My pre-C solution was Go. It bootstrapped pretty fast, but I actually ran into a problem where my large embedded data structures (hash table, bloom filter, etc) took a very long time (hundreds of milliseconds?) to initialize. I guess it has a startup cost for embedded data. Go also has a slightly larger binary size, which may or may not have affected startup time on the CTF worker boxes.