BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
BEGIN:VEVENT
DTSTAMP:20260709T121453Z
UID:d688488f-35ab-4be7-ba60-021772b8512c
DTSTART:20220405T064500Z
DTEND:20220407T140000Z
DESCRIPTION:Overview\n\nMost HPC systems are clusters of shared memory node
 s. To use such systems efficiently both memory consumption and communicati
 on time has to be optimized. Therefore\, hybrid programming may combine t
 he distributed memory parallelization on the node interconnect (e.g.\, wit
 h MPI) with the shared memory parallelization inside of each node (e.g.\, 
 with OpenMP or MPI-3.0 shared memory). This course analyzes the strengths 
 and weaknesses of several parallel programming models on clusters of SMP n
 odes. Multi-socket-multi-core systems in highly parallel environments are 
 given special consideration. MPI-3.0 has introduced a new shared memory pr
 ogramming interface\, which can be combined with inter-node MPI communicat
 ion. It can be used for direct neighbor accesses similar to OpenMP or for 
 direct halo copies\, and enables new hybrid programming models. These mode
 ls are compared with various hybrid MPI+OpenMP approaches and pure MPI. Nu
 merous case studies and micro-benchmarks demonstrate the performance-relat
 ed aspects of hybrid programming.\n\nHands-on sessions are included on all
  days. Tools for hybrid programming such as thread/process placement suppo
 rt and performance analysis are presented in a "how-to" section. This cour
 se provides scientific training in Computational Science and\, in additio
 n\, the scientific exchange of the participants among themselves.\n\nConte
 nt Levels: Beginners = 0:00h (0%) + Intermediate = 1:30h (10%) + Advanced 
 = 13:30h (90%)\n\nThis course is a PRACE training event. It is organized 
 by the VSC Research Center\, TU Wien\, in cooperation with HLRS and NHR@FA
 U.\n\nAgenda &amp\; Content\n\n1st day – 5 April 2022\n\n08:45   Join
  online\n09:00      Welcome\n09:05      Motivation\n09:15      In
 troduction\n09:30      Programming Models\n09:35         - MPI + O
 penMP\n10:00            Practical (how to compile and start)\n10:30
    Break\n10:45         - continue: MPI + OpenMP\n11:30   Break\
 n11:45         - continue: MPI + OpenMP\n12:30            Pract
 ical (how to do pinning)\n13:00   Lunch\n14:00            Practica
 l (hybrid through OpenMP parallelization)\n15:30            Q &amp\;
  A\, Discussion\n16:00   End of first day\n\n2nd day – 6 April 2022\n\
 n08:45   Join online\n09:00         - Overlapping Communication and
  Computation\n09:30            Practical (taskloops)\n10:30   Brea
 k\n10:45         - MPI + OpenMP Conclusions\n11:00         - MPI
  + Accelerators\n11:30      Tools\n11:45   Break\n12:00      Progr
 amming Models (continued)\n12:05         - MPI + MPI-3.0 Shared Memor
 y\n13:00   Lunch\n14:00            Practical (replicated data)\n15
 :30            Q &amp\; A\, Discussion\n16:00   End of second day\
 n\n3rd day – 7 April 2022\n\n08:45   Join online\n09:00         
 - MPI Memory Models and Synchronization\n09:40         - Pure MPI\n10
 :00   Break\n10:15         - Recap - MPI Virtual Topologies\n10:45
          - Topology Optimization\n11:15   Break\n11:30        
    Practical/Demo (application aware Cartesian topology)\n12:30      
    - Topology Optimization (Wrap up)\n12:45       Conclusions\n13:00
    Lunch\n14:00       Finish the hands-on labs\, Discussion\, Q &amp
 \; A\, Feedback\n16:00   End of third day (course)\nhttps://events.prace
 -ri.eu/event/1337/
SUMMARY:[ONLINE] Hybrid Programming in HPC – MPI+X @ VSC Vienna
URL;VALUE=URI:https://events.prace-ri.eu/event/1337/
END:VEVENT
END:VCALENDAR