Shulou(Shulou.com)06/03 Report--

This article introduces the relevant knowledge of "skills related to Unity project optimization". In the operation of actual cases, many people will encounter such a dilemma, so let the editor lead you to learn how to deal with these situations. I hope you can read it carefully and be able to achieve something!

Catalogue

1. Simple optimization

Second, optimize the advanced level

Third, code-related optimization

1. Simple optimization

1. Try to make the texture size 2 to the power of 2 (such as 128, 512, 1024, etc.) so that it can be optimized by unity.

two。 Combine multiple small image textures with a large image texture to reduce drawcall, especially the elements of the same ui panel are put into one atlas, and the background image is made as a separate picture as possible.

3. Share shaders as much as possible for the same reason.

4. Models that do not need to play the animation remove the Animator or animation components because they will be initialized if they do not work or do not use.

5. In the aspect of the model, we should pay attention to reduce the number of vertices and triangles as much as possible, and reuse vertices as much as possible.

6. Texture mipmap switch depending on the situation (generally used in 3D scenes, UI and 2D games), if turned on, it can reduce the jitter and flicker when the texture is overlaid, speed up rendering and reduce image aliasing; but it will take up more memory space and make the project's packaged files larger (after all, a series of layer-by-layer smaller backup images are generated).

7. Set the objects in the scene that will not change (displacement, rotation, etc.) to Static to facilitate unity for static batch processing.

8. Turn on static batch processing (Static Batching) and dynamic batch processing (Dynamic Batching) in player setting: static batch processing needs to check the Batching Static of objects that do not need to be changed in the scene, otherwise it will not have an optimization effect. Note that the use of combined batch processing only reduces the number of transfers from CPU to GPU and SetPassCall and does not need to perform vertex change operation again, but it will increase packet body and memory consumption.

9. Spotlight is very expensive in lighting and should be used as little as possible.

10. Use real-time shadows as little as possible, and use Hard Shadow if necessary.

11. Be careful to use translucent materials and UI to reduce overdraw. For example, image with alpha 0 is not visible, but gpu will still draw it. If there are many translucent interfaces in UI, it is recommended that UI be rendered by a separate camera, and UI should not be superimposed into the rendering range of the scene camera.

Second, optimize the advanced level

1. Use buffer pools to manage object that needs to be frequently seen and destroyed.

two。 Use lod technology as appropriate, which requires multiple sets of models, so it will make the packaged files of the project larger and take up more memory, but it can reduce the number of vertices that need to be drawn, and sudden changes in the model may occur if the distance adjustment is not good.

3. Use the OC (occlusion culling) function according to the situation (usually used when there are a large number of models in the scene). Using this feature can greatly reduce the overdraw (after all, the occluded object is not rendered directly), and you need to pay attention to achieving a balance between the size of the object and the size of the cell; you can control the object to be occluded or occluded by setting its Occluder Static and Occludee Static. The disadvantage of this technique is that it consumes extra CPU.

4. Use light mapping with light probe group to replace real-time lighting. Use this technology to require higher requirements on the model, there can be no UV confusion, normal confusion, otherwise there will be a very strange situation; at the same time, there will be no dynamic shadow generation.

5. Compress the texture for different platforms (for example, IOS is pvrtc;, Android is ETC1, and RGBA 16-bit is used if there is an alpha channel). It should be noted that the quality of the texture is sometimes affected after compression.

6. Different UI interfaces use multiple Canvas, because unity merges all elements under canvas into one Mesh for performance optimization; if all interfaces are put into one canvas, it will cause redrawing when the interface changes, reducing performance. The best solution is to put a static ui into a canvas, and dynamic ui into different canvas according to the situation. Do not divide them into details or it will cause an increase in dc.

7. The skinned mesh animation is changed to vertex animation, exchanging space for time.

8. If the mobile terminal has no specific requirements for rendering frame rate, you can save machine performance by reducing the frame rate. Note that it cannot be set to Don't Sync in QualitySettings.vSyncCount, otherwise the following settings will be invalid, as follows:

Add the following code to the Awake or Start in the project configuration class or main class

Application.targetFrameRate = 30; / / the mobile platform defaults to 30 as the power-saving frame rate, and sets it to 300 to achieve the best performance (so that the machine will try its best to increase the frame rate, but the probability is less than 300)

9. Reasonable use of Profiler for performance bottleneck analysis and optimization.

10. Convert Terrain to Mesh, and large maps use streaming load and unload technology.

11. The skinned mesh animation of the model is changed to vertex animation, exchanging space for time.

twelve。 Optimize a large number of model animations using Animation Instancing.

13. Replace MeshCollider with BoxCollider whenever possible.

Third, in-depth optimization (complicated to operate, requiring some low-level knowledge or situations requiring extreme optimization)

1. Use a custom shader

two。 Merge mesh (here you can learn about the blogs written by Boss Lin)

Delete the GraphicRaycaster component from the Canvas that does not involve X-ray detection in 3.UI

4. Try to use less LayoutGroup and contentSizeFitter, although it is more convenient to use, but in the user operation will lead to the entire ui interface redrawing, capable Daniel suggested that they write an algorithm to achieve local redrawing, do not have the ability only in the use of these two components of the place to use a separate Canvas to avoid the entire UI redrawing.

5. Use single-channel instanced rendering as appropriate: to enable this feature, turn on the Player setting (select Edit > Project Settings__, and then select the Player__ category). In the Player settings, navigate to the bottom XR Settings panel, select the Virtual Reality Supported option, and then select Single Pass Instanced (Preview) from the Stereo Rendering Method drop-down menu.

Third, code-related optimization

1. Use for loop instead of foreach,

two。 Replace gameObject.tag= "tag" with gameObject.CompareTag ("tag").

3.Find and GetComponent methods are called in Start and Awake as far as possible, not in methods such as Update.

4. Replace + concatenation strings with string.format.

5. Physical testing is carried out in fixedUpdate.

This is the end of the content of "skills related to Unity Project Optimization". Thank you for reading. If you want to know more about the industry, you can follow the website, the editor will output more high-quality practical articles for you!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.