Optimize for the current CPU arch and for speed

Use -Ofast -march=native -mtune=native -funroll-loops to improve the
speed without effort :)

Before: around 8.5 seconds per iteration (11200 songs/second)
After: around 2.9~3 seconds per iteration (30000 songs/second)

Author: antlarr

setup.py

@@ -9,7 +9,9 @@
                     libraries=['boost_python-py3'],
                     library_dirs=['/usr/lib'],
                     sources=['bard/bard_ext.cpp'],
-                    extra_compile_args=['-std=gnu++17', '-fopenmp'])
+                    extra_compile_args=['-std=gnu++17', '-fopenmp', '-Ofast',
+                                        '-march=native', '-mtune=native',
+                                        '-funroll-loops'])
 
 setup(
     # Application name: