> pitrou> I think your analysis is wrong. These mmap() calls are for
> pitrou> anonymous memory, most likely they are emitted by the libc's
> pitrou> malloc() to get some memory from the kernel. In other words
> pitrou> they will be blazingly fast.
>
> Are you sure? :-)

Well, it is fast. It's true that mmap is slower than brk, since the
kernel zero-fills the pages, but the overhead probably doesn't come
from this, but more likely from pymalloc or malloc, and also from the
call to _PyBytes_resize in posix_read when the number of bytes read is
smaller than what has been requested.