cuda_iszero_kernel_impl() [2/4]

void syten::CudaDenseTensorImpl::cuda_iszero_kernel_impl	(	std::size_t	sz,
		const float *	inp,
		int *	is_nonzero,
		void *	str
	)

Launcher for the CUDA isZero kernel, float version.

Parameters

sz	number of floats in array
inp	input array, device memory
is_nonzero	a boolean output parameter, true if the array is non-zero, device memory
str	CUDA stream in which the computation will take place