optimize instruction sequence
I'd like to propose Ruby to have a infrastructure optimizing its execution.
My implementation and detailed explanation of how it works can be found at
I guess, I don't necessarily think that above patch is the best way to achieve speed.
You might not want to pull it to ruby-core as-is for reasons. Maybe it's a rough-cut.
That being said, I still think we need some sort of optimization like this.
Because when we want to make ruby 3x faster, there are only two ways to do so;
either make it execute 3 instructions at once, or to make it execute 1/3 less instructions.