Identifying Data 2019/20
Subject (*) HPC Tools Code 614473105
Study programme
Mestrado Universitario en Computación de Altas Prestacións / High Performance Computing (Mod. Presencial)
Descriptors Cycle Period Year Type Credits
Official Master's Degree 1st four-month period
 First Optional 6
Language
English
Teaching method Face-to-face
Prerequisites
Department Enxeñaría de Computadores
Coordinador
Padron Gonzalez, Emilio Jose
 E-mail
emilio.padron@udc.es
Lecturers
Andrade Canosa, Diego
Padron Gonzalez, Emilio Jose
 E-mail
diego.andrade@udc.es
emilio.padron@udc.es
Web http://aula.cesga.es
General description O obxectivo desta materia é que o alumnado se familiarice cos tipos de aplicación máis comúns que son susceptibles de precisar a utilización de HPC, así como que coñezan as ferramentas e as implementacións existentes para cada unha delas, entendendo ademais os retos que hai que abordar para a súa paralelización e mellora do rendemento. Isto permitirá ao estudante adquirir un coñecemento xeral do mundo HPC e das súas diferentes aplicacións.

Ademais, o alumno aprenderá que ferramentas ten á súa disposición para a caracterización do rendemento en contornos HPC, e como se poden empregar estas para abordar o proceso de paralelización e mellora do rendemento dunha aplicación nunha determinada plataforma. Isto permitiralle ao estudante ser capaz de analizar o rendemento esperable desa aplicación nese sistema, identificando ademais os puntos quentes sobre os que centrar os seus esforzos de optimización.

Finalmente, o alumnado aprenderá que alternativas tecnolóxicas existen para despregar unha aplicación HPC de forma rápida e eficiente. Isto permitirá ao alumno ser capaz de distribuír aplicacións HPC de xeito sinxelo e eficaz en distintas contornos.

Study programme competencies
Code Study programme competences
A1 CE1 - Define, evaluate and select the most appropriate architecture and software to solve a problem
A2 CE2 - Analyze and improve the performance of a given architecture or software
A3 CE3 - Know the high performance computing basic concepts
A4 CE4 - Deepen in the knowledge of different programming tools and programming languages in the field of the high performance computing
A5 CE5 - Analyze, design and implement efficient parallel algorithms and applications
B1 CB6 - Possess and understand the knowledge that give a baseline or opportunity to be original in the development and/or application of ideas, often in a research environment
B3 CB8 - The students have to be able to integrate knowledge and face the complexity to make judgments from information, despite being partial and limited, includes reflexions about the social and ethical responsabilities linked to the application of their judgements and knowledge
B4 CB9 - The students have to be able to communicate their conclusions, their knowledge and the reasons that hold them to specialized and non specialized audience in a clear and unambiguous manner
B6 CG1 - Be able to search and select useful information to solve complex problems, using the bibliographic sources of the field
B8 CG3 - Be able to maintain and extend properly funded theoretical hypothesis to allow the introduction and exploitation of novel and advanced technologies in the field
B9 CG4 - Be able to plan and do research, development and innovation tasks in high performance computing related environments
C1 CT1 - Use the basic technologies of the information and computing technology field required for the professional development and the long-life learning
C4 CT4 - Value the importance of research, innovation and the technological development in the socioeconomical and cultural advance of the society

Learning aims
Learning outcomes Study programme competences
Students will know the most common types of applications in which HPC techniques are usually applied. AJ1
AJ2
 BJ1
BJ6
 CJ1
Students will learn to use tools to characterize and represent the performance of applications. AJ3
AJ4
 BJ3
BJ9
 CJ4
Students will learn to use tools to compile, generate and deploy software in HPC environments. AJ3
AJ5
 BJ1
BJ4
BJ8
 CJ1

Contents
Topic Sub-topic
A survey of main application types in HPC. For each type we’ll see: 1. Problem: formal description.
2. Parallelization and performance tuning challenges.
3. Existing approaches.
Tools to measure, characterize and represent the performance of HPC applications. 1. Usage of performance characterization and benchmarking tools, such as software monitoring and hardware counters.
2. Hot spot detection to drive the optimization process.
3. Application of performance models to this process.
4. Tools for application performance representation.
Tools for the compilation, generation and deployment of HPC software. 1. Code compilation, optimization and generation in a compiler.
2. Code optimization with a compiler.
3. Automatic parallelization and vectorization.
4. Software development tools.
5. Leveraging containers for the easy deployment of HPC applications.

Planning
Methodologies / tests Competencies Ordinary class hours Student’s personal work hours Total hours
Guest lecture / keynote speech A3 B1 C4 23 0 23
Laboratory practice A1 A2 A4 A5 C1 18 52 70
Supervised projects B3 B4 B6 B8 B9 0 54 54
Mixed objective/subjective test B4 B6 2 0 2
 
Personalized attention 1 0 1
 
(*)The information in the planning table is for guidance only and does not take into account the heterogeneity of the students.

Methodologies
Methodologies Description
Guest lecture / keynote speech Lectures, discussing the different lessons of the course. Students will have available all the necessary material in advance and the teacher will promote an active attitude in the classroom, asking questions that may clarify specific aspects and leaving open issues for student reflection.
Laboratory practice Lab sessions, allowing the students to become familiar from a practical standpoint with the issues discussed in the lectures.
Supervised projects Guided task fulfillment: students apply the acquired knowledge to solve different problems autonomously.
Mixed objective/subjective test Written test/exam to show that the students have acquired the Degree's competences trained in this course by answering theoretical questions and solving exercises.

Personalized attention
Methodologies
Laboratory practice
Supervised projects
Description
Personalized attention is guaranteed during the development of the laboratory practices and supervised projects, being essential to guide students in the fulfillment of their tasks. This personalized attention is also useful to validate and evaluate the work carried out throughout the different development stages, until finished.

Furthermore, it is recommended for students to leverage the teacher's office hours as a complementary assistance tool.

Assessment
Methodologies Competencies Description Qualification
Mixed objective/subjective test B4 B6 Written test/exam to show that the students have acquired the Degree's competences trained in this course by answering theoretical questions and solving exercises. 30
Supervised projects B3 B4 B6 B8 B9 Guided task fulfillment: students apply the acquired knowledge to solve different problems autonomously. 70
 
Assessment comments

Sources of information
Basic
[1] Computer Architecture: A Quantitative Approach (5th or 6th Ed.). John L. Hennessy, David A. Patterson. Morgan Kaufmann. ISBN 978-0123838728 (5th Ed. 2011) 978-0128119051 (6th Ed. 2017)
[2] Performance Tuning of Scientific Applications. David H. Bailey, Robert F. Lucas, Samuel Williams. CRC Press. ISBN 978-1439815694
Complementary
[3] Intel® C++ Compiler Developer Guide and Reference https://software.intel.com/cpp-compiler-developer-guide-and-reference

[4] A Guide to Vectorization with Intel® C++ Compilers https://software.intel.com/sites/default/files/m/4/8/8/2/a/31848-CompilerAutovectorizationGuide.pdf

[5] Intel® VTune™ Amplifier Help https://software.intel.com/en-us/vtune-amplifier-help

[6] Free Software Foundation, Inc.: Using the GNU Compiler Collection (GCC). https://gcc.gnu.org/onlinedocs

Recommendations
Subjects that it is recommended to have taken before
Parallel Programming/614473102

Subjects that are recommended to be taken simultaneously

Subjects that continue the syllabus

Other comments

Because of the strong interrelation between the lectures and the lab sessions, and the progressive presentation of concepts very related each other in the lectures, it is recommended to dedicate enough time to a daily study or review.

This course will leverage online communication tools in quite an intensive way: videoconference, e-mail, chat, etc.


(*)The teaching guide is the document in which the URV publishes the information about all its courses. It is a public document and cannot be modified. Only in exceptional cases can it be revised by the competent agent or duly revised so that it is in line with current legislation.