BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
BEGIN:VEVENT
DTSTAMP:20260717T191431Z
UID:c30d6524-4cc8-4bd3-b41c-d0ce2ae00bbe
DTSTART:20190128T083000Z
DTEND:20190129T153000Z
DESCRIPTION:Overview\n\nMost HPC systems are clusters of shared memory node
 s. Such SMP nodes can be small multi-core CPUs up to large many-core CPUs.
  Parallel programming may combine the distributed memory parallelization o
 n the node interconnect (e.g.\, with MPI) with the shared memory paralleli
 zation inside of each node (e.g.\, with OpenMP or MPI-3.0 shared memory). 
 This course analyzes the strengths and weaknesses of several parallel prog
 ramming models on clusters of SMP nodes. Multi-socket-multi-core systems i
 n highly parallel environments are given special consideration. MPI-3.0 ha
 s introduced a new shared memory programming interface\, which can be comb
 ined with inter-node MPI communication. It can be used for direct neighbor
  accesses similar to OpenMP or for direct halo copies\, and enables new hy
 brid programming models. These models are compared with various hybrid MPI
 +OpenMP approaches and pure MPI. Numerous case studies and micro-benchmark
 s demonstrate the performance-related aspects of hybrid programming.\n\nHa
 nds-on sessions are included on both days. Tools for hybrid programming s
 uch as thread/process placement support and performance analysis are prese
 nted in a "how-to" section. This course provides scientific training in Co
 mputational Science\, and in addition\, the scientific exchange of the par
 ticipants among themselves. The course is a PRACE training event. It is or
 ganized by LRZ in cooperation with HLRS\, RRZE\, and VSC (Vienna Scientifi
 c Cluster).\n\nAgenda &amp\; Content (preliminary)\n\n1st day\n\n09:30 Reg
 istration\n10:00 Welcome\n10:05 Motivation\n10:15 Introduction\n10:45 Prog
 ramming Models\n           - Pure MPI\n11:05 Coffee Break\n11:25
   - Topology Optimization\n12:05    Practical (application aware Car
 tesian topology)\n12:45  - Topology Optimization (Wrap up)\n13:00 Lunch\n
 14:00  - MPI + MPI-3.0 Shared Memory\n14:30    Practical (replicated
  data)\n15:00 Coffee Break\n15:20  - MPI Memory Models and Synchronizatio
 n\n16:00    Practical (substituting pt-to-pt by shared memory)\n16:45 
 Coffee Break\n17:00    Practical (substituting barrier synchronization 
 by pt-to-pt)\n18:00 End\n\n19:00 Social Event at Gasthof Neuwirt (self pay
 ing)\n\n2nd day\n\n09:00 Programming Models (continued)\n        
    - MPI + OpenMP\n10:30 Coffee Break\n10:50     Practical (how to c
 ompile and start)\n11:30     Practical (hybrid through OpenMP paralleli
 zation)\n13:00 Lunch\n14:00  - Overlapping Communication and Computation\
 n14:20     Practical (taskloops) \n15:00 Coffee Break\n15:20  - MP
 I + OpenMP Conclusions\n15:30  - MPI + Accelerators\n15:45 Tools\n16:00 C
 onclusions\n16:15 Q&amp\;A\n16:30 End\n\n \n\n \nhttps://events.prace-ri
 .eu/event/807/
SUMMARY:Introduction to Hybrid Programming in HPC @ LRZ
URL;VALUE=URI:https://events.prace-ri.eu/event/807/
END:VEVENT
END:VCALENDAR