INFO4110 : Performance Evaluation
- Responsable(s) :
-
- Daniel Hirschkoff
- Enseignant(s) :
-
- Francescomaria Faticanti
Niveau
M1+M2
Discipline
Informatique
Public externe (ouverts aux auditeurs de cours)
Informations générales sur le cours : INFO4110
"Performance evaluation" is a label that covers many topics, from theory to practice, in computer science and computer engineering. It includes the design of mathematical models (very often with probabilistic assumptions), the analysis of such models (mathematically or using simulation), the design of experimental protocols (measuring tools and data collection), the analysis of data (statistics on real experiments or in silicon experiments). It is used to design, compare or tune computer systems and communication networks. Some tools also apply to related systems like transport networks, logistics or automatic control.
The goal of the course is to initiate students to the different possible techniques to evaluate the performance of computing systems. After taking the course, students will be able to define and perform a performance evaluation campaign, from choosing which techniques to employ (models, simulations, real-world experiments) to what components to evaluate, to which performance metrics to use.
Covered topics:
- Queuing theory
- Simulation
- Experimental performance analysis
- Basic knowledge in probability, statistics and in Python programming (for statistics / data analysis)
- Intermediate level in Linux Operating System, including command lines and C/C++ programming (for practical case studies)
- Knowledge on networks acquired through the course L3 Réseaux or similar (http://www.ens-lyon.fr/formation/catalogue-de-cours/info3107/2022)
- Knowledge on systems acquired through the course L3 ArchiSys or similar (http://www.ens-lyon.fr/formation/catalogue-de-cours/info3209/2023)
- Knowledge on probability acquired through the course L3 Probabilités or similar (http://www.ens-lyon.fr/formation/catalogue-de-cours/info3202/2022)
2 grades of equal weight:
- CC
- 35% mid-term
- 5% homeworks
- 60% project
- Final
- 100% final
Performance Evaluation (from theory to practice)
- Performance Modeling and Design of Computer Systems (Queueing Theory in Action), by Mor Harchol-Balter. Cambridge University Press, 2013.
Probability and queuing theory
- Introduction to probability models, by Sheldon K. Ross, Academic Press (11th edition available online http://mitran-lab.amath.unc.edu/courses/MATH768/biblio/introduction-to-prob-models-11th-edition.PDF)
Statistics
- Python for Data Analysis, by W. McKinney. O′Reilly, 2012.