Struct ShaderKernel 

pub struct ShaderKernel<T: Send + Sync> { /* private fields */ }

A named, parameterised GPU kernel ready for parallel dispatch.

Implementations

impl<T: Send + Sync> ShaderKernel<T>

pub fn new( name: impl Into<String>, group_size: usize, f: impl Fn(&ThreadGroupContext, &mut T) + Send + Sync + 'static, ) -> Self

Create a new kernel with an explicit group size.

Panics

Does not panic; group_size = 0 is normalised to 1 at runtime.

pub fn execute(&self, data: &mut [T], work_groups: usize)

Execute the kernel over data using work_groups groups in parallel.

Elements at indices >= work_groups * group_size are silently ignored, matching GPU semantics where excess invocations are masked out.

pub fn group_size(&self) -> usize

The number of threads per work group.

pub fn name(&self) -> &str

The human-readable label for this kernel.