Streaming

章节:

This section contains settings related to resource streaming.本节包含与资源流有关的设置。

Streaming settings

流媒体设置

First Frames Force	number of frames to load immediately after the first frame. It will make the scene appear right at the application start-up. 数值范围: [0, INT_MAX]. 默认值为 : 60. 通过控制台访问: render_streaming_first_frames_force (API控制)
Max Threads	maximum number of threads used for streaming. Higher number of threads results in faster streaming, but may cause spikes in case of excessive consumption of GPU resources. 数值范围: [1, 256]. 默认值为 : 1. 通过控制台访问: render_streaming_max_threads (API控制)
Particles Memory Limit	cache memory limit for vertices of particle systems, in percentage of the total GPU memory. 注意 Setting a too low limit for a huge number of particle systems in the scene may lead to rendering only some of them. 数值范围: [0, 100]. 默认值为 : 3. 通过控制台访问: render_streaming_particles_memory_limit (API控制)

Preload at World Loading#

Shaders	通过控制台访问: render_shaders_preload (API控制)

BudgetsBudgets#

Loading	time limit for loading graphics resources, in milliseconds per frame. Increasing the limit leads to increasing the streaming performance. However, memory consumption also increases due to a decrease in the resource loading speed. 数值范围: [0.0f, inf]. 默认值为 : 1.0f. 通过控制台访问: render_streaming_budget_loading (API控制)
Textures Destroy	time limit for deleting textures, in milliseconds per frame. The Engine will distribute the textures cleanup process by the required number frames in order to not exceed the time limit. The number of frames will also depend on the amount of content in the scene and the computing capacity. Increasing the limit leads to increasing the streaming performance. However, memory consumption also increases due to a decrease in the texture unloading speed. 数值范围: [0.0f, inf]. 默认值为 : 0.1f. 通过控制台访问: render_streaming_budget_destroy_textures (API控制)
Meshes Destroy	time limit for deleting meshes, in milliseconds per frame. The Engine will distribute the meshes cleanup process by the required number frames in order to not exceed the time limit. The number of frames will also depend on the amount of content in the scene and the computing capacity. Increasing the limit leads to increasing the streaming performance. However, memory consumption also increases due to a decrease in the mesh unloading speed. 数值范围: [0.0f, inf]. 默认值为 : 0.1f. 通过控制台访问: render_streaming_budget_destroy_meshes (API控制)

注意

定期检查图形资源是否被修改，以便重新加载或删除。

TexturesTextures#

Streaming Mode	streaming mode for textures. The following modes are available: Async - asynchronous loading of textures. Force - force-loading of textures required for each frame at ones. 选项 #1 默认情况下选择（见上文）。 (API控制)
Cache Resolution	resolution for texture cache elements. These minimized copies of textures are used instead of the originals while the latter are loaded. 以下值之一： 8x8 16x16 (默认) 32x32 64x64 128x128 256x256 512x512 通过控制台访问: render_streaming_textures_cache_resolution (API控制)
Memory Limit	cache memory limit used for textures streaming. This is a recommended value to aim for. So, it can be exceeded if textures are required to render the current frame. The memory limit is associated with the lifetime: textures are deleted from video memory only when both values are exceeded. 数值范围: [0, 100]. 默认值为 : 65. 通过控制台访问: render_streaming_textures_memory_limit (API控制)
Life Time	lifetime of GPU cache used for textures rendering. The engine deletes textures after this time only if the specified memory limit for textures streaming is also exceeded. 数值范围: [1, 60]. 默认值为 : 4. 通过控制台访问: render_streaming_textures_life_time (API控制)

Meshes GPUMeshes GPU#

Streaming Mode	streaming mode for loading meshes to video memory (VRAM). The following modes are available: Async - asynchronous loading of meshes. Force - force-loading of meshes required for the current frame at once. All - loading all meshes available in the project on the application start-up. This mode is good for small projects with few meshes. 选项 #1 默认情况下选择（见上文）。 (API控制)
Memory Limit	memory limit used for loading meshes to video memory (VRAM). This is a recommended value to aim for. So, it can be exceeded if meshes are required to render the current frame. The memory limit is associated with the lifetime: meshes are deleted from VRAM only when both values are exceeded. 数值范围: [0, 100]. 默认值为 : 4. 通过控制台访问: render_streaming_meshes_limit_vram (API控制)
Life Time	lifetime of meshes in video memory since the last time they were accessed. The engine deletes meshes after this time only if the specified VRAM limit is also exceeded. 数值范围: [0, 60]. 默认值为 : 4. 通过控制台访问: render_streaming_meshes_life_time_vram (API控制)

Meshes CPUMeshes CPU#

Streaming Mode	streaming mode for loading meshes to memory (RAM). The following modes are available: Async - asychronous loading of meshes. Force - force-loading of meshes required for the current frame at once. All - loading all meshes available in the project on the application start-up. This mode is good for small projects with few meshes. 选项 #1 默认情况下选择（见上文）。 (API控制)
Memory Limit	memory limit used for loading meshes to memory (RAM). This is a recommended value to aim for. So, it can be exceeded if meshes are required for the current frame. The memory limit is associated with the lifetime: meshes are deleted from RAM only when both values are exceeded. 数值范围: [1, 100]. 默认值为 : 5. 通过控制台访问: render_streaming_meshes_limit_ram (API控制)
Life Time	lifetime of meshes in memory since the last time they were accessed. The engine deletes meshes after this time only if the RAM limit is also exceeded. 数值范围: [1, 60]. 默认值为 : 20. 通过控制台访问: render_streaming_meshes_life_time_ram (API控制)

Prefetch CPUPrefetch CPU#

Collision Mode	mode of asynchronous pre-loading of meshes to memory before they are used. Pre-loading is available only for meshes, which have at least one surface with the Collision flag set. There are 3 modes of loading such meshes to RAM: Disable - loading is disabled. Radius - meshes within the prefetch radius are loaded. Full - all meshes with the Collision flag are loaded. This method should be used when the Async streaming mode for meshes is set. 选项 #1 默认情况下选择（见上文）。 (API控制)
Intersection Mode	mode of asynchronous pre-loading of meshes into memory before they are used. Pre-loading is available only for meshes, which have at least one surface with the Intersection flag set. There are 3 modes of loading such meshes to RAM: Disable - loading is disabled. Radius - all meshes within the prefetch radius are loaded. Full - all meshes with the Intersection flag are loaded. This method should be used when the Async streaming mode for meshes is set. 选项 #1 默认情况下选择（见上文）。 (API控制)
Radius	radius within which meshes are pre-loaded into memory. The value should exceed the physics radius (for collisions) and/or the radius within which intersections are calculated. 数值范围: [0.0f, inf]. 默认值为 : 0.0f. 通过控制台访问: render_streaming_meshes_prefetch_radius (API控制)

For OpenGL OnlyFor OpenGL Only#

The settings below are available for OpenGL API only.以下设置仅适用于OpenGL API。

注意

这些参数应针对每种特定情况和配置进行仔细调整，因为它们很大程度上取决于所使用的硬件和驱动程序。

Async Buffer	Size of an intermediate buffer (between the CPU and new resource) used for mesh and texture streaming, in Mb. The size of this buffer must be equal to the size of the largest resource (mesh/texture), otherwise in case of a larger resource, the buffer will be resized causing a spike. 用于网格和纹理流传输的中间缓冲区（在CPU和新资源之间）的大小，以Mb为单位。此缓冲区的大小必须等于最大资源的大小（网格/纹理），否则，在资源较大的情况下，将调整缓冲区的大小，从而导致峰值。注意 Be aware, that the size of this intermediate buffer will be added to total memory consumption.请注意，此中间缓冲区的大小将添加到总内存消耗中。
Async Buffer Indices	Size of an intermediate buffer, analogous to the Async Buffer above, used for mesh streaming (to store vertex indices of meshes), in Mb.类似于上面的Async Buffer的中间缓冲区的大小，以Mb为单位，用于网格流传输（存储网格的顶点索引）。注意 Be aware, that the size of this intermediate buffer will be added to total memory consumption.请注意，此中间缓冲区的大小将添加到总内存消耗中。
Async Buffer Synchronization	Enable buffer synchronization for transferring data from the streaming thread to the main one. When disabled, both async buffer and async buffer for indices are created anew each time. This reduces the number of buffer synchronizations but increases the number of memory allocations.启用缓冲区同步以将数据从流线程传输到主线程。禁用后，每次都会重新创建索引的异步缓冲区和异步缓冲区。这减少了缓冲区同步的数量，但增加了内存分配的数量。注意 Sometimes (depending on the hardware/driver used, e.g. when the main thread is affected by sychronization primitives in other threads) memory allocation may be faster than synchronizations, in such cases, when streaming becomes unacceptably slow, it is recommended to disable buffer synchronization.有时（取决于所使用的硬件/驱动程序，例如，当主线程受到其他线程中的同步原语影响时），内存分配可能比同步要快，在这种情况下，当流变得令人无法接受的慢时，建议禁用缓冲区同步。

最新更新: 2023-12-19

Help improve this article

Was this article helpful?

Report a problem related to this article

(or select a word/phrase and press Ctrl+Enter)