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

[mr magoo]

And I think the idea that this is becoming a pattern it probably a signal here that they are comfortable pushing XSX to its limits where some dips are acceptable for the overall gain because the dips are either not numerous enough to matter or not significant enough to hurt.

this may be it, tbh if i havent watched DF vid i could swear game is 60fps rock solid. I think average Joes like me will not notice any drops.

 

[iroboto]

this may be it, tbh if i havent watched DF vid i could swear game is 60fps rock solid. I think average Joes like me will not notice any drops.

as per video and text, they drop in different areas if they do drop. So I would lean to that being a bottleneck difference. Something that resolution adjustment may not help, as what drops XSX may not drop PS5 and vice versa.

 

[pjbliverpool]

Higher avarage res on xsx, slightly better perf on ps5 so quite normal tough strange that ps5 has slowest loding times, devs didn’t spend much time to utilize ps5 io probably

It's likely CPU bound which means this is probably a pretty good CPU benchmark for the two consoles.

 

[snc]

It's likely CPU bound which means this is probably a pretty good CPU benchmark for the two consoles.

yes couse as was said in df video, devs didn't use any nextgen api io posibilities and its more cpu benchmark (as in bc mode when often xsx is also faster in loading games), tough here according to df times are very similar

 

[BRiT]

Loading times for Metrox Exodus latest-gen console version within 12% and filesizes within 5%.

SeriesX: 43.06 seconds - 67.4GB
SeriesS: 45.10 seconds - 67.4 GB
PS5: 48.56 seconds - 70.6 GB

https://twitter.com/x/status/1405904800315809797

 

[iroboto]

It's likely CPU bound which means this is probably a pretty good CPU benchmark for the two consoles.

IO is probably indicative of single threaded loading. Not really sure if this game is a CPU benchmark.

 

[Globalisateur]

"- The PC version runs at 2160p using DLSS in quality mode with an RTX 3080.
- SeriesX and PS5 run at a dynamic resolution of 1080p~2160p, but SeriesX maintains a higher average resolution. SeriesS runs at dynamic 576p~1080p resolution.
- Loading times are faster on Series S|X.
- The 3 console versions maintain a good rate of 60FPS. Series S|X suffer random drops of one or two frames, but they are practically imperceptible.
- All versions use Ray-Tracing in global lighting, although it is more limited on consoles. SeriesS in some areas does not seem to render light bounces well.
- Regarding shading, the PS5 / SeriesX versions have a somewhat lower quality than the PC. The SeriesS version must settle for a lower quality and the absence of shadows as in certain vegetation.
- The PC version still has a clear advantage in texturing, especially due to the use of tessellation. Series S has a somewhat lower quality than PS5/SeriesX.
- On PC, the drawing distance and shading in distance are higher compared to SeriesX / PS5. SeriesS, in this case, has a fairly limited draw distance in open areas.
- It would have been interesting to offer a version without RT in SeriesS to obtain a higher resolution, but it's a great work on all platforms."

According to him both versions have a very similar average resolution. Respectively 1620p-1728p for PS5 and 1728p-1800p for XSX (from 8% to 14% more pixels on XSX albeit with a less stable framerate) and both with a min 1080p resolution (which I actually found in both versions in one same scene).

 

[see colon]

Higher avarage res on xsx, slightly better perf on ps5 so quite normal tough strange that ps5 has slowest loding times, devs didn’t spend much time to utilize ps5 io probably

I keep seeing this posted in comments that they didn't use any of the next gen IO. Have the devs confirmed this? I understand that the loading time scales pretty well to the clock speed of the CPUs found in the 3 consoles, but could it be that the data must be processed or set up. Something like compiling the BVH or something.

Also, the PS5 version is larger? That's a first for this gen isn't it? I guess that could point towards the data being not compressed in a format compatible with PS5's IO block.... Or it could mean that both PS5 and XSX versions are compressed in native formats and the Xbox version just compressed a bit better.

 

[mr magoo]

I keep seeing this posted in comments that they didn't use any of the next gen IO. Have the devs confirmed this? I understand that the loading time scales pretty well to the clock speed of the CPUs found in the 3 consoles, but could it be that the data must be processed or set up. Something like compiling the BVH or something.

Also, the PS5 version is larger? That's a first for this gen isn't it? I guess that could point towards the data being not compressed in a format compatible with PS5's IO block.... Or it could mean that both PS5 and XSX versions are compressed in native formats and the Xbox version just compressed a bit better.

https://forum.beyond3d.com/posts/2210669/

 

[Shortbread]

We are back with yet another patch for Jedi Fallen Order, this one though is a true, full Now Gen Update version. What does the force have in store for you now and which is the chosen one?

 

[Globalisateur]

Now that I have a PS5 I might buy this one. It's being a long time since I haven't played a decent Start Wars (Gamecube, Rogue leader).

1440p 60fps seems to be the sweat spot this gen and again a more stable framerate on PS5. Seems to be a common occurrence now.

 

[Deleted member 11852]

bottleneck differences between the two systems. Dynamic resolution is higher because but frame rate worse because the algorithm that determines what size the frame should be detects (based on the inputs) it should be at that frame buffer size. So the bottleneck is elsewhere for them.

This has been a very common theme though. You can pick your poison, resolution or frame rate but what is causing is a bunch of games using a bunch of different engines? The use of dynamic resolution is clear intention that resolution should be sacrificed for frame rate so why do a bunch of mutliplatform games on Series X keep over-reaching on resolution at the expense of hitting 60?

How hard can it be to tweak the algorithm. It's clearly not an ML algorithm because that would learn and adapt, so it's a bunch of values somewhere in the code. Perhaps there is some bottleneck where lowering the resolution wouldn't boost the frame rate. ¯\_(ツ)_/¯.

 

[Globalisateur]

This has been a very common theme though. You can pick your poison, resolution or frame rate but what is causing is a bunch of games using a bunch of different engines? The use of dynamic resolution is clear intention that resolution should be sacrificed for frame rate so why do a bunch of mutliplatform games on Series X keep over-reaching on resolution at the expense of hitting 60?

How hard can it be to tweak the algorithm. It's clearly not an ML algorithm because that would learn and adapt, so it's a bunch of values somewhere in the code. Perhaps there is some bottleneck where lowering the resolution wouldn't boost the frame rate. ¯\_(ツ)_/¯.

Or not enough. They could need to lower much more aggressively the resolution like what they have done with AC Valhalla (the game can drop down to 1080p after a patch on XSX vs 1368p on PS5 according to @VGA ). They are maybe thinking it's not worth it and as people can have VRR on XSX...

 

[mr magoo]

This has been a very common theme though. You can pick your poison, resolution or frame rate but what is causing is a bunch of games using a bunch of different engines? The use of dynamic resolution is clear intention that resolution should be sacrificed for frame rate so why do a bunch of mutliplatform games on Series X keep over-reaching on resolution at the expense of hitting 60?

How hard can it be to tweak the algorithm. It's clearly not an ML algorithm because that would learn and adapt, so it's a bunch of values somewhere in the code. Perhaps there is some bottleneck where lowering the resolution wouldn't boost the frame rate. ¯\_(ツ)_/¯.

As you and irobo said the problem may not be gfx related. For example control and outriders had bad frame issues at launch (or before launch) which were patched without sacrificing resolution or gfx fidelity.
It may be matter or bad api calls or problem with tools.
But i fully agree this seems to be a common problem and often games in comparison seems to have more stable frame rates on ps5. All those problems are eventually fixed by patch and this makes me thinks its a software issue some kind. It would be great to have some feedback on this thou.

 

[Dictator]

Or not enough. They could need to lower much more aggressively the resolution like what they have done with AC Valhalla (the game can drop down to 1080p after a patch on XSX vs 1368p on PS5 according to @VGA ). They are maybe thinking it's not worth it and as people can have VRR on XSX...

there is only so much I can say - but as far as I know (and I have it on reliable source) the reason why Xbox Series X and S are getting drops in the open world (like the ones I highlight in the Xbox video very poignantly), is not a GPU load issue to due proper utilisation (aka GPU is overtasked due to what is happening rendering wise), but something else that is in all likelihood going to be patched. But in a place like Yamantau where it is definitely GPU load and proper utilisation, where in that one scene you see Xbox Series X dipping to 1134p and hitting around 57-58 fps while PS5 is 1015p at 60 fps, XSX could use a small res drop (to 1080p probably) to clear up that small instability for sure. But then again, with VRR that is completely invisible... so...

 

[Allandor]

there is only so much I can say - but as far as I know (and I have it on reliable source) the reason why Xbox Series X and S are getting drops in the open world (like the ones I highlight in the Xbox video very poignantly), is not a GPU load issue to due proper utilisation (aka GPU is overtasked due to what is happening rendering wise), but something else that is in all likelihood going to be patched. But in a place like Yamantau where it is definitely GPU load and proper utilisation, where in that one scene you see Xbox Series X dipping to 1134p and hitting around 57-58 fps while PS5 is 1015p at 60 fps, XSX could use a small res drop (to 1080p probably) to clear up that small instability for sure. But then again, with VRR that is completely invisible... so...

The major difference is windows in the background . Their thread management was never the best but always good enough. So this can now lead to problems when threads are moved between cores. E.g. to much for one core to handle in time. Although even dx12 has still a higher CPU usage overhead than e.g. Vulkan. Dropping res in those situations might just not help. IMHO the Xbox has the faster CPU, but that only helps so much that they can at best getting the same speed for API calls than Sony.

 

[Jay]

But then again, with VRR that is completely invisible... so...

And this is why I think we need some additional options.
Xbox has pushed VRR as a technology, so it's a shame when it's not taken advantage of when available.
Meaning I would like it not to lock to 60 but maintain a higher resolution.

 

[Dictator]

The major difference is windows in the background . Their thread management was never the best but always good enough. So this can now lead to problems when threads are moved between cores. E.g. to much for one core to handle in time. Although even dx12 has still a higher CPU usage overhead than e.g. Vulkan. Dropping res in those situations might just not help. IMHO the Xbox has the faster CPU, but that only helps so much that they can at best getting the same speed for API calls than Sony.

It has nothing to do with that at all - rather something else entirely and which is a Bug. Think about it - on Real Windows PCs with something like a 2600 such massive frametime Spikes do not exist. The Extra OS layer Stuff on Xbox is not suddenly going to cause 66+ MS Spikes, be reasonable.

 

[snc]

I keep seeing this posted in comments that they didn't use any of the next gen IO. Have the devs confirmed this? I understand that the loading time scales pretty well to the clock speed of the CPUs found in the 3 consoles, but could it be that the data must be processed or set up. Something like compiling the BVH or something.

Also, the PS5 version is larger? That's a first for this gen isn't it? I guess that could point towards the data being not compressed in a format compatible with PS5's IO block.... Or it could mean that both PS5 and XSX versions are compressed in native formats and the Xbox version just compressed a bit better.

no confirmation by devs just obvious deduction (df thinks same)

 

[BRiT]

Now that I have a PS5 I might buy this one. It's being a long time since I haven't played a decent Start Wars (Gamecube, Rogue leader).

It's part of EA Play, so that's one option to possibly spend less like get a trial or 1 month or 1 year membership and you get access to so many other games too.