Wissenschaftliche Berichte:

S. Hunold, A. Carpen-Amarie, F. Lübbe, J. Träff:
"PGMPI: Automatically Verifying Self-Consistent MPI Performance Guidelines";
Bericht für CoRR - Computing Research Repository; Berichts-Nr. arXiv:1606.00215, 2016; 13 S.

Kurzfassung englisch:
The Message Passing Interface (MPI) is the most commonly used application programming interface for process communication on current large-scale parallel systems. Due to the scale and complexity of modern parallel architectures, it is becoming increasingly difficult to optimize MPI libraries, as many factors can influence the communication performance. To assist MPI developers and users, we propose an automatic way to check whether MPI libraries respect self-consistent performance guidelines. We introduce the PGMPI framework that can benchmark and detect violations of performance guidelines. Our experimental results show that PGMPI can indeed help to pinpoint undesired and often unexpected performance degradations of collective MPI operations. We demonstrate how to overcome these performance issues of several libraries by adapting the algorithmic implementations of their respective collective MPI calls.

MPI, performance guidelines, benchmarking

Zugeordnete Projekte:
Projektleitung Jesper Larsson Träff:

Projektleitung Jesper Larsson Träff:

Erstellt aus der Publikationsdatenbank der Technischen Universität Wien.