Simulation Kernels

A simulation kernel is an object code library containing the core simulation algorithms and services used for discrete event simulation. Various categories of simulation kernels are available, as follows:

  • Development vs. optimized (“kernel_type” preference)

    Development kernels allow ODB use, profiling, and detailed function call stack (FCS) information in errors and warnings.

    Optimized kernels run much faster than development, but do not allow ODB use. Also, no profiling data is collected for kernel procedures, nor will any Riverbed APIs be listed in the FCS information. Model code can still be profiled or can still include FCS information, if needed.

  • Sequential vs. parallel (“mt” preference)

    Parallel kernels allow multiple events to be executed in parallel.

  • 32-bit vs. 64-bit (“m32” and “m64” preferences)

    64-bit kernels use 64-bit addresses to give a larger space for executables.

Note: The preferences are used with op_runsim, op_mko, op_mkso, and similar programs to specify the intended simulation kernel.