Class ConvolveImageUnrolled_SB_S32_S32

java.lang.Object
boofcv.alg.filter.convolve.noborder.ConvolveImageUnrolled_SB_S32_S32

@Generated("boofcv.alg.filter.convolve.noborder.GenerateConvolvedUnrolled_SB")
public class ConvolveImageUnrolled_SB_S32_S32
extends Object

Unrolls the convolution kernel to reduce array accessing and save often used variables to the stack.

Unrolling the image being convolved resulting in an additional 10% performance boost on a Core i7 processor, see commented out code below. Due to the added complexity it was decided that this performance boost was not worth it. By comparison, unrolling the kernel causes a performance boost between 2 and 3 times.

DO NOT MODIFY. Automatically generated code created by GenerateConvolvedUnrolled_SB