Cube GUI User Guide  (CubeGUI 4.5, revision release-4.5)
Introduction in Cube GUI and its usage
Memory analysis for KNL

In order to analyse quality of the computational code Advisor requires that at least the following metrics are collected:

  1. LLC_MISSES
  2. knl_unc_imc0::UNC_M_CAS_COUNT:ALL:cpu=0,
  3. knl_unc_imc1::UNC_M_CAS_COUNT:ALL:cpu=0,
  4. knl_unc_imc2::UNC_M_CAS_COUNT:ALL:cpu=0,
  5. knl_unc_imc3::UNC_M_CAS_COUNT:ALL:cpu=0,
  6. knl_unc_imc4::UNC_M_CAS_COUNT:ALL:cpu=0 and
  7. knl_unc_imc5::UNC_M_CAS_COUNT:ALL:cpu=0

How to do it see Score-P manual


Cube Writer Library    Copyright © 1998–2020 Forschungszentrum Jülich GmbH, Jülich Supercomputing Centre
Copyright © 2009–2015 German Research School for Simulation Sciences GmbH, Laboratory for Parallel Programming