Current Generation Games Analysis Technical Discussion [2020-2021] [XBSX|S, PS5, PC]

VGTech Immortal Fenyx analysis. PS5 easily wins in both modes (30fps and 60fps).

Performance mode:
- slightly higher performance on PS5 (min is lower on XSX).
- Higher minimum resolution found on PS5 (1280p) while it can drop at 1080p on XSX. At min res PS5 pushes about 39% more pixels.

Quality mode:
- Higher resolution on average on PS5. Locked native 4K on PS5, drops down to 1872p on XSX (PS5 is pushing at least 32% more pixels).
- Both locked at 30fps.

@chris1515 After AC it's the second game with a higher res on PS5. That should alleviate your concerns about PS5 bandwidth.

EDIT: percentage corrections.

 

Almost parity.
But still a last gen game. It even runs on switch !

 

Maybe it is just the GDK I wait the dictator video for have a first idea. The PC version were probably not using the GDK.

 

You should really stop repeating Microsoft own PR excuse.

 

I dont think comparing with the PC is that good of a measurement either. If you decide between consoles, compare with the consoles. The pc is a totally different market. It doesnt matter how PS5 compares to a PC if you want a console

 

Again seeing the spec sheet, this is the only explanation. I don't do technical comparison on rumors. Here devs talk about it since long time. We have people saying this is true.

Until we have solid proof of what going on like a PS5 APU photo. The best explanation is an API problem on Xbox side.

The quest of secret PS5 advantage reminds me the hidden GPU of misterxmedia.

We have an idea how basically an AMD GPU perform against a Turing GPU and the PC version were probably not using the GDK. Taking the software out of the scope, compare with a PC version will be the best way to know where is the PS5 performance. On AC Vahlalla it will be easy to see for example if it performs better than a 5700XT...

I like transparency this is a key, Microsoft was great on this side, Sony much less...

 

We have others data and we already know enough stuff about PS5 design thanks to Cerny (narrow and fast design, number of CU by SA, importance of low latency etc.). And you shouldn't believe Microsoft supposed transparency anyways, like their full RDNA2 "transparent" statement. We have for instance one respected member here saying that XSX GPU could be actually RDNA 1 + RDNA 2, with solid reasoning behind his claims.

You really should not listen to any marketing PR statements as your source of information about those consoles. MS have focused their marketing campain around compute and fixed clocks for a good reason.

 

Aha, how so? missed that in case. Where not both consoles RDNA2?

 

Believe Cerney because.
Don't believe MS because.

 

As of know I guess we can summarise

1. PS5 lower bandwidth spec has not been an issue yet vs XSX, VRR seems to be a bigger one, ref tearing.
2. Xbox GDK might be the reason that it currently it lags behind in certain scenarios, it will get updates that may or may not remedy this.
3. Future games designed for this new gen of platforms might change the current status quo.
4. The difference is currently mostly measurable...

 

Well, both games use the same engine. This makes the results less surprising.

 

the first real meaningful next gen comparison will be GTA6....in 5 years.

 

You think that soon?

 

Only news here is that they proactively applied the same settings as the valhalla patch to another game with the same engine and most likely same asset guidelines for the artists.

 

Sony never promised being the best in some arbitrary metric. They only presented the direction they were going in for their console and why they decided to go that way. That's pretty fucking transparent in my eyes regardless if they produced a die shot to slather over every internet publication.

 

I'm not going to get into this here, apart from just saying an arbitrary metric that has been used for years as a measure of power (across same base architectures) seems very disingenuous.
Whether that power is being realised or has inherent issue being tapped into its another issue.
So because games aren't displaying it, everything they've said and shown in hot chips etc isn't being open, due to the TF metric? Anyway, I'll not be taking part in this conversation anymore as it's only going to go down hill and we can agree to disagree.

 

I really don't understand the "tflops are arbitrary" meme. Floating point operations are most of what gpus do. It's an extremely relevant metric -- so is bandwidth (xbox wins), bus size (xbox wins), architecture (they're mostly the same). Microsoft's "most powerful console" claim is pretty safe, what remains to be seen is whether their power advantage is actually relevant to game tasks -- the ps5 has some different design decisions (such as the variable clock, less CUs, etc) which could turn out to be much more practical for rendering games.

Many suspect microsoft's bets on what hardware nextgen games will need are good, and (after the recent poor showings against ps5 on crossgen titles) many think they aren't! Most of your posts I've seen feel more like console warring than participation in that discussion.

(also, a lot of people are discounting the tools gap, but even with with something as simple as "the profiling tools on xbox suck" you could easily see big framerate dips not get fixed on xbox within the allotted development time -- tools ((not to mention providing a single target for developers)) are a place where i think we can assume sony clearly has the right idea and microsoft is way behind)

 

Thats going to be true for both sides.

They could either choose to go wide/slow or narrow/fast. Not both. Different decisions, faster frontend pays off now atleast.

Welcome to the internet forums

 

20% flops difference is drop in ocean. Much more matters how good developers are, do they have enough budget(time, resources), tools on platform, which platform is lead platform and get's most developer attention etc. 20% difference in flops is pretty meaningless once you think about the big picture. Not all units in gpu's scale like flops. Clockspeed can also matter. Better tools could easily sway which platform is lead platform or which platform is easier to optimize for mitigating minor flops differences completely.