Performance Evaluation of Core Numerical Algorithms: A Tool to Measure Instruction Level Parallelism

• Main Authors: Goossens, Bernard, Langlois, Philippe, Parello, David, Petit, Eric
• Other Authors: Electronique, Informatique, Automatique et Systèmes (ELIAUS), Université de Perpignan Via Domitia (UPVD), Digits, Architectures et Logiciels Informatiques (DALI), Laboratoire d'Informatique de Robotique et de Microélectronique de Montpellier (LIRMM), Université de Montpellier (UM)-Centre National de la Recherche Scientifique (CNRS)-Université de Montpellier (UM)-Centre National de la Recherche Scientifique (CNRS)-Université de Perpignan Via Domitia (UPVD), ANR-06-BLAN-0257,EVA-Flo,Evaluation et Validation Automatique pour le calcul Flottant(2006)
• Format: Conference Object
• Language: English
• Published: HAL CCSD 2010
• Subjects: [INFO.INFO-MS]Computer Science [cs]/Mathematical Software [cs.MS]; Reykjavík; Iceland
• Online Access: https://hal.science/hal-00477541; https://hal.science/hal-00477541v2/document; https://hal.science/hal-00477541v2/file/para10-final.pdf

International audience We measure and analyze the instruction level parallelism that condition the running-time performance of core numerical subroutines. We propose PerPI, a programmer oriented tool to fill the gap between high level algorithm analysis and machine dependent profiling tools and which provides reproducible results.