BEGIN:VCALENDAR VERSION:2.0 PRODID:icalendar-ruby CALSCALE:GREGORIAN BEGIN:VEVENT DTSTAMP:20260705T105117Z UID:a78e75e4-6ee0-44f2-87be-17daff269edc DTSTART:20171023T080000Z DTEND:20171027T150000Z DESCRIPTION:The registration to this course is now open. Please\, bring yo ur own laptop. All the PATC courses at BSC are free of charge.\n\nCourse C onvener: Xavier Martorell\n\nLOCATION: UPC Campus Nord premises.Building C 6\, room 106 and 101\n\nLevel: \n\nIntermediate: For trainees with some t heoretical and practical knowledge\, some programming experience.\n\nAdvan ced: For trainees able to work independently and requiring guidance for so lving complex problems.\n\nAttendants can bring their own applications and work with them during the course for parallelization and analysis.\n\nPre requisites: Fortran\, C or C++ programming. All examples in the course wi ll be done in C\n\nObjectives:  The objectives of this course are to unde rstand the fundamental concepts supporting message-passing and shared memo ry programming models. The course covers the two widely used programming m odels: MPI for the distributed-memory environments\, and OpenMP for the sh ared-memory architectures. The course also presents the main tools develop ed at BSC to get information and analyze the execution of parallel applica tions\, Paraver and Extrae. It also presents the Parallware Assistant tool \, which is able to automatically parallelize a large number of program st ructures\, and provide hints to the programmer with respect to how to chan ge the code to improve parallelization. It deals with debugging alternativ es\, including the use of GDB and Totalview. The use of OpenMP in conjunct ion with MPI to better exploit the shared-memory capabilities of current c ompute nodes in clustered architectures is also considered. Paraver will b e used along the course as the tool to understand the behavior and perform ance of parallelized codes. The course is taught using formal lectures and practical/programming sessions to reinforce the key concepts and set up t he compilation/execution environment..\n\nAttendants can bring their own a pplications and work with them during the course for parallelization and a nalysis.\n\nLearning Outcomes:\n\nThe students who finish this course will be able to develop benchmarks and applications with the MPI\, OpenMP and mixed MPI/OpenMP programming models\, as well as analyze their execution a nd tune their behaviour in parallel architectures.\n\nAgenda:\n\n \n\nDay 1 (Monday)\n\nSession 1 / 10:00 am – 1:00 pm (2 h lectures\, 1 h practi cal)\n\n1. Introduction to parallel architectures\, algorithms design and performance parameters\n\n2. Introduction to the MPI programming model\n\n 3. Practical: How to compile and run MPI applications\n\nSession 2 / 2:00p m – 5:00 pm (2h lectures\, 1h practical)\n\n1. MPI: Point-to-point commu nication\, collective communication\n\n2. Practical: Simple matrix computa tions\n\n3. MPI: Blocking and non-blocking communications\n\n \n\nDay 2 ( Tuesday)\n\nSession 1 / 10:00 am - 1:00 pm (1.5 h lectures\, 1.5 h practic al)\n\n1. MPI: Collectives\, Communicators\, Topologies\n\n2. Practical: H eat equation example\n\nSession 2 / 2:00 pm - 5:00 pm (1.5 h lectures\, 1. 5 h practical)\n\n1. Introduction to Paraver: tool to analyze and understa nd performance\n\n2. Practical: Trace generation and trace analysis\n\n \ n\nDay 3 (Wednesday)\n\nSession 1 / 10:00 am - 1:00 pm (1 h lecture\, 2h p ractical)\n\n1. Parallel debugging in MareNostrumIII\, options from print to Totalview\n\n2. Practical: GDB and IDB\n\n3. Practical: Totalview\n\n4. Practical: Valgrind for memory leaks\n\nSession 2 / 2:00 pm - 5:00 pm (2 h lectures\, 1 h practical)\n\n1. Shared-memory programming models\, OpenM P fundamentals\n\n2. Parallel regions and work sharing constructs\n\n3. Sy nchronization mechanisms in OpenMP\n\n4. Practical: heat diffusion in Open MP\n\n \n\nDay 4 (Thursday)\n\nSession 1 / 10:00am – 1:00 pm (2 h pract ical\, 1 h lectures)\n\n1. Tasking in OpenMP 3.0/4.0/4.5\n\n2. Programming using a hybrid MPI/OpenMP approach\n\n3. Practical: multisort in OpenMP a nd hybrid MPI/OpenMP\n\nSession 2 / 2:00pm – 5:00 pm (1.5 h lectures\, 1 .5 h practical)\n\n1. Parallware: guided parallelization\n\n2. Practical s ession with Parallware examples\n\n \n\nDay 5 (Friday)\n\nSession 1 / 10: 00 am – 1:00 pm (2 hour lectures\, 1 h practical)\n\n1. Introduction to the OmpSs programming model\n\n2. Practical: heat equation example and div ide-and-conquer\n\nSession 2 / 2:00pm – 5:00 pm (1 h lectures\, 2 h prac tical)\n\n1. Programming using a hybrid MPI/OmpSs approach\n\n2. Practical : heat equation example and divide-and-conquer\n\n \n\nEND of COURSE\n\n  \n\nhttps://events.prace-ri.eu/event/640/ SUMMARY:Parallel Programming Workshop @ BSC URL;VALUE=URI:https://events.prace-ri.eu/event/640/ END:VEVENT END:VCALENDAR