BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
BEGIN:VEVENT
DTSTAMP:20260625T072632Z
UID:47beafe7-a8fd-41e4-a0c7-7b1774fa7243
DTSTART:20140317T093000Z
DTEND:20140318T150000Z
DESCRIPTION:In some applications\, the overheads associated with the fundam
 entally two-sided (send and receive) nature of MPI message-passing can adv
 ersely affect performance. This is of particular concern for scaling up to
  extremely large systems. There is a renewed interest in simpler single-si
 ded communications models where data can be written/read directly to/from 
 remote processes.\nThis two-day course covers two single-sided PGAS librar
 ies: the OpenSHMEM standard http://www.openshmem.org/ on day 1\, and the n
 ew open-source GASPI library http://www.gaspi.de/ on day 2. Hands-on pract
 ical sessions will play a central part in the course\, illustrating key is
 sues such as the need for appropriate synchronisation to ensure program co
 rrectness. All the exercises can be undertaken using C or Fortran. The Ope
 nSHMEM material will be delivered by the ARCHER CSE team\; the GASPI mater
 ial will be delivered by Christian Simmendinger (T-Systems Solutions for R
 esearch) and Mirko Rahn (Fraunhofer ITWM). Further details of the GASPI tr
 aining are available below.Efficient Parallel Programming with GASPI\nThe 
 HPC programmers of tomorrow will have to write codes\, which are able to d
 eal with systems hundreds of times larger than the top supercomputers of t
 oday. In this Tutorial we present an asynchronous dataflow programming mod
 el for Partitioned Global Address Spaces (PGAS) as an alternative to the p
 rogramming model of MPI.\nGASPI\, which stands for Global Address Space Pr
 ogramming Interface\, is a partitioned global address space (PGAS) API. Th
 e GASPI API is designed as a C/C++/Fortran library and focused on three ke
 y objectives: scalability\, flexibility and fault tolerance. In order to a
 chieve its significantly improved scaling behaviour GASPI aims at asynchro
 nous dataflow with remote completion\, rather than bulk-synchronous messag
 e exchanges. GASPI follows a single/multiple program multiple data (SPMD/M
 PMD) approach and offers a small\, yet powerful API (also see http://www.g
 aspi.de). GASPI today is used in academic and industrial simulation applic
 ations.\nThe Tutorial gives an overview over the key concepts of elements 
 of GASPI\, such as synchronization primitives\, synchronous and asynchrono
 us collectives\, fine grained control over one-sided read and write commun
 ication primitives\, global atomics\, passive receives\, communication gro
 ups and communication queues. GASPI aims at multi-threaded execution\, off
 ers a thread-safe API and can be used in combination with all current thre
 ading models (OpenMP\, Pthreads\, MCTP\, and others). GASPI provides its p
 artitioned global address space in the form of configurable memory segment
 s and features support for heterogeneous memory architectures. All GASPI s
 egments can directly read and write from/to each other. By spawning a GASP
 I segment across e.g. the main memory of a distributed Xeon Phi system and
  a segment across the memory of the corresponding x86 host\, the GASPI API
  hence can provide a consistent and cohesive view of this hybrid distribut
 ed memory architecture. The flexibility of the configurable GASPI segments
  also allows developers to both leverage multiple memory models within a s
 ingle application and/or to globally tightly couple different applications
  (e.g. multi-physics solvers). GASPI is failure tolerant and allows for a 
 dynamic (shrinking or growing) node set. All non-local procedures feature 
 timeout parameters and provide a well defined exit status.\nThe Tutorial w
 ill provide a hands-on introduction (in C and Fortran) which features exam
 ples and use-cases for all its key concepts (segment creation\, synchroniz
 ation primitives\, read and write communication primitives\, global atomic
 s\, passive receives\, collectives\, communication groups and communicatio
 n queues). Case studies which demonstrate the various aspects of the GASPI
  API and application categories that can take advantage of these aspects a
 re identified. This Tutorial also includes a discussion of the current GPI
 -2 release\, the first release which supports the GASPI standard. Benchmar
 k results on different platforms are discussed. Tools for programming with
  GASPI/GPI such as profiling tools are presented in a "howto" section.\n\n
 https://events.prace-ri.eu/event/279/
SUMMARY:Single-sided PGAS Communications Libraries @ EPCC
URL;VALUE=URI:https://events.prace-ri.eu/event/279/
END:VEVENT
END:VCALENDAR