BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
BEGIN:VEVENT
DTSTAMP:20260622T082604Z
UID:1e9372a0-1cff-4430-a6e3-06798e7c5dec
DTSTART:20130326T080000Z
DTEND:20130328T160000Z
DESCRIPTION:This course will present the state of the art in the developmen
 t of parallel direct methods for sparse linear systems. Some of these meth
 ods are being implemented as black box solvers\; the corresponding toolkit
 s will be experimented in dedicated hand-on sessions.\nThe course will als
 o present recent work on hybrid (direct / iterative) methods that better e
 xploit the two level structure of current parallel architectures. \nDirec
 t solvers:\n		Dense matrices (serial and parallel\, 1D and 2D distribution
 s)\n	\n		Sparse matrices (graph model\, renumbering\, elimination graph\, 
 symbolic and numeric factorizations)\n	\n		Parallel methods: data distribu
 tion\, scheduling of computations and communications\n	\n		Hands-on with t
 he MUMPS and Pastix solvers\n\nHybrid methods\n		Basics on Krylov subspace
  methods\n	\n		Basics on algebraic domain decomposition methods (Schur/Sch
 warz)\n	\n		Hybrid direct/iterative methods: motivation and description of
  methods\n	\n		Hierarchical parallel implementation and scalability issues
 \n\nHands-on session with the MaPhys and HIPS solvers. Learning outcomes\
 nParticipants will understand the main features of the main direct and ite
 rative methofd\, their tradeoffs and when using them is most advisable.\nT
 hey will also gain practical experience with some state of the art solvers
 Prerequisites\nBasic knowledge of linear algebra and parallel algorithms\n
 Knowledge of a programming language (Fortran\, C\, C++)\nUse of Linux (Uni
 x) \n\nhttps://events.prace-ri.eu/event/121/
SUMMARY:Parallel linear algebra @ MdS
URL;VALUE=URI:https://events.prace-ri.eu/event/121/
END:VEVENT
END:VCALENDAR