this is for holding javascript data
Xavier Andrade edited Parallelization.tex
over 9 years ago
Commit id: 7baafcf45b47d7dd17c0a5f196bd6ce9535aa392
deletions | additions
diff --git a/Parallelization.tex b/Parallelization.tex
index 9325ee8..8e77580 100644
--- a/Parallelization.tex
+++ b/Parallelization.tex
...
structure codes must execute efficiently in modern computational
platforms. This implies support for massively parallel platforms and
modern parallel processors, including graphics processing units
(GPUs).
Octopus has been shown to perform efficiently
both on parallel
supercomputers~\cite{Andrade_2012,Alberdi_2014}. Octopus supercomputers with scaling to hundreds of thousands of cores~\cite{Andrade_2012,Alberdi_2014}. The code also has
a very efficient an implementation of GPU
acceleration~\cite{Andrade_2012,Andrade_2013}. acceleration~\cite{Andrade_2012_gpus,Andrade_2012} that has shown to be competitive in performance with Gaussian DFT running on GPUs~\cite{Andrade_2013}.
Performance is not only important for established methods, but also
for the implementation of new features. Ideally, developers should be
isolated as much as possible from the optimization and parallelization
requirements. The simplicity of real-space grids allows us to provide Octopus developers with building blocks that they can use to produce
highly efficient code without caring about the details of the
implementation. In most cases, this building blocks allow developers to write
code that is automatically parallel, efficient, and that can transparently