Or "How I sped up Neat by a factor of 2x". If you check the related_post_gen benchmark, you will find that Neat, my language, has moved up a few spots. How did I achieve this? Did I implement new high-level optimizer passes that use language details to expose hidden optimization potential? I changed arrays to be passed as three pointer parameters instead of one parameter consisting of a three-pointer struct. That's it. This problem has been vexing me for a long time. Neat seemed weirdly slower t...