[Gems-users] Different CPI's for Ruby and Opal

["Jonathan Brooks" <jonathan@xxxxxxxxxxxxx>]

I am running a CMP system (2 procs on 1 chip) with Serengeti/Solaris target (mostly default configuration).  I use both Opal and Ruby for a simulation of running a few binaries for 1 million steps.  When I look at the statistics in Ruby, I see Instructions per cycle per processor, which I assume I can take the inverse to get Cycles Per Instruction).  However these numbers do not match the CPI’s I get in Opal’s report.  Why are they different?