It is a bump allocator. I don't know why it's so much faster than mine, but my hypothesis was that CF Bolz was testing on a faster machine. The speedup over ma

It is a bump allocator. I don't know why it's so much faster than mine, but my hypothesis was that CF Bolz was testing on a faster machine. The speedup over malloc/free is because bumping a pointer is much faster than calling a subroutine.