Module sparse_fft_gpu_cuda 

GPU-accelerated sparse FFT implementation using scirs2-core abstractions

This module provides GPU implementations of sparse FFT algorithms through the scirs2-core::gpu module. All direct GPU API calls are forbidden.

Structs

BufferDescriptor

Placeholder for GPU buffer descriptor - to be implemented with core GPU abstractions

FftGpuContext

FFT-specific GPU context wrapping scirs2-core GPU context

GpuDeviceInfo

GPU device information using core abstractions

GpuMemoryManager

Placeholder memory manager - to be implemented with core GPU abstractions

GpuSparseFFT

CUDA-accelerated sparse FFT implementation

GpuStream

Placeholder for GPU stream - to be implemented with core GPU abstractions

Enums

BufferLocation

Placeholder for buffer location - to be implemented with core GPU abstractions

BufferType

Placeholder for buffer type - to be implemented with core GPU abstractions

Functions

cuda_batch_sparse_fft

Perform batch CUDA-accelerated sparse FFT

cuda_sparse_fft

Perform CUDA-accelerated sparse FFT

ensure_gpu_available

Check if GPU is available through core platform capabilities

get_cuda_devices

Initialize GPU subsystem and get available GPU devices

get_global_memory_manager

Placeholder for global memory manager - to be implemented with core GPU abstractions