diff --git a/docs.it4i/software/karolina-compilation.md b/docs.it4i/software/karolina-compilation.md index 3045196c432d3c1b1988a8c9554cce684eaba2ee..ac6148c77bafbc366c1e9b699d97304375da4cc6 100644 --- a/docs.it4i/software/karolina-compilation.md +++ b/docs.it4i/software/karolina-compilation.md @@ -40,9 +40,12 @@ OMP_NUM_THREADS = 64 GOMP_CPU_AFFINITY=0:63:1 ``` -However, to get full performance, you have to execute two jobs on the two Karolina sockets at the time. +However, to get full performance, you have to execute two jobs on two Karolina sockets at the time. Other BLAS libraries may be used, however none performs as well as the MKL. +!!! note + Most MPI libraries do the binding automatically. The binding of MPI ranks can be inspected for any MPI by running `$ mpirun -n num_of_ranks numactl --show`. However, If the ranks spawn threads, binding of these threads should be done via the environment variables described above. + The choice of BLAS library and its performance may be verified with our benchmark, see [Lorenz BLAS performance benchmark][a].