The architecture of supercomputers is evolving to expose massive parallelism. MPI and OpenMP are widely used in application codes on the largest supercomputers in the world. The community primarily focused on composing MPI with OpenMP before its version 3.0 introduced task-based programming. Recent advances in OpenMP task model and its interoperability with MPI enabled fine model composition and seamless support for asynchrony. Yet, OpenMP tasking overheads limit the gain of task-based applications over their historical loop parallelization (parallel for construct). This paper identifies the OpenMP task dependency graph discovery speed as a limiting factor in the performance of task-based applications. We study its impact on intra and inter-node performances over two benchmarks (Cholesky, HPCG) and a proxy-application (LULESH). We evaluate the performance impacts of several discovery optimizations, and introduce a persistent task dependency graph reducing overheads by a factor up to 15 at run-time. We measure 2x speedup over parallel for versions weak scaled to 16K cores, due to improved cache memory use and communication overlap, enabled by task refinement and depth-first scheduling.
The adoption of ARM processor architectures is on the rise in the HPC ecosystem. Fugaku supercomputer is a homogeneous ARM-based machine, and is one among the most powerful machine in the world. In the programming world, dependent task-based programming models are gaining tractions due to their many advantages: dynamic load balancing, implicit expression of communication/computation overlap, early-bird communication posting,...MPI and OpenMP are two widespreads programming standards that make possible task-based programming at a distributed memory level. Despite its many advantages, mixed-use of the standard programming models using dependent tasks is still under-evaluated on large-scale machines. In this paper, we provide an overview on mixing OpenMP dependent tasking model with MPI with the state-of-the-art software stack (GCC-13, Clang17, MPC-OMP). We provide the level of performances to expect by porting applications to such mixed-use of the standard on the Fugaku supercomputers, using two benchmarks (Cholesky, HPCCG) and a proxy-application (LULESH). We show that software stack, resource binding and communication progression mechanisms are factors that have a significant impact on performance. On distributed applications, performances reaches up to 80% of effiency for task-based applications like HPCCG. We also point-out a few areas of improvements in OpenMP runtimes.
Parallélisation sur un moteur exécutif à base de tâches des méthodes itératives pour la résolution de systèmes linéaires creux sur architecture multi et many coeurs : application aux méthodes de types décomposition de domaines multi-niveaux: Les méthodes en simulation numérique dans le domaine de l’ingénierie pétrolière nécessitent la résolution de systèmes linéaires creux de grande taille et non structurés. La performance des méthodes itératives utilisées pour résoudre ces systèmes représente un enjeu majeur afin de permettre de tester de nombreux scénario.Dans ces travaux, nous présentons une manière d'implémenter des méthodes itératives parallèles au dessus d’un support exécutif à base de tâches. Afin de simplifier le développement des méthodes tout en gardant un contrôle fin sur la gestion du parallélisme, nous avons proposé une API permettant d’exprimer implicitement les dépendances entre tâches : la sémantique de l'API reste séquentielle et le parallélisme est implicite.Nous avons étendu le support exécutif HARTS pour enregistrer une trace d'exécution afin de mieux exploiter les architectures NUMA, tout comme de prendre en compte un placement des tâches et des données calculé au niveau de l’API. Nous avons porté et évalué l'API sur les processeurs many-coeurs KNL en considérant les différents types de mémoires de l’architecture. Cela nous a amené à optimiser le calcul du SpMV qui limite la performance de nos applications.L'ensemble de ce travail a été évalué sur des méthodes itératives et en particulier l’une de type décomposition de domaine. Nous montrons alors la pertinence de notre API, qui nous permet d’atteindre de très bon niveaux de performances aussi bien sur architecture multi-coeurs que many-coeurs.
High-Performance Computing (HPC) is currently facing significant challenges. The hardware pressure has become increasingly difficult to manage due to the lack of parallel abstractions in applications. As a result, parallel programs must undergo drastic evolution to effectively exploit underlying hardware parallelism. Failure to do so results in inefficient code. In this constrained environment, parallel runtimes play a critical role, and their testing becomes crucial. This paper focuses on the MPI interface and leverages the MPI binding tools to develop a multi-language test suite for MPI. By doing so and building on previous work from the Forum document editors, we implement a systematic testing of MPI symbols in the context of the Parallel Computing Validation System (PCVS), which is an HPC validation platform dedicated to running and managing test suites at scale. We first describe PCVS, then outline the process of generating the MPI API test suite, and finally, run these tests at scale. All data sets, code generators, and implementations are made available in open-source to the community. We also set up a dedicated website showcasing the results, which self-updates thanks to the Spack package manager.