Benchmarks

The run-time inputs are set as follows

  1. cubeRootN is varied to record the assemble/initialization time, matrix-vector product time and the realtive forward error in the solution.

  2. nParticlesInLeafAlong1D is set to 6

  3. L is set to 1

  4. TOL_POW is set to 7

  5. Qchoice is set to 7

All the following benchmarks have been recorded by running the code on an Intel Xeon Gold 2.5GHz processor with 8 OpenMP threads. The compiler flags that were utilized are the same are those mentioned in the CMakeLists.txt file.

_images/assembleTime.png _images/matvecTime.png _images/error.png