In high performance computing, benchmarks evaluate architectures, compilers and optimizations. Standard benchmarks are mostly issued from the industrial world and may have a very long execution time. So, evaluating a new architecture or an optimization is costly. Most of the benchmarks are composed of independent kernels. Usually, users are only interested by a small subset of these kernels. To get faster and easier local optimizations, we should find ways to extract kernels as standalone executables. Also, benchmarks have redundant computational kernels. Some calculations do not bring new informations about the system that we want to study, despite that we measure them many times. By detecting similar operations and removing redundant kern...
International audienceEvaluating the strong scalability of OpenMP applications is a costly and time-...
Since performance is not portable between platforms, engineers must fine-tune heuristics for each pr...
Achieving high application performance depends on the combination of memory footprint, instruction m...
In high performance computing, benchmarks evaluate architectures, compilers and optimizations. Stand...
Dans le domaine du calcul haute performance, de nombreux programmes étalons ou benchmarks sont utili...
The complexity of CPUs has increased considerably since their beginnings, introducing mechanisms suc...
A crucial step in the process of application performance analysis is the accurate detection of progr...
Comprendre l'interaction existante entre applications, compilateurs, et architecture est fondamental...
International audienceSystem selection aims at finding the best architecture for a set of programs a...
A methodology is introduced to reduce the overall simulation time of large benchmarking suites. Prev...
Un choix efficace des optimisations de compilation améliore notablement la performances des applicat...
In this dissertation, we show that source-to-source optimization is an efficient method to generate ...
International audienceMany computationally-intensive algorithms benefit from the wide parallelism of...
Because of physical limits, hardware designers have switched to parallel systems to exploit ...
Conference of 17th Design, Automation and Test in Europe, DATE 2014 ; Conference Date: 24 March 2014...
International audienceEvaluating the strong scalability of OpenMP applications is a costly and time-...
Since performance is not portable between platforms, engineers must fine-tune heuristics for each pr...
Achieving high application performance depends on the combination of memory footprint, instruction m...
In high performance computing, benchmarks evaluate architectures, compilers and optimizations. Stand...
Dans le domaine du calcul haute performance, de nombreux programmes étalons ou benchmarks sont utili...
The complexity of CPUs has increased considerably since their beginnings, introducing mechanisms suc...
A crucial step in the process of application performance analysis is the accurate detection of progr...
Comprendre l'interaction existante entre applications, compilateurs, et architecture est fondamental...
International audienceSystem selection aims at finding the best architecture for a set of programs a...
A methodology is introduced to reduce the overall simulation time of large benchmarking suites. Prev...
Un choix efficace des optimisations de compilation améliore notablement la performances des applicat...
In this dissertation, we show that source-to-source optimization is an efficient method to generate ...
International audienceMany computationally-intensive algorithms benefit from the wide parallelism of...
Because of physical limits, hardware designers have switched to parallel systems to exploit ...
Conference of 17th Design, Automation and Test in Europe, DATE 2014 ; Conference Date: 24 March 2014...
International audienceEvaluating the strong scalability of OpenMP applications is a costly and time-...
Since performance is not portable between platforms, engineers must fine-tune heuristics for each pr...
Achieving high application performance depends on the combination of memory footprint, instruction m...