Trait PushKernelArg 

pub unsafe trait PushKernelArg<T> {
    // Required method
    fn arg(&mut self, arg: T) -> &mut Self;
}

Something that can be copied to device memory and turned into a parameter for result::launch_kernel.

See the cuda docs on argument processing.

Safety

This is unsafe you need to ensure that T can be represented in CUDA and also references to it can be properly passed to CUDA.

Most implementations will be required to use #[inline(always)] to ensure that references are handled properly.

Required Methods

fn arg(&mut self, arg: T) -> &mut Self

Dyn Compatibility

This trait is not dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementors

impl<'a, 'b, 'c, T> PushKernelArg<&'b CudaView<'c, T>> for LaunchArgs<'a>

where 'b: 'a, 'c: 'b,

impl<'a, 'b, 'c, T> PushKernelArg<&'b UnifiedView<'c, T>> for LaunchArgs<'a>

where 'b: 'a, 'c: 'b,

impl<'a, 'b, 'c, T> PushKernelArg<&'b UnifiedViewMut<'c, T>> for LaunchArgs<'a>

where 'b: 'a, 'c: 'b,

impl<'a, 'b, 'c, T> PushKernelArg<&'b mut CudaViewMut<'c, T>> for LaunchArgs<'a>

where 'b: 'a, 'c: 'b,

impl<'a, 'b, 'c, T> PushKernelArg<&'b mut UnifiedViewMut<'c, T>> for LaunchArgs<'a>

where 'b: 'a, 'c: 'b,

impl<'a, 'b, T> PushKernelArg<&'b CudaSlice<T>> for LaunchArgs<'a>

where 'b: 'a,

impl<'a, 'b, T> PushKernelArg<&'b T> for LaunchArgs<'a>

where 'b: 'a, T: DeviceRepr,

impl<'a, 'b, T> PushKernelArg<&'b UnifiedSlice<T>> for LaunchArgs<'a>

where 'b: 'a,

impl<'a, 'b, T> PushKernelArg<&'b mut CudaSlice<T>> for LaunchArgs<'a>

where 'b: 'a,

impl<'a, 'b, T> PushKernelArg<&'b mut UnifiedSlice<T>> for LaunchArgs<'a>

where 'b: 'a,