BEGIN:VCALENDAR VERSION:2.0 PRODID:icalendar-ruby CALSCALE:GREGORIAN BEGIN:VEVENT DTSTAMP:20260703T195706Z UID:d1f9ab17-99d3-4abf-8c0a-427a832ca7c2 DTSTART:20211018T073000Z DTEND:20211022T153000Z DESCRIPTION:The registration to this course will open soon. \n\nPlease\, br ing your own laptop. All the PATC courses at BSC are free of charge.\n\nCo urse Convener: Xavier Martorell\n\nCourse Lecturers:\n\n\nJudit Giménez - Performance Tools - Group Manager\nGerman Llort - Performance Tools - Sen ior Researcher\nMarc Jordà - Accelerators and Communications for High Per formance Computing - Research Engineer\nAntonio Peña - Accelerators and C ommunications for High Performance Computing - Senior Researcher\nJavier T eruel -  Best Practices for Performance and Programmability  - Group Coo rdinator\nXavier Martorell - Programming Models - Parallel programming mod el - Group Manager\n\n\nLevel: Intermediate: For trainees with some theor etical and practical knowledge\, some programming experience.\n\nAdvanced: For trainees able to work independently and requiring guidance for solvin g complex problems.\n\nAttendants can bring their own applications and wor k with them during the course for parallelization and analysis.\n\n\nPrere quisites: Fortran\, C or C++ programming. All examples in the course will be done in C\n\nSoftware requirements: Zoom (recommended)\, SSH client (t o connect HPC systems)\, X Server (enabling remote visual tools).\n\nObjec tives: The objectives of this course are to understand the fundamental co ncepts supporting message-passing and shared memory programming models. Th e course covers the two widely used programming models: MPI for the distri buted-memory environments\, and OpenMP for the shared-memory architectures . The course also presents the main tools developed at BSC to get informat ion and analyze the execution of parallel applications\, Paraver and Extra e. It also presents the Parallware Assistant tool\, which is able to autom atically parallelize a large number of program structures\, and provide hi nts to the programmer with respect to how to change the code to improve pa rallelization. It deals with debugging alternatives\, including the use of GDB and Totalview.\n\nThe use of OpenMP in conjunction with MPI to better exploit the shared-memory capabilities of current compute nodes in cluste red architectures is also considered. Paraver will be used along the cours e as the tool to understand the behavior and performance of parallelized c odes. The course is taught using formal lectures and practical/programming sessions to reinforce the key concepts and set up the compilation/executi on environment.\n\nAttendants can bring their own applications and work wi th them during the course for parallelization and analysis.\n\nLearning Ou tcomes: The students who finish this course will be able to develop bench marks and applications with the MPI\, OpenMP and mixed MPI/OpenMP programm ing models\, as well as analyze their execution and tune their behaviour i n parallel architectures.\n\nAgenda: \n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nSess ions will be in October 18th-22nd 2021 from 9:30 – 13:00 and from 14:30 to 17:30 CET with 20’ break in between sessions and 1h30' lunch break - delivered online via Zoom\n\nDay 1 (Monday October 18th)\n\nSession 1 / 9: 30 – 13:00 (20`rest in between)\n1. Introduction to parallel architectur es\, algorithms design and performance parameters\n2. Introduction to the MPI programming model\n3. Practical: How to compile and run MPI applicatio ns\n\n13:00 - 14:30 Lunch Break\n\nSession 2 / 14:30 – 17:30 (20`rest in between)\n1. MPI: Point-to-point communication\, collective communication \n2. Practical: Simple matrix computations\n3. MPI: Blocking and non-block ing communications\n\nDay 2 (Tuesday October 19th)\n\nSession 1 / 9:30 – 13:00 (20`rest in between)\n1. MPI: Collectives\, Communicators\, Topolog ies\n2. Practical: Heat equation example\n\n13:00 - 14:30 Lunch Break\n\nS ession 2 / 14:30 – 17:30 (20`rest in between)\n1. Introduction to Parave r: tool to analyze and understand performance\n2. Practical: Trace generat ion and trace analysis\n\nDay 3 (Wednesday October 20th)\n\nSession 1 / 9: 30 – 13:00 (20`rest in between)\n1. Parallel debugging in MareNostrumIII \, options from print to Totalview\n2. Practical: GDB and IDB\n3. Practica l: Totalview\n4. Practical: Valgrind for memory leaks\n\n13:00 - 14:30 Lun ch Break\n\nSession 2 / 14:30 – 17:30 (20`rest in between)\n1. Shared-me mory programming models\, OpenMP fundamentals\n2. Parallel regions and wor k sharing constructs\n3. Synchronization mechanisms in OpenMP\n4. Practica l: heat diffusion in OpenMP\n\nDay 4 (Thursday October 21st)\n\nSession 1 / 9:30 – 13:00 (20`rest in between)\n1. Tasking in OpenMP 3.0/4.0/4.5\n2 . Programming using a hybrid MPI/OpenMP approach\n3. Practical: multisort in OpenMP and hybrid MPI/OpenMP\n\n13:00 - 14:30 Lunch Break\n\nSession 2 / 14:30 – 17:30 (20`rest in between)\n1. Parallware: guided parallelizat ion\n2. Practical session with Parallware examples\n\nDay 5 (Friday Octobe r 22nd)\n\nSession 1 / 9:30 – 13:00 (20`rest in between)\n1. Introductio n to the OmpSs programming model\n2. Practical: heat equation example and divide-and-conquer\n\n13:00 - 14:30 Lunch Break\n\nSession 2 / 14:30 – 1 7:30 (20`rest in between)\n1. Programming using a hybrid MPI/OmpSs approac h\n2. Practical: heat equation example and divide-and-conquer\n\nEND OF TR AINING COURSE\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nhttps://events.prace-ri.eu/even t/1230/ SUMMARY:[ONLINE] Parallel Programming Workshop @ BSC URL;VALUE=URI:https://events.prace-ri.eu/event/1230/ END:VEVENT END:VCALENDAR