-
Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
-
We have made minor adjustments to how the search bar works on ResetEra. You can read about the changes here.
Digital Foundry - Red Dead Redemption II PC: Every Setting Explained (Dictator's Magnum Opus)
- Thread starter RadioactiveLobster
- Start date
Yeah I know, but it was very brief and the final settings weren't confirmed at that point. I generally preferred Alex's videos when he used a 1070 in his system as I think that was far more in-line with what your average PC gamer had and even if you didn't, it was easier to extrapolate.
I get having the 2080ti in there, I want to see its performance as what's possible, but it's so out of the range of anyone but the most hardcore enthusiasts, and also has more vram than any other card, so in a video focusing on what you can get with the 'PC' and how it relates to a console, it's a little too tightly focused on hardware the vast, vast majority don't have.
I finally got this game working after updating my BIOS and completely redoing all my computer's settings afterwards. With my 2080 Ti I'm able to get it mostly all ultra and still get 4K60, so I'm happy.
As I was saying in the other topic, texture VRam requirement doens't make any sense and need to be fixed.
Ultra quality (Xbox One X setting) takes 6-8GB of VRam... and that's mental, especially if you think that Xbox One X use 8GB of shared Ram, so both Ram and VRam.
This NEED to be fixed by R*.
Ultra quality (Xbox One X setting) takes 6-8GB of VRam... and that's mental, especially if you think that Xbox One X use 8GB of shared Ram, so both Ram and VRam.
This NEED to be fixed by R*.
Lovely video, although I could personally care less about "Optimised Xbox One X settings". Would like it even more if this was a side by side comparison between Low and Ultra and the FPS impact on each, then it'd be up to the viewer to choose the best settings for his/her scenario. The summary at the end of the video with the settings to reflect X1X IQ is great though.
I do understand why his videos use such template, not complaining, just my personal opinion.
I do understand why his videos use such template, not complaining, just my personal opinion.
I'm going to use this video as a guide to get the highest settings I can @1080p/60fps on my 9900k/1080 Ti.
Yeah, I get that
Should be easily done tbf
I'm holding out for the Steam version. Hopefully, there will be a patch or two by then to iron some of the issues out.
I mean ultimately it's better with more granularity but yeah I get it sometimes I just wanna set a preset and go.
And the preset slider seems a bit lacking in this.
Using the settings in the attached image is the baseline on PC if you have similar hardware. Better hardware with these settings will give you a better image at 60+ FPS depending on how much better your hardware is. Some of the settings marked as "Low" are even lower on Xbox 1X, but DF wasn't sure exactly how.
I would also check out the Gamers Nexus GPU video. With their High settings the 1080 Ti was getting 97fps at 1080p. You can use that to tweak things up to Ultra from there.
I'm a big fan of Gamers Nexus. I have that video downloaded as well.
nvidia doesnt perform well in this game. At this point people shouldnt be surprised anymore. Nvidia cards just wont age well in the big titles as long as both consoles use AMD. The more impressive the game is, the worse nvidia typically performs.
This could have an explanation.
Nvidia is superior to AMD in video drivers. Vulkan/DX12 allow developers to bypass most of driver-level optimization to go directly to the "metal" optimization.
On the other hand Nvidia has shown to be able to deliver much better performance on DX11, because their engineers still have a better knowledge of the internal behavior of the hardware.
So when you remove from the equation this driver-level boost that Nvidia still retains, things get equalized with AMD.
Yeah, I am just now messing with it. If I want 60fps at 4k with my 2080ti, I pretty much have to match the Xbox X settings provided here. Anything more and 60FPS is not happening.
I wanna see where I might be going wrong with trying to maintain 60fps at 1440p with mostly ultra, with my 3700x and 2080ti.
I didn't fuck with it a ton, but in my short struggle to 60FPS, TAA was the deciding factor. I went from maxed to off, so there might be an in between that would do it, too. Didn't do a ton of testing.
The game runs quite poorly on the 1000 series sadly, not sure why.
Great video. I run the game without any performance issues on my rig and I'm only annoyed about the constant crashing but your vid highlighted pretty much everything.
Though my personal opinion is the game on ultra settings doesn't differ that much from the Xbox One X if you aren't playing in first person and looking for the obvious differences.
In the end having finished the game on ps4 pro, and playing it now on pc the visual differences are mostly in lighting (though I do remember the resolution being fuzzy on Playstation) as I don't have hdr going from TV to monitor and I'm enjoying the smoothness of the game with higher framerates and Gsync, that and much faster loading times of course.
Though my personal opinion is the game on ultra settings doesn't differ that much from the Xbox One X if you aren't playing in first person and looking for the obvious differences.
In the end having finished the game on ps4 pro, and playing it now on pc the visual differences are mostly in lighting (though I do remember the resolution being fuzzy on Playstation) as I don't have hdr going from TV to monitor and I'm enjoying the smoothness of the game with higher framerates and Gsync, that and much faster loading times of course.
User banned (2 weeks): Trolling over a series of post. History of related behavior.
Because I can.
And the GTX 1080 Ti, specs wise, is faster in basically every single way. How many times must I repeat myself?
The only mechanism we have to measure the relative power of GPUs is by benchmarking it, and different benchmarks run differently on different architectures. We saw on day 1 that Turing thrived on games that that were traditionally said to be AMD favoring - things like Wolfenstein 2, or Strange Brigade or whatever all run great on both Turing and GCN.
At the time of the 2080's release, aggregate game benchmarks put it roughly on par with the 1080ti - a touch slower depending on which specific games you looked at. The 1080ti has slightly more memory bandwidth than the 2080, and it has more overall memory. It has roughly the same FP32 compute. It has more Rops and Cuda cores, but fewer transistors. Architectural differences are what allow the 2080 to compete with the 1080ti despite paper specs where the numbers are higher on 1080ti. Considering transistor counts that doesn't necessarily mean it's much more efficient, but it is different.
Unless we get a patch or driver update tomorrow that totally improves performance on Pascal cards, we won't be able to confirm whether the game is designed in a way that heavily utilizes Turing's differences, or whether it's some random last minute glitch, or because it hasn't had much attention paid to these cards from Rockstar, or whether Nvidia hasn't put much effort into performance on those cards for this game, or whatever. There's lots of things it could be (possibly multiple things in conjunction).
I will say that this game is a very abnormal data point in terms of the sheer difference between Pascal and Turing. Like there's a ~25% advantage that the 2080 has over the 1080ti here, whereas in Wolfenstein 2 (a game famous for running much better on Turing, generally the one with the largest difference when the reviews came out) only has a ~12-15% performance advantage on 2080 over 1080ti.
At the time of the 2080's release, aggregate game benchmarks put it roughly on par with the 1080ti - a touch slower depending on which specific games you looked at. The 1080ti has slightly more memory bandwidth than the 2080, and it has more overall memory. It has roughly the same FP32 compute. It has more Rops and Cuda cores, but fewer transistors. Architectural differences are what allow the 2080 to compete with the 1080ti despite paper specs where the numbers are higher on 1080ti. Considering transistor counts that doesn't necessarily mean it's much more efficient, but it is different.
Unless we get a patch or driver update tomorrow that totally improves performance on Pascal cards, we won't be able to confirm whether the game is designed in a way that heavily utilizes Turing's differences, or whether it's some random last minute glitch, or because it hasn't had much attention paid to these cards from Rockstar, or whether Nvidia hasn't put much effort into performance on those cards for this game, or whatever. There's lots of things it could be (possibly multiple things in conjunction).
I will say that this game is a very abnormal data point in terms of the sheer difference between Pascal and Turing. Like there's a ~25% advantage that the 2080 has over the 1080ti here, whereas in Wolfenstein 2 (a game famous for running much better on Turing, generally the one with the largest difference when the reviews came out) only has a ~12-15% performance advantage on 2080 over 1080ti.
Here goes:
Resolution: 3840x2160
Screen Type: Windowed Borderless
VSync: On
Triple Buffering: On
Texture Quality: Ultra
Anisotropic Filtering: X16
Lighting Quality: Ultra
Global Illumination Quality: Ultra
Shadow Quality & Far Shadow Quality: High
SSAO: High
Reflection & Mirror Quality: Ultra
Water, Volumetrics, Particle, Tessellation, TAA: All High
FXAA & MSAA: Off
Graphics API: Directx 12
Near & Far Volumetric Resolutions: High
Volumetric Lighting Quality: High
Unlocked Volumetric Raymarch Resolution: Off
Particle Lighting Quality: Medium
Soft Shadows: High
Grass Shadows: Medium
Long Shadows: On
Full Resolution SSAO: Off
Water Refraction & Reflection Quality: High
Water Physics Quality: 80%
Resolution Scale: Off
TAA Sharpening: Full
Motion Blue: Off
Reflection MSAA: Off
Geometry LOD: Full
Grass LOD: 80%
Tree, Parallax Occlusion, Decal, Fur Quality: All High
My GPU is also boosted a bit by MSI Afterburner's overclock feature.
The performance advantage using Vulkan in Wolfenstein 2 is around 35% performance advantage in favor of the 2080 over the 1080ti.
15% at 4k is the benchmark I was looking at (PCPer). I think that must represent bandwidth constraints then.
Yeah that's bad. I mean it seems the 1080Ti (Pascal) has been nerfed in this game anyway, but your results are very low.
By comparison, I am running a 1080Ti (oc'ed a good amount 2076Mhz core, +750 on the ram) but with a much worse CPU (6700k @ 4.6) and at 4k with these XBX settings I get (two run average):
Min : 45.211083
Avg : 54.299125
Absolutely no way my fps drops to 32 (did you mean 42?) in the final scene.
Let's be honest, most companies these days cannot be trusted (from my experience at least) it seems like they want to force consumers to upgrade on a yearly cycle, now I don't want to accuse Nvidia of nerfing but I am simply saying, I would not put it past them, just like I wouldn't put it past any company (Samsung with TVs and the auto updates via software which even when you turn them off, still update and you have no ability to roll back the updates and no patch notes to see what changed for example.)
I mean, I just recently updated windows to 1903 and my performance has dropped, 1 min longer for boot ups to windows, lower benchmark scores on 3Dmark both via cpu and gpu with the newest Nvidia drivers, we are only talking about 4-5fps difference but it's still lower and slower than it was before and I can see that when comparing benchmarks, some people might be inclinced to update their hardware when they notice slowdowns, that's all I am saying.
I posted this in the performance thread but I think it fits this thread better.
I was testing the settings and found something interesting. Reflection setting isn't only for the resolution of reflections on windows and stuff like I thought.
You can clearly see the difference in quality on the wooden surface of the wagon. There's also a blue light bounce/reflection on the ceiling of the barn which is more prominent on ultra settings. Global illumination is the same(High) for both screenshots.
ULTRA reflection quality:
LOW reflection quality:
I hope Dictator can explain what's happening here.
I was testing the settings and found something interesting. Reflection setting isn't only for the resolution of reflections on windows and stuff like I thought.
You can clearly see the difference in quality on the wooden surface of the wagon. There's also a blue light bounce/reflection on the ceiling of the barn which is more prominent on ultra settings. Global illumination is the same(High) for both screenshots.
ULTRA reflection quality:
LOW reflection quality:
I hope Dictator can explain what's happening here.
Slightly OT but not really.... over 1 minute Windows boot? That's insane, if you're still on a regular HDD you NEED to go SSD, it'll accelerate everything including games like RDR2. If you already did something else might be wrong. Newer Windows versions have patches against all those CPU vulnerabilities integrated which will hit older CPUs so slight performance degradation is expected if you're coming from an unpatched version but it shouldn't be that bad.
Last edited:
Yup, over 1 minute, moving from 1607 (didn't upgrade for a long time for this very reason but was essentially forced to for RDR2.)
Got it on my SSD already plus I have already disabled spectre because it was hammering my CPU and degrading performance.
I have isolated startup programs as well and they aren't the issue, I honestly cannot figure out what is causing the bootup issues. My bootup time on the older version of windows was 6 seconds from bios to login, so yeah, 6 seconds to 68.5 seconds