Suscribete

  • Suscribete a Novas CESGA

HPC User Portal

  • HPC User Portal

Estado de los sistemas

  • Solo visible desde Firefox o Chrome.

Compromiso con la igualdad

  • Igualdad CESGA

6º Foro Transfiere 2017

Síguenos ...

  • Twitter FacebbokFlickrYouTube CESGA

Servicios PYME's

  • Servicios para Empresas

díxitos Diciembre

Training Intel Parallel Studio XE 2018

 0 voto(s)

lunes 15/01/2018 12:39

The objective of this course is to provide support for compiling appications with Intel compilers and optimize, debug and improve their performance by using the tools made available by Intel as: Advisor, Inspector, analyzer and vtune.

 The course is recomended for users who develop their own applications and/or want to improve their parallel performance.

Dates: January 29th - 31st 

Times: 10:00 - 13:00h. 14:00 - 17:00h.

REGISTRATION

Day 1 : The goal of this training is to be able to compile, port an application, scalar, MPI or hybrid with last intel compiler and execute it on the cluster. We will speak about intel options and runtime option to compile and optimize a fully parallel application.

  • Intel Compiler (New Options and features, How to port, How to report, How to optimiz, labs)
  • MPI 3 (Introduction, Nonblocking and Collective Communication, Neighborhood Collectives, One Sided Communication)
  • Intel MPI Library (Introduction, Tuning, Collective performance, Labs)

Day 2 : Advisor and Inspector help developers to debug and optimize code. Intel Inspector is a power tools for C/C++ developer to analyses thread and memory debugging, Intel Advisor help developers (C/C++ or fortran) to understand where they have to work to vectorize.

  • Intel Advisor (Advisor: Introduction, Compiler diagnostics/Survey Report, Detect problem and recommend how to fix it, Dependencies Analysis, Memory Access Patterns, Roofline, labs)
  • Intel Inspector (Inspector: Introduction, Memory & Thread Debugger, User Intreface, labs)

Day 3 : After the two first days, we will show how to profile code, with light tool and scalable (Snapshot) but also with very precise tool for the both dimension scalability through interconnect with intel trace analyzer and high optimization inside processor with vtune

  • Intel Vtune (Vtune: Introduction, The Software Optimization Process, GUI, labs)
  • Intel Trace Analyzer and Collector(ITAC: Introduction, How to use ITAC, MPI Performance Snapshot, MPI Application Correctness)

Trainer : Dr Christophe Berthelot work for Bull an Atos Company since 1998. From two years he is a member of Atos’s Center for Excellence in Parallel Programming CEPP He deliver training about Intel product from 10 years, worldwide.

Valorar: