I have a UNIX-based server with dedicated processors. The dedicated processors are tied to compiled code. On code execution, they are restricted to used up to 100% of the processor's resources. I am looking for a way to isolate the processor load in respects to data throughput and threshold with out actual data driving the test.

Any ideas?