Creates a single kernel object using a Metal Shading Language (MSL) kernel function.
- iOS 11.0+
- macOS 10.13+
- Mac Catalyst 13.0+Beta
- tvOS 11.0+
- Core Image
The name of the function in the Metal shading language.
A metallib file compiled with the Core Image Standard Library.
This method allows you to use MSL as the shader language for a Core Image kernel. Since MSL based kernels are precompiled, initializing them is faster than their than Core Image Kernel Language (CIKL) counterparts and Xcode can provide error diagnostics during development rather than at runtime. MSL is a more modern language than CIKL, and you can write shader code that uses arrays, structs and matrices.
MSL based kernels still support concatenation and tiling and can work in the same filter graph as traditional CIKL kernels.
Specifying Compiler and Linker Options
To use MSL as the shader language for a
CIKernel, you must specify some options in Xcode under the Build Settings tab of your project's target. The first option you need to specify is an
-fcikernel flag in the Other Metal Compiler Flags option. The second is to add a user-defined setting with a key called
MTLLINKER with a value of
Creating a General Kernel in Swift
The following code shows how you can create a general kernel based on a Metal function named
The first step is to create a
Data object that represents the the default Metal library. If you have built this in Xcode, it will be called
default and can be loaded using the Bundle type's
Using the representation of the Metal library and the function name
my, you initialize a
The code to create general, color, warp and blend filters is identical.
Metal Shading Language Extensions
Core Image provides a set of language extensions to MSL in
CIKernel. These extensions include three new data types for working with images:
sampler (for accessing the input image),
sample (represents a single color sample from the input image), and
destination (for accessing the output image). The extensions also include convenience functions such as color conversion and premultiply / unpremultiply.
Whereas in CIKL, you typically called global functions when working with, for example, coordinates and samples, these functions are implemented as member functions on the new types.
The following table shows a summary of CIKL global functions and their equivalent MSL functions.
Core Image Kernel Language
Metal Shading Language
Get destination coordinate
Transform coordinate to sampler space
Get destination coordinate in sampler space
Sample from source image
Get extent of source image