Differences

This shows you the differences between two versions of the page.

--- doku:likwid [2015/09/21 14:48] – sh
+++ doku:likwid [2017/03/17 14:20] – removed ir
@@ Line 1: / Line 1: @@
-===== Likwid 4.0 =====
-==== Background: ====
-It is proving increasingly difficult to exert control over how different threads are assigned to the available CPU cores in multi-threaded OpenMP applications. Particularly troublesome are hybrid MPI/OpenMP codes where usually the developer has a clear idea of which regions to run in parallel, but relies on the OS for optimal assignment of different physical cores to the individual threads. A variety of methods do exist to explicitly state which CPU core should be linked to what particular thread, however, in practice many of these recommended ways of configuration turn out to be either non-functional, dependent on MPI versions, or frequently ineffective and overruled by the queuing system (e.g. [[doku:slurm|SLURM]]). In the following we describe the auxiliary tool ''likwid-pin'' that has shown promise in successfully managing arbitrary thread assignment to individual CPU cores in a more general way.
-==== Example: ====
-Assume we have the following little test program, test_mpit_var.c, and want to run it with 8 threads on a single compute node using the following thread assignment to physical cores: 3, 4, 2, 1, 6, 5, 7, 9. So after compiling it like, ''mpigcc -fopenmp ./test_mpit_var.c'' we could use the following submit script to [[doku:slurm|SLURM]]
-   #!/bin/bash
-   #
-   #SBATCH -J prflng
-   #SBATCH -N 2
-   #SBATCH --time=00:10:00
-   module purge
-   module load intel-mpi/5
-   module load allinea/5.10
-   perf-report mpirun -np 32 a.out
-This will result in the creation of two summary files in *.txt and *.html format providing an overview of the relative time spent in MPI, I/O, OpenMP etc.
-==== Further Reading: ====
-Explanations, examples, typical cases at [[http://content.allinea.com/downloads/getting-started.pdf]]