1
Observations on optimizing scientific computing applications, or What to do if there is no bigger hammer ? Dr Axel Kohlmeyer, Sr. Scientific Computing Expert, International Centre for Theoretical Physics, Trieste, Italy Tuesday May 14, 16:00, Room MXC 315 Abstract: Computational tools are becoming increasingly common in many areas of research and at the same time researchers are working on increasingly complex problems, which require using highperformance computing resources and optimized and parallelized applications. However, making an application for a given problem run faster may not always be straightforward. This talk will discuss some example cases that highlight the different possible approaches ranging from postcompilation optimizations to performing a complete rewrite from scratch.

Observationsonoptimizingscientific, …‰COLE POLYTECHNIQUE FÉDÉRALE DE LAUSANNE THEORY AND SIMULATION OF MATERIALS Title Microsoft Word - 2013_05_Seminar.Kohlmeyer.docx Author

  • Upload
    vanthu

  • View
    219

  • Download
    2

Embed Size (px)

Citation preview

                                       

Observations  on  optimizing  scientific    computing  applications,  or    

What  to  do  if  there  is  no  bigger  hammer  ?      

Dr  Axel  Kohlmeyer,  Sr.  Scientific  Computing  Expert,    International  Centre  for  Theoretical  Physics,  Trieste,  Italy  

     

Tuesday  May  14,  16:00,  Room  MXC  315          

Abstract:  Computational  tools  are  becoming  increasingly  common  in  many  areas  of  research  and  at  the  same  time  researchers  are  working  on  increasingly  complex  problems,  which  require  using  high-­‐performance  computing  resources  and  optimized  and  parallelized  applications.      However,  making  an  application  for  a  given  problem  run  faster  may  not  always  be  straightforward.  This  talk  will  discuss  some  example  cases  that  highlight  the  different  possible  approaches  ranging  from  post-­‐compilation  optimizations  to  performing  a  complete  rewrite  from  scratch.