Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
Finished Stray at 1080p120 on an RX 5700 XT + R5 3600 (with UE4 TSR and 67% rez scale enabled in ini).
Disable RT in ini also if you have a DX12U GPU.
What are you talking about - there's a MULTITUDE of videos that prove that Dx12 runs WORSE and has more frame stutter than Dx 11.
https://www.youtube.com/watch?v=fIk6x8ZJ-oc
I even tested this myself back when I first got the game.
Stick to Directx 11.
Basically you're never CPU bound in DX12 in UE4 titles.
In Stray, the game absolutely will randomly become CPU bound in DX11, on my 5700 XT + R5 3600. Fps would drop from 120 to 70-90 even.
In DX12? 120 fps locked, not even a blip of a CPU bound scenario.
UE4 DX12 doesn't have judder. CPU load is LOWER at both same fps or higher fps than DX11, not higher. And I said GPU performance should be identical in a GPU bound scenario.
Something haywire is going in that vid.
Your statement about Dx 12 being better than Dx 11 is only half right, for the games that PROPERLY support it and utilise it. Resident Evil 2 remake was BRUTALISED by its players for having worse performance in Dx 12 than Dx 11:
https://www.reddit.com/r/pcgaming/comments/vkfn5x/digital_foundry_resident_evil_2_remake_pc_patch/
...so much so it allowed users to implement the Dx 11 Patch through using a beta patch through Steam.
https://www.youtube.com/watch?v=hEJICLXaHEs (skip to 3:06)
Also, when Hitman (2016) was released and its episodes were released episodically, you'd have to benchmark both Dx 11 and Dx 12 to find out which one was smoother and it would change from episodic patch to patch because a lot of the time they're break something with Dx 12 and you'd have to rely on Dx 11 till they fixed whatever was wrong with Dx 12.
You are technically right - Dx 12 is MEANT to run smoother, but it doesn't always.
At launch, this was true. This also is no longer relevant today. DX12 in Resident Evil games is now 1:1 as performant as DirectX11.
Hitman situation was different. Early DX12 release, iffy driver support, no Async Compute on Nvidia. Now DX12 is absolutely superb performance wise on all GPUs in Hitman 3.
So basically DX 12 in Resident evil games is essentially just as good as DX 11... that's not really a selling point, and you're not really winning your case here.
If DIrectx 12 is only giving 100% the performance as Directx11 - what good is Directx 12 then?
Here's another example of practically no difference in performance (Witcher 3):
https://youtu.be/4EYGa3R0Guo?t=421
I'll also admit - I don't have Hitman 3 installed anymore, so I can't benchmark Dx 11 vs 12 but a lot of games don't offer better performance in one direct over the other unfortunately.
As I said, I don't think it's meant to be that way - put it down to poor optimisation and coding on the developers part. Maybe they constant dumbing down of PC games because they were ported from other consoles has to play a part in that, but the problem is that we're not really seeing any games with major performance benefits due solely because they are Directx 12 games.
Plus, I think Vulkan is generally superior in many cases to Dx 12 anyway, so that doesn't really help their case. Doom Eternal dropped Dx entirely.
Anyway, here's a recent video from someone - 2 weeks old testing the games.
https://www.youtube.com/watch?v=HgGod-K0Vy8
Long story short:
Battlefield V: Dx 11 on par with 12 (1% negligible diffence)
Borderlands 3: Slight win to Dx 11 (so again, 2-3% better, so nothing impressive)
Fortnite: Dx 11 BEATS THE ♥♥♥♥ out of Dx 12.
Hellblade: Dx 12 for the first time beats Dx 11 (possibly good optimisation and coding from the developers)
PUBG: Dx 11 whoops the ♥♥♥♥ out of Dx 12 (20-25% better than Dx12)
Resident Evil: As you said 1:1 performance with both Dx 11 or 12.
SotTR: Again, better on Dx 12, but could be brilliant optimisation.
I'm not saying you're wrong. Directx 12 SHOULD be better than Directx 11, but in most of the examples we have, it either is on par, or is beaten by Dx 11.