BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
BEGIN:VEVENT
DTSTAMP:20260720T133952Z
UID:9d57dd93-cb70-4f40-b057-b331e247cff4
DTSTART:20200420T073000Z
DTEND:20200421T143000Z
DESCRIPTION:As stipulated by the Bavarian State Ministry of Science and the
  Arts on 10 March 2020\, we have to cancel or postpone all upcoming on-sit
 e courses and workshops at LRZ until  –  as of now  –  20 April 2020
 . \n\nWe have joined forces with VSC Vienna\, HLRS Stuttgart and RRZE Erla
 ngen and will offer this course online on 17-19 June 2020\, see https://ev
 ents.prace-ri.eu/event/1009/ for details and registration.\n\nOverview\n\n
 Most HPC systems are clusters of shared memory nodes. Such SMP nodes can b
 e small multi-core CPUs up to large many-core CPUs. Parallel programming m
 ay combine the distributed memory parallelization on the node interconnect
  (e.g.\, with MPI) with the shared memory parallelization inside of each n
 ode (e.g.\, with OpenMP or MPI-3.0 shared memory). This course analyzes th
 e strengths and weaknesses of several parallel programming models on clust
 ers of SMP nodes. Multi-socket-multi-core systems in highly parallel envir
 onments are given special consideration. MPI-3.0 has introduced a new shar
 ed memory programming interface\, which can be combined with inter-node MP
 I communication. It can be used for direct neighbor accesses similar to Op
 enMP or for direct halo copies\, and enables new hybrid programming models
 . These models are compared with various hybrid MPI+OpenMP approaches and 
 pure MPI. Numerous case studies and micro-benchmarks demonstrate the perfo
 rmance-related aspects of hybrid programming.\n\nHands-on sessions are inc
 luded on both days. Tools for hybrid programming such as thread/process p
 lacement support and performance analysis are presented in a "how-to" sect
 ion. This course provides scientific training in Computational Science\, a
 nd in addition\, the scientific exchange of the participants among themsel
 ves. The course is a PRACE training event. It is organized by LRZ in coope
 ration with HLRS\, RRZE\, and VSC (Vienna Scientific Cluster).\n\nAgenda &
 amp\; Content (preliminary)\n\n1st day\n\n09:30   Registration\n10:00  
   Welcome\n10:05    Motivation\n10:15    Introduction\n10:45    Pro
 gramming Models\n10:50       - MPI + OpenMP\n11:30   Coffee Break\n11:
 50       - continue: MPI + OpenMP\n12:40          Practical (how
  to compile and start)\n13:00   Lunch\n14:40          Practical (
 continued) \n15:00          Practical (hybrid through OpenMP paral
 lelization)\n16:00   Coffee Break\n16:20       - Overlapping Communica
 tion and Computation\n16:40          Practical (taskloops)\n17:20   
     - MPI + OpenMP Conclusions\n17:30       - MPI + Accelerators\n17:
 45    Tools\n18:00   End of first day\n\n19:00 Social Event at Gasthof 
 Neuwirt (self paying)\n\n2nd day\n\n09:00    Programming Models (continu
 ed)\n09:05       - MPI + MPI-3.0 Shared Memory\n09:45          Pra
 ctical (replicated data)\n10:30   Coffee break\n10:50          conti
 nue: Practical (replicated data)\n11:50       - MPI Memory Models and S
 ynchronization\n12:30   Lunch\n13:30       - Pure MPI\n13:50       
 - Topology Optimization\n14:30   Coffee Break\n14:50          Practi
 cal (application aware Cartesian topology)\n15:45       - Topology Opti
 mization (Wrap up)\n16:00     Conclusions\n16:15     Q &amp\; A\n16:30
    End of second day (course)\nhttps://events.prace-ri.eu/event/902/
SUMMARY:[POSTPONED] Introduction to Hybrid Programming in HPC @ LRZ
URL;VALUE=URI:https://events.prace-ri.eu/event/902/
END:VEVENT
END:VCALENDAR