Module algorithm_selector 

FFT Algorithm Auto-Selection Module

This module provides intelligent automatic selection of FFT algorithms based on:

• Input size characteristics (power-of-2, prime, smooth numbers)
• Hardware capabilities (cache size, SIMD support, core count)
• Memory constraints
• Historical performance data

Features

• Input Analysis: Detects optimal algorithm based on input size properties
• Cache-Aware Selection: Considers L1/L2/L3 cache sizes for optimal performance
• Memory Optimization: Selects memory-efficient algorithms for large inputs
• Hardware Detection: Adapts to available SIMD instructions and core count
• Performance Profiling: Tracks and learns from execution history

Example

use scirs2_fft::algorithm_selector::{AlgorithmSelector, SelectionConfig};

let selector = AlgorithmSelector::new();
let recommendation = selector.select_algorithm(1024, true).expect("Selection failed");
println!("Recommended: {:?}", recommendation.algorithm);

Structs

AlgorithmRecommendation

Algorithm recommendation with metadata

AlgorithmSelector

Main algorithm selector

CacheInfo

CPU cache information

HardwareInfo

Hardware information for algorithm selection

InputCharacteristics

Detected input characteristics for algorithm selection

PerformanceEntry

Performance history entry

PerformanceHistory

Performance history database

PerformanceStats

Performance statistics

SelectionConfig

Configuration for algorithm selection

SimdCapabilities

SIMD capability detection

Enums

FftAlgorithm

FFT Algorithm variants available for selection

SizeCharacteristic

Input size characteristics