Module simd 

SIMD Optimizations Module

Vectorized operations using AVX2/SSE instructions for high-performance tensor operations.

Features

• Vectorized dot products
• SIMD matrix multiplication
• Element-wise operations (add, mul, relu)
• Reduction operations (sum, max, min)

Platform Support

• x86_64 with AVX2: Full SIMD acceleration
• Fallback: Pure Rust implementation for other platforms

Functions

add_elementwise_simd

Adição elemento-wise com SIMD

dot_product_simd

Dot product vetorizado usando AVX2

mul_elementwise_simd

Multiplicação elemento-wise com SIMD

mul_scalar_simd

Multiplicação escalar com SIMD

relu_simd

ReLU com SIMD (max(0, x))

sum_simd

Soma de todos os elementos com SIMD