nlcpy.sca.create_kernel

nlcpy.sca.create_kernel(description desc_i, description desc_o=None)

Creates a SCA kernel.

Creates a SCA kernel from a stencil description that denotes stencil shapes. If the keyword argument desc_o is omitted, the SCA kernel’s output array is automatically created when the SCA kernel is executed.

Parameters

desc_inlcpy.sca.description.description

The stenil description that is associated with the input ndarray.

desc_onlcpy.sca.description.description, optional

The stenil description that is associated with the output ndarray. If not given or None, the kernel creates a new output ndarray when the kernel execution is done.

Returns

kernelnlcpy.sca.kernel.kernel

The SCA kernel.

See also

nlcpy.sca.kernel.kernel.execute

Executes the created SCA kernel and returns the result of stencil computations.

nlcpy.sca.destroy_kernel

Destroy a SCA kernel. It is recommended that the kernel be properly destroyed by the function after you finish using the kernel.

Examples

>>> import nlcpy as vp
>>> xin = vp.arange(10, dtype='f4')
>>> xout = vp.zeros_like(xin)
>>> dxin, dxout = vp.sca.create_descriptor((xin, xout))
>>> desc_i = dxin[-1] + dxin[0] + dxin[1]
>>> desc_o = dxout[0]
>>> kern = vp.sca.create_kernel(desc_i, desc_o=desc_o)
>>> kern.execute()
array([ 0.,  3.,  6.,  9., 12., 15., 18., 21., 24.,  0.], dtype=float32)