Slidecast: Introduction to hybrid Cray XC30 course

CSCS organized last March (24-27 March) a 4-day training course on “Piz Daint”, CSCS hybrid Cray XC30 system. “Piz Daint” has 5.272 compute nodes (with Intel® Xeon® E5-2670 and NVIDIA® Tesla® K20X) and a peak performance of 7.8 Petaflops. The presentations have being given by experts from CrayNVIDIA and Allinea.

The Cray Programming Environment on Cray XC30, Hart Alistair (Cray)

Using the Cray XC30 Piz Daint, Themis Athanassiadou (CSCS)

New User & Usage Interfaces on Piz Daint, Sadaf Alam (CSCS)

An introduction to GPU hybrid programming on Cray XC30, Hart Alistair (Cray)

Understanding Communication and MPI on Cray XC30, Hart Alistair (Cray)

Understanding Cray MPI, Hart Alistair (Cray)

Hybrid Programming on Cray XC30, Hart Alistair (Cray)

Overview of Compilers on Cray XC30, Alfio Lazzaro (Cray)

Serial Performance Optimizations, Alfio Lazzaro (Cray)

Introduction to application analysis with CrayPAT, Alfio Lazzaro (Cray)

Using CrayPAT to understand parallel performance, Alfio Lazzaro (Cray)

Introduction to at-scale GPU debugging with Allinea DDT, Patrick Wohlschlegel (Allinea)

Introduction to at-scale GPU debugging with Allinea DDT: Examples (1/2), Patrick Wohlschlegel (Allinea)

Introduction to at-scale GPU debugging with Allinea DDT: Examples (2/2), Patrick Wohlschlegel (Allinea)

A brief compendium of GPU-accelerated numerical libraries (1/2), William Sawyer (CSCS)

A brief compendium of GPU-accelerated numerical libraries (2/2), William Sawyer (CSCS)

Cray Scientific Libraries: Overview and Performance, Alfio Lazzaro (Cray)

Petascale resources and CP2K: Enabling sampling, large scale models or correlations beyond DFT, Joost VandeVondele (ETH Zurich)

GPUs and DCA+: a match made in heaven, Peter Staar

Getting COSMO ready for Piz Daint, Oliver Fuhrer (MeteoSwiss)