Mandelbrot optimizaton with SIMD instructions

Usage

1. Processor

   AMD Ryzen 5 4500U

2. Instructions

   AVX, AVX2

3. Language

   C, C++

4. Graphic library

   SFML

5. Compiler

   GCC

Controls

1. Move up

   UP

2. Move down

   DOWN

3. Move left

   LEFT

4. Move right

   RIGHT

5. Zoom in

   PGUP

6. Zoom out

   PGDN

Compile

•   g++ main.cpp -O2 -mavx -mavx2 -I"C:\Program Files\MinGW64\SFML-2.5.1\include" \
                                  -L"C:\Program Files\MinGW64\SFML-2.5.1\lib" \
                                  -lsfml-graphics -lsfml-window -lsfml-system
  

Results

Without SIMD instructions

Using AVX & AVX2

CONCLUSION

As seen, an increase in

fps is about 6 times.

The result has been gotten using

1.  __mm256 (8 32 single-precision float variables)
   

2.  AVX & AVX2 instructions