execute_sparse_fft_kernel

scirs2_fft::sparse_fft_gpu_kernels

Function execute_sparse_fft_kernel 

Source 
pub fn execute_sparse_fft_kernel<T>(
    signal: &[T],
    sparsity: usize,
    algorithm: SparseFFTAlgorithm,
    window_function: WindowFunction,
    gpu_arch: &str,
    compute_capability: (i32, i32),
    available_memory: usize,
) -> FFTResult<(Vec<Complex64>, Vec<usize>, KernelStats)>
where
    T: NumCast + Copy + Debug + 'static,
Expand description

Execute sparse FFT on GPU using optimized kernels

This function provides a high-level interface to the GPU kernel implementation, handling memory allocation, kernel execution, and result collection.

§Arguments

  • signal - Input signal
  • sparsity - Expected number of significant frequency components
  • algorithm - Sparse FFT algorithm to use
  • window_function - Window function to apply
  • gpu_arch - GPU architecture name
  • compute_capability - GPU compute capability
  • available_memory - Available GPU memory in bytes

§Returns

  • Result containing sparse frequency components and kernel statistics