Skip to content

Latest commit

 

History

History
32 lines (26 loc) · 2.37 KB

README.md

File metadata and controls

32 lines (26 loc) · 2.37 KB
Raw

Selalib Example

Build Selalib Dependency

  • clone and check out right version (see above)
  • depends on fftw and [email protected] or higher (in spack notation)
  • create build directory, load dependencies
  • in the build directory, run
cmake -DCMAKE_BUILD_TYPE=Release -DOPENMP_ENABLED=ON -DHDF5_PARALLEL_ENABLED=ON  -DUSE_FMEMPOOOL=OFF -DCMAKE_INSTALL_PREFIX=$(pwd)/install $SELALIB_DIR
make test_cpp_interface
make sll_m_sim_bsl_vp_3d3v_cart_dd_slim_interface
make sim_bsl_vp_3d3v_cart_dd_slim
make install

where test_cpp_interface can be used to test the gerenal C/Fortran interface, sim_bsl_vp_3d3v_cart_dd_slim is the mononlithic solver for our test case, and sll_m_sim_bsl_vp_3d3v_cart_dd_slim_interface builds the libraries needed for this selalib_distributed example. (This may take some time and usually fails if tried in parallel with make -j).

Build Selalib+DisCoTec example

  • run DisCoTec's cmake with -DDISCOTEC_WITH_SELALIB=1 -DSELALIB_DIR= flags (SELALIB_DIR should be set to selalib's CMAKE_INSTALL_PREFIX folder, appended with /cmake, or wherever you find SELALIBConfig.cmake accompanied by libselalib.a and libsll_m_sim_bsl_vp_3d3v_cart_dd_slim_interface.a)
  • then make selalib_distributed

Run Selalib+DisCoTec example

  • update combination technique parameters in ctparam, selalib parameters in template/param.nml (be careful with haveDiagnosticsTask -- it interpolates onto lmax to do the diagnostics, which can be extremely costly)
  • load dependencies and run with mpi, like shown in sbatch.sh -- the number of MPI tasks has to match the process group number and size in ctparam. Make sure that OpenMP runs with 1 process per MPI rank.

Postprocessing for Selalib+DisCoTec example

  • run ./combine_selalib_diagnostics ctparam or update paths in ./plot_landau.py and run (which also invokes./combine_selalib_diagnostics but then plots the quantities of interest)
  • analyze the sparse grid "surplusses" (maximum hierarchical coefficients per level) by running ../../tools/visualize_sg_minmax.py plot.dat_selalib_sg_${TIMESTEP}_0.txt. It shows a correlation matrix and plots projections of maxima on the different level combinations.