dappelha

MultiGridMatlab

Easy to understand Matlab code for geometric multigrid. Great for learning multigrid, with plots at every step.

gpu-tips

Fortran examples of CUDA and directives tips and tricks for IBM Power + Nvidia Systems

fft-overlap

Examples of efficient implementations of ffts on multiple GPUs and across multiple nodes. Overlapping data transfer on multiple levels.

openmp-laplace

Progressive tutorial on how to use OpenMP 4.X to solve laplace equation on GPUs.

summit-scripts

Example submission scripts for LSF and jsrun on the SUMMIT and Sierra supercompters.

hpgmg-fv

Improvements to hpgmg-cuda and the spechpc hpgmg openacc benchmark

BoxLib-omp4-Hackathon2016

IBM OpenMP4 hackathon repository

cuda-tutorial

Example of progressive steps of CUDA parallelization and optimization of a kernel.

dappelha.github.io

fortran-templating

Example of how to use python preprocessor to created function templating in Fortran. Especially useful for runtime selection of one of several functions generated with compile time constants (better compiler codegen).

nvtx-examples

Collection of simple examples and tests for NVIDIA's SDK nvtx range markers.