当前位置:网站首页>Ue5 uses DLSS (super sampling) to improve the FPS of the scene away from the optimization scheme of Caton
Ue5 uses DLSS (super sampling) to improve the FPS of the scene away from the optimization scheme of Caton
2022-07-27 20:56:00 【Blindfold to see the world】
One 、 What is a graphics card DLSS technology ?
from : What is a graphics card DLSS technology ? - You know
DLSS yes Nvidia One of the most innovative features released in recent years . It was originally designed to compensate for the loss of graphics card performance when using ray tracing , With this unique AI technology ,DLSS It can significantly improve the performance of the game while maintaining the display quality , Because of this , Released in subsequent iterations DLSS 2.0 Also become NVidia An independent function of the graphics card .
1、 to the end DLSS What technology is it ?
DLSS Full name Deep Learning Super Sampling( Deep learning supersampling ). It is a video rendering technology ,Nvidia Using neural network supercomputers , Train their AI through deep machine learning , The specific way is to constantly let AI automatically compare ultra-high resolution still images (SSAA Super sampling anti aliasing version ) And low resolution images , Make artificial intelligence have the ability to infer the method to improve the image resolution from low to high . in other words , This technology renders graphics at a lower resolution to maintain high performance , Then apply various effects to output the overall effect of high-resolution image .
This technology is needed hfr ( Greater than 60 FPS) and high The resolution of the (4K) Running games are especially useful . for example , Use DLSS function , With 1080p Resolution rendering game frames , This load makes it easier to achieve a higher frame rate , Then zoom in to 4K Resolution output , So as to bring Ultra HD image display quality .
DLSS Technology depends on RTX Special hardware of graphics card Tensor The core . So at present, only RTX 20 Series and RTX 30 Series has this function . however RTX 30 The series is configured with Nvidia Second generation Tensor The core ,DLSS Better performance .
2、DLSS Performance improvements
DLSS The main advantage of technology is that it can improve the frame rate of the game . How much performance improvement can be achieved depends on the original resolution 、GPU、DLSS Specific settings . Users can choose different DLSS Pattern , such as quality 、 Balance and Performance mode .
Use Quality model , You can get the best visual quality and the lowest performance improvement , Performance mode Can bring up to 70% Performance improvement of , But the visual quality will be greatly reduced , Balance mode The options are in between .
3、DLSS Development stage
DLSS 1.0 And Nvidia Of NGX Use with supercomputers , For training based on AI Rendered RTX Video card Tensor The core .
AI After the network completes the training of the new game ,AI The model will be distributed to RTX In the video card . such ,RTX Video card Tensor The core has the same AI function , As long as the game supports DLSS function , Can play DLSS Performance of .
DLSS 1.0 The performance can be improved about 2 times , But many games didn't support .
2020 year 3 month ,Nvidia Released DLSS 2.0, This is a DLSS Updated version of , It uses a new deep learning neural network , Just render 25-50% The pixel (DLSS 2.1 Super performance mode requires only rendering 11% The pixel ), And use new time feedback technology , And DLSS 1.0 Compared with clearer details and better stability . Because it can be used more effectively Tensor The core , So the speed ratio DLSS 1.0 It's getting close 2 times , It can achieve higher frame rate and eliminate pair GPU、DLSS Setting and resolution limits .
because DLSS 2.0 Not through a single game AI Training , Its implementation method is more general , in other words DLSS 2.0 Of AI The quality of the algorithm can be continuously improved with continuous training , therefore DLSS 2.0 Better compatibility with games .
2020 year 9 month ,Nvidia Released DLSS 2.1, Added for ultra high resolution Games Super performance mode 、 Yes VR Game support and support dynamic resolution . As long as the game's rendering engine supports dynamic resolution , It allows the source frame to dynamically change the resolution , The output image size remains unchanged .
4、DLSS 2.0 Optional mode
DLSS 2.0 share 3 Optional image quality modes : quality 、 Balance and performance ,DLSS 2.1 Also added Super performance Pattern . performance Mode improves performance but also reduces image quality .
Use 2.0, performance Mode improves performance most significantly , The game can be changed from 1080p Upgrade to 4K, This is a 4 Zoom in (2 Times the width and 2 Times the height ). Balance The mode reaches 3 Zoom in , and quality Mode use 2 Zoom in .DLSS 2.1 Introduced Super performance The pattern is 16 Zoom in , You can configure RTX 3090 The machine of the graphics card is 8K The resolution of the (7680 x 4320) Run the game .
DLSS The iteration of makes it possible to display better image quality in the image area with moving objects . In the fan blade image above DLSS 2.1 This version of rendering looks better than using DLSS 1.0 The image of is much better .
5、 Which graphics cards support DLSS
Nvidia Of RTX 20 Series and 30 The series fully supports DLSS.
- GeForce RTX 2060
- GeForce RTX 2060 Super
- GeForce RTX 2060 12GB
- GeForce RTX 2070
- GeForce RTX 2070 Super
- GeForce RTX 2080
- GeForce RTX 2080 Super
- GeForce RTX 2080 Ti
- GeForce RTX 3060
- GeForce RTX 3060 Ti
- GeForce RTX 3070
- GeForce RTX 3070 Ti
- GeForce RTX 3080
- GeForce RTX 3080 Ti
- GeForce RTX 3090
RTX 30 series DLSS Better performance than the RTX 20 series , because RTX 30 Series of graphics cards are equipped with the second generation Tenso The core , It can provide higher single core performance .
6、DLSS The future development of Technology
DLSS Technology may eventually become Nvidia RTX One of the most influential functions of the graphics card . The ray tracing effect is very beautiful , But this function consumes the computing power of the graphics card , It often has a considerable impact on the performance of the graphics card , and DLSS At present, technology is not as powerful as we hope , but DLSS It provides us with a balanced solution , Make the game run with higher resolution and higher frame rate .
DLSS When it was first released, it seemed to be a niche function of low-end graphics cards , But that's not the case . contrary ,DLSS It can also improve cyberpunk on high-end graphics cards 2077 And control the visual quality of games , While maintaining a high frame rate .DLSS Technology not only improves the performance of low-end graphics cards , At the same time, it has also become the future development direction of high-end graphics cards .
NVidia Of RTX 3090 The video card has 24GB Explicit memory of , When raytracing and DLSS When the function , Can be 8K Resolution rendering 《 Wolfenstein :Young Blood》 Such games . Even though 8K Resolution is far from universal , but 4K However, the display of has become more and more popular . Gamers can 1080p or 1440p Rendering , And use DLSS Output 4K High resolution and keep 50-60fps, That is, higher frame rate without obvious loss of image quality .
DLSS Technology will continue to iterate , At the same time, there will be more games and DLSS Technical compatibility .
7、AMD Is there a similar technology ?
AMD stay 2021 In, a new solution was announced called FidelityFX Super Resolution abbreviation FSR, This is for NVIDIA Of DLSS( Deep learning supersampling ) Technical response .
Nvidia Of DLSS and AMD Of FSR All help the graphics card balance between performance and resolution , But these two technologies are implemented in completely different ways .NVidia DLSS Technology depends on Neural Networks 、 Supercomputer processing and artificial intelligence , and FidelityFX Use software algorithms to achieve similar results . in addition DLSS Technology can only be owned Tensor The core RTX Work on the graphics card , and AMD Based on software implementation , therefore FSR Compatible with various graphics options .
Two 、UE5 combination DLSS Solutions for
2.1 download DLSS Shanyueji sharing
2.2 Will download DLSS Copy to the engine directory

2.3 open UE5 The project of , Choose to install DLSS plug-in unit :

2.4 Then restart the project , Set parameters :

Quality For the highest quality .
Run to see the effect ,end......
边栏推荐
- R语言使用epiDisplay包的lroc函数可视化logistic回归模型的ROC曲线并输出诊断表(diagnostic table)、可视化多条ROC曲线、使用legend函数为可视化图像添加图例
- Jetpack compose performance optimization guide - compilation metrics
- SQL高级技巧CTE和递归查询
- 洋葱集团携手OceanBase实现分布式升级,全球数据首次实现跨云融合
- 【Dart】一门为跨端开发而生的编程语言
- Programmer growth Chapter 18: project launch
- Software test interview question: count the number in a queue, how many positive numbers and how many negative numbers, such as [1, 3, 5, 7, 0, -1, -9, -4, -5, 8]
- Innovative cases | the growth strategy of digitalization of local life services and upgrading of Gaode brand
- Arduino开发(二)_基于Arduino UNO开发板的RGB灯光控制方法
- UE5使用DLSS(超级采样)提升场景的 FPS 远离卡顿的优化方案
猜你喜欢

Why are TS slice files generated when easycvr platform turns off video recording?

82.(cesium篇)cesium点在3d模型上运动

API Gateway介绍
![[dart] a programming language for cross end development](/img/e1/1167a322bb9f276f2e00fb12414d17.png)
[dart] a programming language for cross end development

品牌列表案例

Kingbasees heterogeneous database migration guide (4. Application migration process)
![[Numpy] 数组索引和切片](/img/ce/34db7aef3fefe8a03e638d0838492f.png)
[Numpy] 数组索引和切片

The variable "lattice" or class "lattice.latticeeasy" (matlab) is not defined

推荐一款强大的搜索工具Listary

JVS基础介绍
随机推荐
MYSQL设计优化生成列
[efficiency] abandon notepad++, this open source substitute is more awesome!
Where is the program?
UE5使用DLSS(超级采样)提升场景的 FPS 远离卡顿的优化方案
JVS基础介绍
Hcip day 5
JVs basic framework function list
一文了解Pycharm快捷键
用户登录切换案例
Laboratory management system implemented by SSM framework +jsp [source code + database + system paper]
Hexagon_V65_Programmers_Reference_Manual(6)
Express WEB服务器的简单使用
Recognized by international authorities! Oceanbase was selected into the Forrester translational data platform report
vi工作模式(3种)以及模式切换(转换)
JVs privatization deployment start failure handling scheme
Slim: self supervised point cloud scene flow and motion estimation (iccv 2021)
从0开始写bootloader
Software test interview question: given a queue, such as: [1, 3, 5, 7], how to put the first number into the third position to get: [3, 5, 1, 7]
用户组织架构的管理
One week activity express | in simple terms, issue 8; Meetup Chengdu station registration in progress