For native benchmark items such as libmicro, we execute the programs and
collect the results through UNIX pipe. However, we should always perform
sanity checks in advance.
At least, we need to avoid executing native benchmarks when programs are not
present.
Original issue reported on code.google.com by jserv.tw@gmail.com on 15 Jul 2011 at 5:11
Original issue reported on code.google.com by
jserv.tw@gmail.com
on 15 Jul 2011 at 5:11