing that the lack of an unlocked 60fps mode is due to performance limitations, and not equipping XSX with that mode is logical when the fps data and VRR features on XS match perfectly.
I mean, there's no technical reason for XSX not to have that unlocked 60fps quality mode, which is always a better option than a locked 40fps, even more so when the resolution is fixed and DRS doesn't kick in. Exactly the same case as Wukong, although in this case, without being an official mode by GameSciencie, you can force that mode via XSX OS.
Last edited:
Duchess
Member
Fingers crossed that the FSR4 Pro update next year makes things even better.
It would be great if Sony were able to have it as a drop-in replacement for the existing PSSR tech, so devs don't need to rewrote existing titles.
What has your stupidly incorrect statement saying "not a single fps below 40fps" got to do with PS5? It drops below 40 and you can compare with fps caps gone.
I don't even know wtf you're saying anymore but I showed you with evidence that when the cap doesn't apply the game is running 43% faster than the XSX on the PS5 Pro. It's the reason why the cap is 60fps on PS5 Pro and 40fps on XSX. It is NOT "closer to the performance mode" which is ~5% on average better on Pro.You are saying that I have expressly said that the difference with Pro is 5% in that quality mode and I have said that given the references between PS5 base and XSX in that quality mode and the data given in performance mode, it is PROBABLY much closer to what is seen in performance mode than the 50% that you defend. It is a matter of reading that paragraph and deciding who is distorting it
Because depending on the scene there are different bottlenecks. I didn't ignore the results in performance mode at all. When it's running at high fps like 90fps it can run into other bottlenecks that are not GPU but in the graphics mode (with RT) it is mostly GPU limited where the XSX is much weaker.Lol. Then explain why you have areas where XSX runs at a locked 40fps (which means it's above those 40fps) and PRo is at 40fps. It's funny that according to you, the only meaningful reference is the one you want, and you also ignore the results in performance mode.
Because it is.
What? If you say so bud but again it doesn't and LFC has to kick in when it isn't.
no 46-50fps with 33ms frametime spikes is not "always better than 40fps". They did a good job of keeping a consistent frametime of 25ms on their engine with no sudden jumps. 16ms for 20 frames then 33ms for the other 20 to get a 45fps+ average or whatever is not a smooth experience even with the existence of VRR. That's poor frametime spikes and add to that there are the occasional 37fps drops and you're out of the range too. Their consistent framepaced 40fps is a good choice given the hardware cannot guarantee <25ms at all times with RT due to the hardware limitations on XSX.
Last edited:
Panajev2001a
GAF's Pleasant Genius
Ray reconstruction and neural radiance cache (plus FSR4 of course) could be a big boost for PS5 Pro (and PS6).
Still think that mid generation upgrades are less and less useful as we go, but it could help PS6's development and it is helping AMD as a competitor. I do not disagree with those observing that the architectures where Sony was involved in ended up being better than the others from PS4 onwards (RDNA2, RDNA4, RDNA5, etc…), just saying they have a good partnership more than just Sony picking up something from a laundry list of components from your roadmap and calling it a day.
It benefits AMD having a platform provider, with a good HW background too, that has a very strong stable of developers making their custom engines as well as using popular third party ones and the console R&D team is lead by someone with hands on knowledge of the entire game development pipeline.
winjer
Member
Thanks for your capture, Adam
It looks almost identical to the reflection on the PS5 Pro.
I've also included a zoomed comparison shot to show the difference in image quality between PSSR and TAA, which everyone is curious about.
PSSR(PS5 Pro) vs. TAA(XSX)
PSSR(PS5 Pro) vs. TAA(XSX)
PS5 Pro Quality Mode(RT) unlocked frame rate / VRR 120Hz screen shot *The file size was too large so it was compressed into JPEG format.
For curiosity sake, I took a picture of the game on the same area with FSR4. Max settings, RT to medium. 1440p upscaled to 4k.
If we talk about stupid things said in this thread the only in this category is someone denying the possibility and viability of an unlocked 60fps mode on XSX despite the data and evidence.....
PS5? Well, it has to do with the fact that it's also reference benchmark in both modes, and if the differences between PS5 and XSX in quality mode are similar to those we see in performance mode... that means they'll be similar in the rest of the game, favoring XSX (without it needing to be in the same %).
You want to argue that this possibility doesn't exist and that the framerate behavior is different in that mode VS performance mode.. Good luck convincing anyone on this, but OK
No, you've only shown a specific moment that coincides with a specific part of the game that is known to deviate greatly from the overall performance of the rest of the game as evidence while denying the parts where the framerate and data indicate otherwise.
I mean, you're basing your argument on a specific moment when you know full well that the rest of the game on XSX runs above the VRR measurements on XS.
The fact is that generally, a 60fps mode, especially on these consoles, will always show more bottlenecks than a 30fps mode. And yes, you're ignoring performance mode, where you have the same effect on "death row," while the rest of the framerates show their highest figures. It's logical to assume that the behavior in quality mode is similar, not the opposite. Starting with the fact that 40fps is maintained, while on PS5 (the other benchmark), it drops to the low 30s.
No, is not. Surely if in Wukong and other cases it were not possible to force that 60fps mode unlocked via VRR in the system you would be here defending the same thing even if the data said the same as Control.
It is
I've had the opportunity to play countless games in these conditions, and the experience has always been better than with 40fps modes. Not only with the responsiveness, but also because in many cases DRS is in effect, and you avoid graphical and resolution options that cut down on balanced modes. And no, it's not me. The vast majority of people have the same impression, and that's why the opportunity for this extra option, which is definitely more than viable here on XSX, is missed and in demand.
Last edited:
Skifi28
Member
I can't tell much from a static screenshot, but the game's RT reflections are extremely noisy even with DLSS on quality.For curiosity sake, I took a picture of the game on the same area with FSR4. Max settings, RT to medium. 1440p upscaled to 4k.
All this discussion for a 18-22% of more raw power which show mostly in unlocked fps jesus christ
DF should really stop to remark such stuff as a sort of unexpected revelation. Or at least they should stop to complain about the fanboys complaints when they repeatedly feed their narrative with nothing burger, it's the consequences you have if your main target is such audience.
Last edited:
It never ran better on ps5 lol. What you are talking about was the XSX I/O issues not fps perfomance. They already showed in photo mode XSX gpu had a notable performance advantage.
The irony of this sentence is that when Control first got its PS5/Series update, the engine actually ran better on PS5.
Last edited:
ShaiKhulud1989
Gold Member
Honestly, I find their temporal choices for Control a bit annoying even on PC. But at least consoler verisons recieved at least any semblance of clarity, finally.
FalconPunch
Member
Must be a joke post because control is anything but amazing. Very bad story, repetitive gameplay. The only good things are art direction and physics.Good thing then that we're talking about Control, an amazing game.
The point is GPUs needed improvements because flow control for massive parallelism they employ sucked, not support for non existing sparsity 'feature'.
That wasn't the case for vector accelerators like the ps3 that were designed for such flexibility to begin with.
The "in hardware" nomenclature is frequently misused (especially so by Nvidia PR that tacked feature names to driver improvements more often than ppl realise), but here its specific to classic gpu weakness not some general purpose hw acceleration fixed block.
Negotiator101
Member
I have tried a number of times to like this game but I'm with you on this one. AW2 is infinitely better.
lol aw2 is a glorified walking sim with even worse, nonsensical story.
generally sam lake is way too overrated as a writer
control sucks in narrative department as well, but at least has some fun combat
Last edited:
Yes, as you say, it's normal that between consoles so similar in power and architecture that there are cases in every sense because many circumstances come into play in game development that decide the balance.
Then, we've had a year where (could it be circumstantial or not, we will see the trend on future games launch) a majority of technically relevant games on differents graphix engine have consistently shown a lead in favor of the XSX over the PS5 base. In some cases, these advantages are extremely clear compared to the general norm between the two consoles in previous years. That this happens at this point in the generation, and when it comes to consoles that will evidently be affected in some way in game optimization and development by being the platform with the smallest user base, I think is not defining of course, but significant and relevant.
If that advantage can lead to the implementation of a game mode that improves the gaming experience for many users on a concrete platform... then the discussion is lawful and legitimate.All this discussion for a 18-22% of more raw power which show mostly in unlocked fps jesus christ
Why can't we request, for example, a performance mode with Lumen on XSX in Wukong if there are no technical limitations?
If this type of discussion and complaints didn't exist, studios would surely not bother, on most occasions, to fix bugs and optimize things that aren't suited to the hardware of the different platforms on which they release their game.
Therefore, you may believe that DF's work only leads to arguments between users of different platforms defending the performance of their product or complaining about the use of its capabilities, leading to console wars (which can be true), but I see a window to expres complaint and doing pressure on Studios that haven't done their job well to fix what isn't working well on games than cost up to 80€. A large number of cases have been corrected after and thanks to DF analysis.
Last edited:
Zathalus
Member
Yes, modern GPUs are great at uniform workloads and struggled with sparse data. Vector units in the PS2 and Cell could handle data sparsity relatively easily, but still required it to be manually managed in software.
The sparsity in Ampere onwards is not just mere driver improvements, the Tensor cores have new instruction encodings for for sparse matrix multiplication.
So yes, while nearly any programmable vector processor can handle sparsity through software logic the difference is whether the hardware itself includes instruction formats and control logic to skip zero operations automatically without software intervention. It's explicitly a hardware level feature that brings enhancements.
Another one joins the AW2>>Control group
PS. Although it's not that I think Control is a bad game, I just much prefer AW2 if I have to choose between the two.
adamsapple
Or is it just one of Phil's balls in my throat?
Sure but that's more general purpose than sparsity (or even matrix processing, depending on implementation).
Ie. most CPU SIMDs have always had at least parts of this logic dating back into the 90ies - and lots of research out there shows that past a point, having a handful of control instructions to perform something vs. complex arithmetic instruction that does it all has statistically insignificant impact on performance*
But I digress - point being that PS5 Pro does indeed run sparse matrices about 2x faster than if they aren't, and the rest of implementation details seem to be irrelevant to general public (ie. noone ever cared that PSP could do register memory aliasing to read/write/operate on data in memory as multiple different data structures simultaneously with 0 performance overhead - no matter how assembler math nerds like me were fawning over it).
*it does impact performance - but that of a software engineer if they have to manually fiddle with the stuff - eg. I can accept that SPEs did indeed hit their peak compute throughput without access to Vector components as IBM engineers surmised - but it was hell of a lot more painful to write code that way.
Yea but if you stand perfectly still, wait for camera to stop shaking, and pause the game, the table jaggies go away in the TAA too.
I think Remedy really needs to work on just how damn polarizing their output is - there's something wrong here.
Personally I am not a huge fan of control (long laundry list of what didn't work for me) but AW2 was just a complete dud.
TLDR - AW1 >>> Control >> AW2
Last edited:
I tried the update to see the improvements to graphics mode and wasn't really impressed. But then I bought a new TV where I have the PS5 Pro and the difference was drastic. VRR w/120hz is really a game changer for modes like this where the FPS hangs in the 40-50s. I could not really tell the difference in fluidity between performance and graphics mode after playing on the new TV.
adamsapple
Or is it just one of Phil's balls in my throat?
I will not abide by this Control slander this last page has taken.
Diego Brasil
Member
diffusionx
Gold Member
When Control was on PS+ I accidentally redeemed the PS4 version. So I am very annoyed now as I reallly enjoyed the game on PC and would want to play it again on PS5 Pro. Hopefully it goes on sale soon (probably not).
AW1 is awful.
AW1 is awful.
Last edited:
Soltype
Member
They're smoking crack, they can say what they want about the story or aesthetics, but the gameplay was good. I do recommend people play it on the mouse and keyboard though, especially when you start flying.I will not abide by this Control slander this last page has taken.
alabtrosMyster
Member
Yes, this is why MS raised the price of the Series X, they were actually making it a pro console, but their marketing team forgot to tell anyone!
Radical_3d
Member
I'm trying II starting today since it's on plus. If this ends being a good game I'm going to regret not giving my money to the devs. I needed some linear 20-hours adventure after my 209+ hours Elden Ring + DLC play.
18/22% don't lead to anything of meaningful in a multiplat development, imagine even a new perfomance mode for that and do you think they really need of DF to spot relevant differences in performance? Come on be serious, they have their tools for that. No need of this childish cheerleading of the XSX hardware capability; such difference is almost irrelevant and knew from awhile.Yes, as you say, it's normal that between consoles so similar in power and architecture that there are cases in every sense because many circumstances come into play in game development that decide the balance.
Then, we've had a year where (could it be circumstantial or not, we will see the trend on future games launch) a majority of technically relevant games on differents graphix engine have consistently shown a lead in favor of the XSX over the PS5 base. In some cases, these advantages are extremely clear compared to the general norm between the two consoles in previous years. That this happens at this point in the generation, and when it comes to consoles that will evidently be affected in some way in game optimization and development by being the platform with the smallest user base, I think is not defining of course, but significant and relevant.
If that advantage can lead to the implementation of a game mode that improves the gaming experience for many users on a concrete platform... then the discussion is lawful and legitimate.
Why can't we request, for example, a performance mode with Lumen on XSX in Wukong if there are no technical limitations?
If this type of discussion and complaints didn't exist, studios would surely not bother, on most occasions, to fix bugs and optimize things that aren't suited to the hardware of the different platforms on which they release their game.
Therefore, you may believe that DF's work only leads to arguments between users of different platforms defending the performance of their product or complaining about the use of its capabilities, leading to console wars (which can be true), but I see a window to expres complaint and doing pressure on Studios that haven't done their job well to fix what isn't working well on games than cost up to 80€. A large number of cases have been corrected after and thanks to DF analysis.
Last edited:
rnlval
Member
That's a red herring. AMD added hardware 4:2 sparsity feature with RDNA 4, not just NVIDIA.
PS3 can do ML with sparsity...
PS2... can compute ML with sparsity.
Out of all the SIMD compute accelerators in consoles released for last 25 years, more of them do sparsity before NVidia even 'invented' ML than after.
Accelerating Generative AI on AMD Radeon™ GPUs - AMD GPUOpen
Discover AMD-optimized ONNX models on Hugging Face for AMD Ryzen™ AI APUs and Radeon™ GPUs and incredible performance with the AMD Radeon RX 9000 Series' advanced AI accelerators.
- AMD Radeon™ RX 9000 Series (AMD RDNA™ 4 architecture) Advancements
- Second-generation AI accelerators doubles fp16, and quadruples int8 matrix multiplication operations compared to the AMD RDNA™ 3 architecture.
- Hardware support for 4:2 structured sparsity, improving throughput when exploited in models and libraries.
- With Sparsity up to 8x int and 4x fp16 operation per cycle improvement vs AMD RDNA™ 3 architecture
NVIDIA's argument is hardware-accelerated 4:2 structured sparsity, not some weak sauce CELL.
2:4 Sparse Llama FP8: SOTA performance for NVIDIA Hopper GPUs | Red Hat Developer
Advancing AI efficiency is more critical than ever, and sparsity has proven to be a cornerstone in this pursuit
- Hardware-accelerated sparsity: Achieves an average of 30% lower latency and 20% higher throughput from sparsity alone on NVIDIA Hopper GPUs.
- FP8 quantization compatible: Supports NVIDIA's FP8 format with sparsity, enabling an average of 1.7X lower latency and 1.5X faster throughput.
Last edited:
FalconPunch
Member
All 3 are awful. I've said this many times, Remedy should really be closed down soon. They've skirted by on deals like epic game store exclusives, nvidia promo deals, and services like ps plus. If they had to survive on the merit of their games sales in this era, they'd have been shut down or bought out.
They make story driven games with bad stories and poor gameplay.
adamsapple
Or is it just one of Phil's balls in my throat?
Diego Brasil
Member
Yes, as you say, it's normal that between consoles so similar in power and architecture that there are cases in every sense because many circumstances come into play in game development that decide the balance.
Then, we've had a year where (could it be circumstantial or not, we will see the trend on future games launch) a majority of technically relevant games on differents graphix engine have consistently shown a lead in favor of the XSX over the PS5 base. In some cases, these advantages are extremely clear compared to the general norm between the two consoles in previous years. That this happens at this point in the generation, and when it comes to consoles that will evidently be affected in some way in game optimization and development by being the platform with the smallest user base, I think is not defining of course, but significant and relevant.
If that advantage can lead to the implementation of a game mode that improves the gaming experience for many users on a concrete platform... then the discussion is lawful and legitimate.
Why can't we request, for example, a performance mode with Lumen on XSX in Wukong if there are no technical limitations?
If this type of discussion and complaints didn't exist, studios would surely not bother, on most occasions, to fix bugs and optimize things that aren't suited to the hardware of the different platforms on which they release their game.
Therefore, you may believe that DF's work only leads to arguments between users of different platforms defending the performance of their product or complaining about the use of its capabilities, leading to console wars (which can be true), but I see a window to expres complaint and doing pressure on Studios that haven't done their job well to fix what isn't working well on games than cost up to 80€. A large number of cases have been corrected after and thanks to DF analysis.
Let's analyze the 2025 games (eliminating competitive games, for example, Microsoft games making PS5 or Sony for Xbox) as this could be seen as the competing company benefiting from their game on your device
Identical: 6
Better on Xbox: 5
Better on PS5: 3
Undefined, as each has an advantage: 2
As a result, of the 16 games analyzed, only 5 are better on Xbox.I consider the similar games a victory for PS5, as it has 10 teraflops and Xbox Series has 12. (less powerful console having the same performance)
Battlefield 6 virtually identical
slightly better frame rate on xbox
That was the difference, 95fps versus 93fps
Dying Light: The Beast
This is tricky to measure performance because the X series doesn't have full SSR coverage.And this affects performance.
xbox has an advantage in performance and ps5 in ssr
Borderlands 4
Here we have a positive point for the Xbox and another for the PS5.The resolution on the Xbox is a little higher... but in most scenes the difference is very small.On the other hand, the Xbox SX suffers from screen tearing, while the PS5 doesn't.
virtually equal have a small advantage
Cronos: The New Dawn
Here, the series has better performance and some scenes have slightly higher resolution.
We have an advantage over the Series X.
Assassin's Creed Shadows
identical
Metal Gear Solid Delta
Both have performance issues here, a horrible experience for both.
But the Series X has a slight performance advantage.
Wuchang: Fallen Feathers
Virtually identical, sometimes the PS5 is ahead in performance, sometimes the Xbox Series.The average of both is the same.
Mafia: The Old Country
better performance on the x series, very significant difference
MindsEye
Here it seems that the Xbox has an advantage in frame rate, but both are horrible (PS5 drops to 18fps and Series X to 24fps)
Elden Ring Nightreign
slightly better frame rate on ps5
Clair Obscur: Expedition 33
identical
Atomfall
identical
Split Fiction
identical
Monster Hunter Wilds
ps5 has slightly better performance
Kingdom Come Deliverance 2
here ps5 has better performance
Sniper Elite: Resistance
identical
Last edited:
Yes - calling 'sparsity' a fixed function feature is.
Cell is obviously not comparable to modern AI accelerator in raw performance (8-way Cell would output 1TOP Int8/2TOPs Int4) but matrix multiplication on a 50% sparse matrix 'will' run 2x as fast when optimized for it.
Research papers demonstrating this fact on Cell are decades old (including scaling on larger than 8 SPE arrays).
rnlval
Member
CELL is still a weak sauce and a dead end.
Sparse Matrix-Vector Multiplication software implementation on CUDA GTX 280. This is not hardware-accelerated.
Sparse Matrix-Dense Vector Multiply on G80: Probing the CUDA Parameter Space - SlideServe
Sparse Matrix-Dense Vector Multiply on G80: Probing the CUDA Parameter Space. Comp 790 GPGP Project Stephen Olivier. Currently…. Have a working “naïve” implementation in which each thread computes one dot product (Similar to Sashi’s implementation) 1.26 GFLOPs, 7.56 GB/s for n=32k, nz/row=20
www.slideserve.com
GeForce 8800 GTX was released a few weeks before the PS3.
Leveraging Dynamic Sparsity in Training DNNs on General-Purpose SIMD Processors.
AMD RDNA 4's and NVIDIA Ampere's 4:2 sparsity are hardware-accelerated implementations.
NVIDIA CUDA survives to this day, while CELL is a dead end. NVIDIA >>>>>>>>>>>> IBM.
I'm game for mid-2000s gaming PC with the CUDA vs teh CELL debate. You're not going to win this debate when the CELL family is dead along with PowerCrap (PowerPC).
Last edited:
The Series X 2TB cost $800 in the US.
blastprocessor
The Amiga Brotherhood
A blip in the space time continuum?
There is much more relevant games was launched this year LOL, more if you count complete year. Where are, Oblivion remaster, WH 40K, Wukong, HB2, Doom, Indiana Jones, FH5, Gears, SW Outlaw.... . All these games have a clear winner , very clear in some case ("clear winner" is understood as identifying which version has the advantage. The perception of differences in games for the average user is another matter. From that point of view, surely 99% would be "identical"Let's analyze the 2025 games (eliminating competitive games, for example, Microsoft games making PS5 or Sony for Xbox) as this could be seen as the competing company benefiting from their game on your device
Identical: 6
Better on Xbox: 5
Better on PS5: 3
Undefined, as each has an advantage: 2
As a result, of the 16 games analyzed, only 5 are better on Xbox.I consider the similar games a victory for PS5, as it has 10 teraflops and Xbox Series has 12. (less powerful console having the same performance)
Battlefield 6 virtually identical
slightly better frame rate on xbox
That was the difference, 95fps versus 93fps
Dying Light: The Beast
This is tricky to measure performance because the X series doesn't have full SSR coverage.And this affects performance.
xbox has an advantage in performance and ps5 in ssr
Borderlands 4
Here we have a positive point for the Xbox and another for the PS5.The resolution on the Xbox is a little higher... but in most scenes the difference is very small.On the other hand, the Xbox SX suffers from screen tearing, while the PS5 doesn't.
virtually equal have a small advantage
Cronos: The New Dawn
Here, the series has better performance and some scenes have slightly higher resolution.
We have an advantage over the Series X.
Assassin's Creed Shadows
identical
Metal Gear Solid Delta
Both have performance issues here, a horrible experience for both.
But the Series X has a slight performance advantage.
Wuchang: Fallen Feathers
Virtually identical, sometimes the PS5 is ahead in performance, sometimes the Xbox Series.The average of both is the same.
Mafia: The Old Country
better performance on the x series, very significant difference
MindsEye
Here it seems that the Xbox has an advantage in frame rate, but both are horrible (PS5 drops to 18fps and Series X to 24fps)
Elden Ring Nightreign
slightly better frame rate on ps5
Clair Obscur: Expedition 33
identical
Atomfall
identical
Split Fiction
identical
Monster Hunter Wilds
ps5 has slightly better performance
Kingdom Come Deliverance 2
here ps5 has better performance
Sniper Elite: Resistance
identical
)So, among the ones you've mentioned (that include DLC from past games
), the question is whether it's correct to call games that aren't ,"better on PS5" or "identical". Totally subjective for your part.For example, calling AC Shadows identical when on XSX there's a resolution advantage, slightly more stable frame rates, and even higher graphical settings (vegetation/detail LOD in the distance) in performance mode... NO, is not identical.
Sniper elite is not identical. XSX have much higher average resolution in this Game VS PS5.Thanks to other websites (Brazilpixel), games have been discovered where XSX has a clear advantage in average DRS resolution, which DF sometimes does not detail and only show lower bound and higher.
Mindeye is a total disaster, yes, but is better on XSX. You're the one who decided to include it, so include it where appropriate
Battlefield 7 identical?? XSX have a consisten adventage in frame rate. Also a sligth DRS average. For much less you give some games on you list "better on PS5"
.Monster Hunter Wild? You say better on PS5 when in It is in one mode and XSX is better on other. Subjetive
MGS Delta "identical" when is not. Is more in "better on XSX".
You included Elden Ring DLC but not this Control patch
Anyway, you can make the table however you want now, but it's far from the "equality" you're talking about and "a success for PS5" as you say.
Not only because there are significantly more games that show an advantage on XSX, but also because some of them, almost the most technologically advanced, show a consistent advantage on XSX (sometimes above the supposed 18% reported on paper).
At this point in the generation, given the sales situation, when the hardware is supposedly being squeezed to a greater extent... A year ago, if some predicted this situation here today reaction was mockery. The general idea was that things would get worse on XSX when the sales gap increased, and that studios would dedicate even less time to optimizing for XSX... Well, even with this circumstance affecting surely XSX versions and PS5 being the common base platform for most developments, XSX is proving more consistent than ever in its versions. It's clear that this is a relevant and significant situation.
Then it's up to everyone to draw their own conclusions, but these are the results. We'll see if this trend continues next year or not, who knows.
Last edited:
SKYF@ll
Member
There is much more relevant games was launched this year LOL, more if you count complete year. Where are, Oblivion remaster, WH 40K, Wukong, HB2, Doom, Indiana Jones, FH5, Gears, SW Outlaw.... . All these games have a clear winner , very clear in some case ("clear winner" is understood as identifying which version has the advantage. The perception of differences in games for the average user is another matter)
So, among the ones you've mentioned (that include DLC from past games
For example, calling AC Shadows identical when on XSX there's a resolution advantage, slightly more stable frame rates, and even higher graphical settings (vegetation/detail LOD in the distance) in performance mode... NO, is not identical.
Sniper elite is not identical. XSX have much higher average resolution in this Game VS PS5.Thanks to other websites (Brazilpixel), games have been discovered where XSX has a clear advantage in average DRS resolution, which DF sometimes does not detail and only show lower bound and higher.
Mindeye is a total disaster, yes, pero is better on XSX. You're the one who decided to include it, so include it where appropriate
Battlefield 7 identical?? XSX have a consisten adventage in frame rate. Also a sligth DRS average. For much less you give some games on you list "better on PS5"
Monster Hunter Wild? You say better on PS5 when in It is in one mode and XSX is better on other. Subjetive
MGS Delta "identical" when is not. Is more in "better on XSX".
You included Elden Ring DLC but not this Control patch
Anyway, you can make the table however you want now, but it's far from the "equality" you're talking about and "a success for PS5" as you say.
Not only because there are significantly more games that show an advantage on XSX, but also because some of them, almost the most technologically advanced, show a consistent advantage on XSX (sometimes above the supposed 18% reported on paper).
At this point in the generation, given the sales situation, when the hardware is supposedly being squeezed to a greater extent... A year ago, if some predicted this situation here today reaction was mockery. The general idea was that things would get worse on XSX when the sales gap increased, and that studios would dedicate even less time to optimizing for XSX... Well, even with this circumstance affecting surely XSX versions and PS5 being the common base platform for most developments, XSX is proving more consistent than ever in its versions. It's clear that this is a relevant and significant situation.
Then it's up to everyone to draw their own conclusions, but these are the results. We'll see if this trend continues next year or not, who knows.
I'm sorry to interrupt your discussion, but having played both for five years, I would say the PS5 and XSX are a draw.
The difference in resolution of 100-200p and the difference in frame rate of 5-10fps are so small that you wouldn't be able to tell the difference in a blind test.
Most games have the same target resolution and target frame rate for PS5 and XSX.
Also, when running UE5 games at 60fps, both consoles struggle.
I hope that with the upcoming UE5.6 and later games, both consoles will run more stably.
It's getting difficult to run the latest games on a five-year-old console.