Guías Docentes Electrónicas
1. General information
Course:
HIGH PERFORMANCE COMPUTING
Code:
311049
Type:
CORE COURSE
ECTS credits:
6
Degree:
2359 - MASTERS DEGREE PROGRAMME IN COMPUTER ENGINEERING (CR-2019)
Academic year:
2021-22
Center:
108 - SCHOOL OF COMPUTER SCIENCE OF C. REAL
Group(s):
20 
Year:
1
Duration:
C2
Main language:
Spanish
Second language:
English
Use of additional languages:
Technical documentation in English
English Friendly:
Y
Web site:
https://campusvirtual.uclm.es
Bilingual:
N
Lecturer: JESUS BARBA ROMERO - Group(s): 20 
Building/Office
Department
Phone number
Email
Office hours
Fermín Caballero/3.09
TECNOLOGÍAS Y SISTEMAS DE INFORMACIÓN
3708
jesus.barba@uclm.es
Available at https://esi.uclm.es/categories/profesorado-y-tutorias

Lecturer: FRANCISCO PASCUAL ROMERO CHICHARRO - Group(s): 20 
Building/Office
Department
Phone number
Email
Office hours
Fermin Caballero / 3.17
TECNOLOGÍAS Y SISTEMAS DE INFORMACIÓN
5748
franciscop.romero@uclm.es
Available at https://esi.uclm.es/categories/profesorado-y-tutorias

2. Pre-Requisites
Not established
3. Justification in the curriculum, relation to other subjects and to the profession

The field of High Performance Computing (HPC) and its application has become one of the most dynamic in the ICT world. Therefore, it is mandatory to know them, their features and posibilities. Starting from a basic nowledge of the infrastructure (nodes + network) supporting this computing facilities, we will dig into the techiques and methods to benchmark supercomputers, as well as the design and development of parallel applications. HPC is preset in a miriad of engineering (i.e. complex simulations of physical and chemical processes) applications and business processes (i.e. Big Data processing). So, mastering HPC is key for the ICT professionals of the future.


4. Degree competences achieved in this course
Course competences
Code Description
CE09 Ability to design and assess operating systems and servers, plus applications and systems based on distributed computing.
CE10 Ability to understand a apply advanced knowledge on high performance computing and numerical or computational methods to engineering problems.
INS01 Analysis, synthesis and assessment skills.
INS04 Problem solving skills by the application of engineering techniques.
INS05 Argumentative skills to logically justify and explain decisions and opinions.
PER01 Team work abilities.
SIS03 Autonomous learning.
5. Objectives or Learning Outcomes
Course learning outcomes
Description
Manage tasks of all elements involved in the running of a high-performance distributed data processing system
Design and engineer high-performance and high-availability data processing equipment, including hardware, software and human resources
Evaluate and exploit the system, including socio-economic aspects
Additional outcomes
Description
Provide the student with the ability to make professional and business decisions that will improve the performance and competitiveness of his organization's ICT infrastructure.
Teach the student in the diverse paradigms of parallel computer programming, influence software techniques for the design and implementation of efficient parallel algorithms and applications, and apply these techniques in a practical way for the programming of parallel computers with different architectures, using supercomputing resources
6. Units / Contents
  • Unit 1: Introduction to High Performance Computing
  • Unit 2: Performance Analysis and Benchmarking
  • Unit 3: Paralell Programming Models for HPC
  • Unit 4: Platforms and Model
  • Unit 5: Application Deployment
ADDITIONAL COMMENTS, REMARKS

The practical sessions will consist of adjusting a theoretical model of system runtimes, determining the performance of our systems and developing distributed applications using the MPI and OpenMP libraries


7. Activities, Units/Modules and Methodology
Training Activity Methodology Related Competences ECTS Hours As Com Description
Class Attendance (theory) [ON-SITE] Combination of methods CE09 CE10 0.75 18.75 N N Theory Master Classes
Laboratory practice or sessions [ON-SITE] Practical or hands-on activities CE09 CE10 INS04 0.57 14.25 Y Y Practices with HPC Systems
Individual tutoring sessions [ON-SITE] Self-study INS05 0.16 4 N N
Study and Exam Preparation [OFF-SITE] Self-study SIS03 2.4 60 N N
Problem solving and/or case studies [ON-SITE] Case Studies CE09 CE10 INS04 0.6 15 Y N
Practicum and practical activities report writing or preparation [OFF-SITE] Self-study INS01 PER01 1.2 30 Y N Report Writing
Final test [ON-SITE] CE09 CE10 INS01 0.32 8 Y Y
Total: 6 150
Total credits of in-class work: 2.4 Total class time hours: 60
Total credits of out of class work: 3.6 Total hours of out of class work: 90

As: Assessable training activity
Com: Training activity of compulsory overcoming (It will be essential to overcome both continuous and non-continuous assessment).

8. Evaluation criteria and Grading System
Evaluation System Continuous assessment Non-continuous evaluation * Description
Final test 40.00% 40.00% Test to be carried out within the planned exam dates of the final exam call (convocatoria ordinaria).
Assessment of problem solving and/or case studies 20.00% 20.00% Resolution of different practical cases proposed in class (INF)
Laboratory sessions 30.00% 30.00% Carrying out practices and preparing a report on the laboratory worf. (LAB)
Oral presentations assessment 10.00% 10.00% Presentation of solutions to problems and cases raised in class (PRES)
Total: 100.00% 100.00%  
According to art. 6 of the UCLM Student Evaluation Regulations, it must be provided to students who cannot regularly attend face-to-face training activities the passing of the subject, having the right (art. 13.2) to be globally graded, in 2 annual calls per subject , an ordinary and an extraordinary one (evaluating 100% of the competences).

Evaluation criteria for the final exam:
  • Continuous assessment:
    In compulsory activities, a minimum mark of 40% is required in order to pass that activity and have the possibility to therefore pass the entire subject. The evaluation of the activities will be global and therefore must be quantified by means of a single mark. In the case of the activities that may be retaken (i.e., rescheduling), an alternative activity or test will be offered in the resit/retake exam call (convocatoria extraordinaria).

    A student is considered to pass the subject if she/he obtains a minimum of 50 points out of 100, taking into account the points obtained in all the evaluable activities, and also has passed all the compulsory activities.

    For students who do not pass the subject in the final exam call (convocatoria ordinaria), the marks of activities already passed will be conserved for the resit/retake examcall (convocatoria extraordinaria). In the case of the passed recoverable activities, the student will have the opportunity to receive an alternative evaluation of those activities in the resit/retake exam call and, in that case, the final grade of the activity will correspond to the latter grade obtained.

    The mark of the passed activities in any call, except for the final exam, will be conserved for the subsequent academic year at the request of the student, provided that mark is equal or greater than 50% and that the activities and evaluation criteria of the subject remain unchanged prior to the beginning of that academic year.

    The failure of a student to attend the final exam will automatically result in her/him receiving a "Failure to attend" (no presentado). If the student has not passed any compulsory evaluation activity, the maximum final grade will be 40%.
  • Non-continuous evaluation:
    Students may apply at the beginning of the semester for the non-continuous assessment mode. In the same way, the student may change to the non-continuous evaluation mode as long as she/he has not participated during the teaching period in evaluable activities that together account for at least 50% of the total mark of the subject. If a student has reached this 50% of the total obtainable mark or the teaching period is over, she/he will be considered in continuous assessment without the possibility of changing to non-continuous evaluation mode.

    Students who take the non-continuous evaluation mode will be globally graded, in 2 annual calls per subject, an ordinary and an extraordinary one (evaluating 100% of the competences), through the assessment systems indicated in the column "Non-continuous evaluation".

    In the "non-continuous evaluation" mode, it is not compulsory to keep the mark obtained by the student in the activities or tests (progress test or partial test) taken in the continuous assessment mode.

Specifications for the resit/retake exam:
Evaluation tests will be conducted for all recoverable activities
Specifications for the second resit / retake exam:
Same characteristics as the resit/retake exam call.
9. Assignments, course calendar and important dates
Not related to the syllabus/contents
Hours hours

General comments about the planning: Four hours of class per week.
10. Bibliography and Sources
Author(s) Title Book/Journal Citv Publishing house ISBN Year Description Link Catálogo biblioteca
Ananth Grama, George Karypis, Vipin Kumar y Anshul Gupta Introduction to Parallel Computing Addison Wesley 978-0201648652 2003 Acceso a la versión digital a través de la web de la biblioteca de la UCLM  
Michael J. Quinn Parallel Programming in C with MPI and OpenMP McGraw Hill Higher Education 978-0072822564 2003  
Peter Pacheco An Introduction to Parallel Programming Morgan Kaufmann 978-0-12-374260-5 2011 http://proquest.safaribooksonline.com/book/programming/9780123742605 Ficha de la biblioteca
Rohit Chandra Leonardo Dagum Dave Kohr Dror Maydan Jeff McDonald Ramesh Menon Parallel Programming in OpenMP San Francisco Morgan Kaufmann Publishers 1-55860-671-8 2001  
Thomas Sterling High Performance Computing: Modern Systems and Practices Morgan Kauffman 2017  



Web mantenido y actualizado por el Servicio de informática