Struct KernelLauncher 

pub struct KernelLauncher<R: Runtime> {
    pub settings: KernelSettings,
    /* private fields */
}

Prepare a kernel for launch.

Fields

settings: KernelSettings

Implementations

impl<R: Runtime> KernelLauncher<R>

pub fn register_tensor(&mut self, tensor: &TensorArg<'_, R>)

Register a tensor to be launched.

pub fn register_tensor_map(&mut self, tensor: &TensorMapArg<'_, R>)

Register a mapped tensor to be launched.

pub fn register_array(&mut self, array: &ArrayArg<'_, R>)

Register an input array to be launched.

pub fn register_scalar<C: CubeScalar>(&mut self, scalar: C)

Register a scalar to be launched.

pub fn register_scalar_raw(&mut self, bytes: &[u8], dtype: StorageType)

Register a scalar to be launched from raw data.

pub fn launch<K: CubeKernel>( self, cube_count: CubeCount, kernel: K, client: &ComputeClient<R>, ) -> Result<(), LaunchError>

Launch the kernel.

pub unsafe fn launch_unchecked<K: CubeKernel>( self, cube_count: CubeCount, kernel: K, client: &ComputeClient<R>, ) -> Result<(), LaunchError>

Launch the kernel without check bounds.

Safety

The kernel must not:

• Contain any out of bounds reads or writes. Doing so is immediate UB.
• Contain any loops that never terminate. These may be optimized away entirely or cause other unpredictable behaviour.

Trait Implementations

impl<R: Runtime> Default for KernelLauncher<R>

fn default() -> Self

Returns the “default value” for a type.