forked from BilalY/Rasagar
30 lines
1.9 KiB
Markdown
30 lines
1.9 KiB
Markdown
|
# What's new in SRP Core version 12 / Unity 2021.2
|
||
|
|
|
||
|
|
This page contains an overview of new features, improvements, and issues resolved in version 12 of the Core Render Pipeline package, embedded in Unity 2021.2.
|
||
|
|
|
||
|
|
## Improvements
|
||
|
|
|
||
|
|
### RTHandle System and MSAA
|
||
|
|
|
||
|
|
The RTHandle System no longer requires you to specify the number of MSAA samples at initialization time. This means that you can now set the number of samples on a per texture basis, rather than for the whole system.
|
||
|
|
|
||
|
|
In practice, this means that the initialization APIs no longer require MSAA related parameters. The `Alloc` functions have replaced the `enableMSAA` parameter and enables you to explicitly set the number of samples.
|
||
|
|
|
||
|
|
### New API to disable runtime Rendering Debugger UI
|
||
|
|
|
||
|
|
It is now possible to disable the Rendering Debugger UI at runtime by using [DebugManager.enableRuntimeUI](https://docs.unity3d.com/Packages/com.unity.render-pipelines.core@latest/api/UnityEngine.Rendering.DebugManager.html#UnityEngine_Rendering_DebugManager_enableRuntimeUI).
|
||
|
|
|
||
|
|
## Added
|
||
|
|
|
||
|
|
### High performance sorting algorithms in CoreUnsafeUtils
|
||
|
|
|
||
|
|
New high performance sorting algorithms in the CoreUnsafeUtils helper methods. The new sorting algorithms include:
|
||
|
|
|
||
|
|
* RadixSort - ideal for very large lists, more then 512 elements.
|
||
|
|
* MergeSort (non recursive) - ideal for mid size lists, less than 512 elements.
|
||
|
|
* InsertionSort - ideal for very small lists, less than 32 elements.
|
||
|
|
|
||
|
|
The sorting algorithms only work on uint elements. They include methods that support standard c# arrays, NativeArray objects or raw pointers.
|
||
|
|
RadixSort and MergeSort require support array data, which can be allocated by the user, or allocated automatically via ref parameter passing. InsertionSort is in-place and does not require support data.
|
||
|
|
These algorithms are compatible with burst kernels when using raw pointers or NativeArray. Currently HDRP utilizes them to sort lights in the CPU lightloop.
|