Module simd_optimize 

CPU feature detection and SIMD optimization for NumRS

DEPRECATED: This module is deprecated in favor of scirs2_core::simd_ops::SimdUnifiedOps. All new code should use SimdUnifiedOps which provides:

• Platform-independent SIMD (AVX2, AVX-512, NEON auto-detection)
• 100+ SIMD operations with automatic fallback
• Better performance characteristics

The functions in this module remain for backwards compatibility but will be removed in a future major release.

Migration Guide

Instead of:

use numrs2::simd_optimize::avx2_enhanced::EnhancedSimdOps;
let result = EnhancedSimdOps::vectorized_exp_f32(&arr);

Use:

use scirs2_core::simd_ops::SimdUnifiedOps;
let nd_arr = scirs2_core::ndarray::Array1::from_vec(arr.to_vec());
let result = f32::simd_exp(&nd_arr.view());

This module provides functionality for detecting available CPU features and selecting the most efficient SIMD implementation for the current hardware.

Re-exports

pub use feature_detect::detect_cpu_features;

pub use feature_detect::CpuFeatures;

pub use simd_select::select_simd_implementation;

pub use simd_select::SimdImplementation;

pub use simd_traits::SimdPerformanceHints;

pub use unified_dispatcher::global_dispatcher;

pub use unified_dispatcher::optimized;

pub use unified_dispatcher::UnifiedSimdDispatcher;

Modules

avx2_enhanced

Enhanced AVX2 SIMD operations with advanced vectorization

avx2_ops

AVX2-optimized SIMD operations

feature_detect

CPU feature detection for SIMD optimization

neon_enhanced

Enhanced NEON SIMD operations for ARM processors

simd_select

SIMD implementation selection based on CPU features

simd_traits

SIMD-optimized trait implementations

unified_dispatcher

Unified SIMD dispatcher for cross-platform optimization

Functions

avx2_optimized_add_f32Deprecated

AVX2-optimized array addition for f32

avx2_optimized_add_f64Deprecated

AVX2-optimized array addition for f64

avx2_optimized_mul_f32Deprecated

AVX2-optimized array multiplication for f32

avx2_optimized_mul_f64Deprecated

AVX2-optimized array multiplication for f64

avx2_optimized_sqrt_f32Deprecated

AVX2-optimized square root for f32

avx2_optimized_sqrt_f64Deprecated

AVX2-optimized square root for f64

avx2_optimized_sum_f32Deprecated

AVX2-optimized sum for f32

avx2_optimized_sum_f64Deprecated

AVX2-optimized sum for f64

detect_and_selectDeprecated

CPU feature detection and SIMD implementation selection in one step