))
9/10 - Distributed Memory and Parallel Computing
Manage episode 473331704 series 3624244
Contenu fourni par Peter Schmidt. Tout le contenu du podcast, y compris les épisodes, les graphiques et les descriptions de podcast, est téléchargé et fourni directement par Peter Schmidt ou son partenaire de plateforme de podcast. Si vous pensez que quelqu'un utilise votre œuvre protégée sans votre autorisation, vous pouvez suivre le processus décrit ici https://fr.player.fm/legal.
Marc Hartung and I will discuss distributed memory in parallel computing in this episode, with tools like OpenMPI. We also discuss some of the hardware aspects around HPC systems and how shared and distributed memory computations differ.
Links:
- https://www.open-mpi.org OpenMPI homepage
- https://docs.open-mpi.org/ the docs for OpenMPI
- https://www.mpi-forum.org The MPI Forum (who write the MPI standard)
- http://openshmem.org/site/ OpenSHMEM
- https://en.wikipedia.org/wiki/Distributed_memory summary page on distributed memory
- https://en.wikipedia.org/wiki/InfiniBand InfiniBand network solution
- https://www.nextplatform.com/2022/01/31/crays-slingshot-interconnect-is-at-the-heart-of-hpes-hpc-and-ai-ambitions/ Slingshot network solution
- https://en.wikipedia.org/wiki/Partitioned_global_address_space
- https://www.techtarget.com/whatis/definition/von-Neumann-bottleneck the bottleneck named after John von Neumann
- https://en.wikipedia.org/wiki/Floating_point_operations_per_second overview of FLOPS (floating point operations per second)
- https://www.openmp.org/wp-content/uploads/HybridPP_Slides.pdf OpenMP and OpenMPI working together in a hybrid solution
- https://blogs.fau.de/hager/hpc-book Georg Hager/Gerhard Wellein book on HPC for scientists and engineers
This podcast is brought to you by the Advanced Research Computing Centre of the University College London, UK.
Producer and Host: Peter Schmidt
20 episodes
Partager
