• 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.

[NxGamer] Spider-Man Remastered PC vs PS5 vs Steam Deck vs PS4 vs 750Ti - Technical Review & Comparison

Tqaulity

Member
Once again a detailed and informative technical analysis by NxGamer on the Spiderman PC port. I mentioned this before in this forum that Spiderman would A) be one of the best examples of PS5 vs PC performance in terms of an engine developed and optimized for console and B) would show that you need a much more powerful PC to match or exceed the performance of a PS5 due not only raw GPU power but overall system utilization and CPU load. I mentioned that in fact we will see a PS5 match more closely with a RTX 3070 based PC system which is generally the best case scenario for PS5 performance and many thought I was crazy. Now that the PC game has been out for some time, there are a number of examples of users demonstrating how their 3070 GPUs are pretty much a match for the PS5. In this analysis, NxGamer literally specifies that the PS5 performance is within 5-10% of a 3070 level on average and is showing to be 20-30% faster than his OC 2070 (virtually a 2070 Super). Again, best case scenario and no the PS5 GPU isn't equivalent to a 3070 on it's own. But Spiderman is a classic case of how the architecture and efficiency differences between PC and console can really make a difference in total system performance when the software is optimized for it. All that constant talk about how it's not just about TFLOPs and the GPU alone but the entire system matters...wasn't coming from nowhere. Carry On! :messenger_beaming:



Thanks for the follow up. It's really not that surprising if you really think about it. Again, never said PS5 GPU is equivalent to a 3070, but it can deliver a similar experience when the game/engine it optimized for it.

So to wrap up, now that we have Digital Foundry's analysis to go with IGN's, Tom's Hardware, DSOGaming, Digital Trends, and others let's see where we're at:

S2QAk43.png


Can we agree with our eyes that Alex's optimized settings are actually a bit lower than both PS5 settings (even the Performance RT)? PS5 settings are pretty much a match for High + RT High as discussed.

So again, where I'm coming from is simply this: if I'm looking at a PC sku of a console game, I'm looking at what will it take to get a similar or better level of performance beyond the console. Both skus are running the same game but have very different methods to get there when you consider the unified memory and much more efficient data throughput on a PS5. The key point here for me is identifying a PC that will deliver the same consistency in performance as the console equivalent and that's where the 99th % frametime/1% FPS values come in to play. You need that to compare to locked (or min) FPS on console.

With that in mind, the question is what PC GPU/CPU will it take to deliver (for example) 1440p + DRS + RT high + 1% 60fps to match or exceed the PS5's performance RT mode? Or similarly 4K + +RT High + 30fps (or 40fps with VRR)?

Here is just one example of Spiderman Remastered running on a RTX 3070 Ti | Ryzen 9 5900X system with DF optimized settings @1440p:

jOISPZt.png


So I would look at that and note that 1% Low being just at 60fps even though the average is 96fps. Meaning that if I wanted to cap the FPS to at least 60fps, it will be mostly stable there. Again this is a 3070TI and a Ryzen 9 5900X CPU. Both supposedly way above a PS5's level of performance. But the PS5 Performance RT is actually slightly higher than these settings and practically never drops a frame from 60fps.

Of course, every PC is different and even multiple PCs with a 3070 or 3070Ti will vary in perf depending on their other components. But there has already been a number of tests that show that if I were looking to build a PC for Spiderman to match or exceed the PS5's level of performance, I'd be looking at at least a 3070 to deliver a consistently on par experience (cinematics, swinging, action, ground level walking etc). Call me crazy if you want...

BTW, I can tell you from experience that those PC recommended specs don't come from nowhere and there are many months of testing to settle on what they are. Not just random circumstance that the "Amazing Ray Tracing" (i.e. close to PS5 Performance RT mode) specs mentioned a RTX 3070 / R7 3700X to deliver that.
 
Once again a detailed and informative technical analysis by NxGamer on the Spiderman PC port. I mentioned this before in this forum that Spiderman would A) be one of the best examples of PS5 vs PC performance in terms of an engine developed and optimized for console and B) would show that you need a much more powerful PC to match or exceed the performance of a PS5 due not only raw GPU power but overall system utilization and CPU load. I mentioned that in fact we will see a PS5 match more closely with a RTX 3070 based PC system which is generally the best case scenario for PS5 performance and many thought I was crazy. Now that the PC game has been out for some time, there are a number of examples of users demonstrating how their 3070 GPUs are pretty much a match for the PS5. In this analysis, NxGamer literally specifies that the PS5 performance is within 5-10% of a 3070 level on average and is showing to be 20-30% faster than his OC 2070 (virtually a 2070 Super). Again, best case scenario and no the PS5 GPU isn't equivalent to a 3070 on it's own. But Spiderman is a classic case of how the architecture and efficiency differences between PC and console can really make a difference in total system performance when the software is optimized for it. All that constant talk about how it's not just about TFLOPs and the GPU alone but the entire system matters...wasn't coming from nowhere. Carry On! :messenger_beaming:


Again it shows how biased DF are. So if I understood correctly they have compared PC uncapped vs PS5 capped while completely disregarding the unstable framerate (sub 60fps drops) on PC with a bit of cherry picking with the selected scenes.
It's borderline dishonest. While on the other hand NXGamer is doing the most fair comparison as usual. Like in Death stranding he found out the PS5 was often near 3700 level.
 

Panajev2001a

GAF's Pleasant Genius
Again it shows how biased DF are. So if I understood correctly they have compared PC uncapped vs PS5 capped while completely disregarding the unstable framerate (sub 60fps drops) on PC with a bit of cherry picking with the selected scenes.
It's borderline dishonest. While on the other hand NXGamer is doing the most fair comparison as usual. Like in Death stranding he found out the PS5 was often near 3700 level.
No… Battaglia making an unfair PC biased comparison? No way…

shocked philip j fry GIF
 
Last edited:

Panajev2001a

GAF's Pleasant Genius
that's a long video to excuse a sub par PC port
Really, a bad port? With Nixxes, a very respected studio whose optimisation (and PC ports) is something that made them famous, being given quite a long time to do it and a very optimised codebase to begin with?

Maybe it just shows that no matter what… porting games optimised for consoles, the other side of the console optimisations factor, porting games around is not a flick of the switch affair and costly (not dissimilar to making a game for a faster HW or in this case a custom one and porting it around).
 
Last edited:

SlimySnake

Flashless at the Golden Globes
3070 is interesting, but this could be due to the ray tracing tech that was designed to take advantage of the PS5 hardware. Pretty much every other ray tracing game offers way better performance than even the 2080 which is on par if not better than the PS5 in standard rasterization games.
 

01011001

Banned
Really, a bad port? With Nixxes, a very respected studio whose optimisation (and PC ports) is something that made them famous, being given quite a long time to do it and a very optimised codebase to begin with?

Maybe it just shows that no matter what… porting games optimised for consoles, the other side of the console optimisations factor, porting games around is not a flick of the switch affair and costly (not dissimilar to making a game for a faster HW or in this case a custom one and porting it around).

a game that looks like a slightly enhanced PS4 game running worse on some rigs than Cyberpunk... mhhh, yeah I'd call that a but of a miss
 

M1chl

Currently Gif and Meme Champion
PS5 has GPU capabilities equivalent of a 2070 Super. Not bad but not cutting edge PC stuff either.

Wait for PS5 Pro to reach 3070 levels of performance.

If you are looking for 1440p gaming, just get a 3060ti and middle of line CPU like 12400 and you are good to go for next 3 years
In raster, maybe it's still AMD GPU, thus...it's more like 2060 with other aspect like RT.

Xbox isn't better, that wasn't my argument
 
Last edited:

ChiefDada

Gold Member
PS5 owners still fixated on proving their Spiderman version is better lol.

PS5 in Fidelity mode vs 2019’s 2070 with a Ryzen 2700X. But it’s more realer that way.

The desperation from PS5 owners to claim their console is the equivalent of a $1000000 PC!


New Girl Facepalm GIF by HULU



Wait for PS5 Pro to reach 3070 levels of performance.

Lol, no need to wait. The PS5 already outclasses the 3070 at w/max texture settings in this game.
 

RoboFu

One of the green rats
Again it shows how biased DF are. So if I understood correctly they have compared PC uncapped vs PS5 capped while completely disregarding the unstable framerate (sub 60fps drops) on PC with a bit of cherry picking with the selected scenes.
It's borderline dishonest. While on the other hand NXGamer is doing the most fair comparison as usual. Like in Death stranding he found out the PS5 was often near 3700 level.
“ I’m no console gamer.., hur hur hur”
 

killatopak

Gold Member
which goes to show how fucked this port is, on Alex's end it didn't look like this on these settings

which it's why it's so hilarious that people call this a good port
An unfair comparison since it was optimized for 2060. Both consoles generally perform higher than a 2060 on other titles. Even goes as far as 2070 super if it’s more raster heavy and less RT is used.
 

IFireflyl

Gold Member
I own this on both PS5 and PC (Steam). It was just as fun playing through it on PC as it was when I played it on PS5. I didn't notice any frame rate issues when playing it on max settings at 1440p, but I also have 64GB of RAM and an RTX 3090. I thought the game was beautiful, but I don't remember it being significantly better or worse than the PS5. Except for reflections. I felt like the reflections were way better on PC than the PS5.

I can't speak to lower-tier hardware, but at least for my hardware it was a satisfactory port. Either there is a lot of nit-picking going on (for both sides), or any issues with PC are seen with lower-tier hardware.

I'm still deciding if I want to trade in my 3090 for a 4090 when it comes out. The performance looks sick (in a good way), but that power draw looks more sick (in a bad way). I think the 40XX series cards are where we'll be able to get a stable 4k/60fps across the board. Games are so hit or miss in this area that I just set everything to 1440p by default.
 

01011001

Banned
An unfair comparison since it was optimized for 2060. Both consoles generally perform higher than a 2060 on other titles. Even goes as far as 2070 super if it’s more raster heavy and less RT is used.

that's great and all, but the issue seems to be that options don't scale as they should and apparently they don't work correctly either.

Alex sets density to low and it looks like PS4, here it's set to low and you see no crowd... can't be normal
 

killatopak

Gold Member
that's great and all, but the issue seems to be that options don't scale as they should and apparently they don't work correctly either.

Alex sets density to low and it looks like PS4, here it's set to low and you see no crowd... can't be normal
Honestly, never played Spiderman or watched DF’s take so I wouldn’t know the difference but if that’s true, it does seem suspect.
 

Sosokrates

Report me if I continue to console war
I'm confused when he was talking about he PCI express bandwidth. PCI express transfer data to and from the CPU right? So if PCIe 3.0 is 16gb/s what's the PS5/Xbox series?

I wish Micheal would not go over these things so quickly without really explaining in detail.
 

Sosokrates

Report me if I continue to console war
While that could be a bottleneck, his PC is even slower than it should be even when accounting for that potential bottleneck.



Micheal tends to create these scenarios to try an prove the PS5 has some special secret sauce that give it some significant advantage beyond comparable PC hardware. Then there is just the case of driver efficiency etc.
 
Last edited:
On many comparisons we can see PC (or steam deck) loading in 5 secs while PS5 loads in 1.5 seconds, but on PC they are cheating as the loadings are actually partial.

NXgamers shows PC with a modest CPU (as his zen 3600) needs much more time (here about 20 seconds vs PS5 1.5 sec) to load all the assets. You need a high end CPU costing more than the PS5 to get near (but not beating) PS5 performance in that area.

8r7v251.png
 

Loxus

Member
I think he may lose quite a bit of creditability when comparing PC Vs PS5 but he using a zen+ CPU on the PC Vs a zen 2 CPU on the PS5.
Unless a 3600x does not give better performance then a 2700?

Micheal tends to create these scenarios to try an prove the PS5 has some special secret sauce that give it some significant advantage beyond comparable PC hardware. The only issue is the PC's CPU is not comparable. Then there is just the case of driver efficiency etc.
It would be better if he used a comparable AMD GPU and a 3700 CPU.
He's using a 3600 Zen 2 CPU paired with a 6800.

93M51xr.jpg
 

Sosokrates

Report me if I continue to console war
He's using a 3600 Zen 2 CPU paired with a 6800.

93M51xr.jpg
Yeah I spoke to soon, he uses both a 2700 and 3600.
I still think the angle he is approaching this video is somewhat pointless. Making conclusions based on one game with different parts, drivers and api's is never going to give you good results.

And he mentions the Ps5 HUMA memory system is some significant advantage.
Huma was how the ps4 + x1 did memory and yet it didn't really prove that advantageous compared to comparable PCs back then.

zrWNEaU.jpg
 
Last edited:

Cryio

Member
I don't know what to say about NXG review.

He pushes the PCIe bandwidth narrative, but he's testing with a 2070, that has neither ReBar, nor PCIe 4.0. Then he's continuing with an RX 6800 and R5 3600, but it's been shown RDNA2 benefits the most from Zen3 or Intel 11th gen or newer for ReBar.

He's also testing 2070 + R7 2700X. All fine and good, but Turing and Ampere kinda hate pre-Zen3 CPUs for whatever reason. Testing a 5700 XT for pure rasterized with 2700X or 6600 XT and 2700X would've been a more accurate take on the situation.

Then again, with how heavy the game streams and decompresses data, it seems that it completely overwhelms/invalidates RDNA2 Infinity Cache and the cards start acting like pure GDDR6 cards.

What a conundrum.
 
Last edited:

TxKnight7

Member
I will always believe DF over anyone else

vKO3cdf.jpg

"The biggest upgrade of all? Hardware-accelerated ray tracing support. I'm happy to see a range of granular settings here, but very high exceeds the high setting typically used by PS5. Building geometry boiled down into flat textures on consoles are fully modelled on PC, texture quality is on an altogether different level quality, while the amount of associated world detail - and shadows - also get fully reflected, sometimes where PS5 has no reflections at all. Nixxes also offers PC users the chance to push out draw distance on objects/crowds/traffic beyond the console standard - and that ties into the level of detail/crowd/traffic settings I've already discussed. Just remember that the more you push RT, the higher the load on both GPU and CPU.

The biggest upgrade offered by PC concerns ray tracing, where we're looking at a night and day improvement in every single way.
 
Miles Morales will be interesting. I started playing it again and find the 120hz RT Performance mode completely unplayable as the framerate is all over the place. Even with VRR you easily notice when it drops from 100fps to 58fps which it does every few seconds when web slinging.

The last boss constantly drops below 48fps so it seems way more demanding than anything in this Spider-Man.
 
Top Bottom