Intel® Advisor Help
Collect data for applications with Data Parallel C++, OpenMP* target, and OpenCL™ code running on CPU.
--jit --no-jit |
Off (no-jit)
collect=projection
Use this option to model performance of a Data Parallel C++ (DPC++), C++/Fortran with OpenMP* target, or OpenCL™ application offloaded to a CPU.
Set up system environment to offload your GPU-enabled application to a CPU before using this option.
Model performance of your DPC++, C++/Fortran with OpenMP target, or OpenCL application on a target device.
$ advisor --collect=projection --project-dir=./advi –jit