Equivalent Red Hat Enterprise 6 systems produce vastly different runtime results during remote MySQL Database testing

Issue

• We have two very similar RHEL 6.3 systems. When we run a program that commits 3+ million queries to a remote MySQL database against each system, the results are wildly different in terms of %sys usage and time to complete.

• The following is seen on these systems:
  ** Exceptional network throughput. As tested by iperf, the results are what we expect: 945 Mb/s over gigabit infrastructure.
  ** The MySQL application itself is identical between the two systems as verified with md5sum verification across all libraries involved.
  ** The list of RPMs installed on the systems are the same.
  ** As far as I've checked briefly, sysctl.conf and kernel command line are identical.
  ** BIOS configuration and revisions on the servers.