v8: shift heap space for aslr on 64bit

Previously we were only shifting the address space for ASLR on 32bit
processes, apply the same shift for 64bit so processes don't
get artificially limited native heap.