lunduniversity.lu.se

Denna sida på svenska This page in English

Installation

Installation

OverviewInstallationRunning examples BenchmarkAndre group

Installation

Please note that fiber diffraction modules will be available as soon as Rosetta weekly builds will be restored!

Download Rosetta files

Download Rosetta files using the following link:

Download Rosetta

For smaller systems this version may provide satisfactory computation times and one can go directly to the Compile Rosetta section at the bottom of this page. For larger systems and faster performance it is recommended to follow instructions below:

Install NFFT library [OPTIONAL]

NFFT ((Nonequispaced fast Fourier transform) requires FFTW3 library that can be downloaded from FFTW3 webpage and installed following instructions. FFTW3 should be installed with -fPIC, which can be done by passing following arguments to configure script.

./configure --prefix=/path/to/fftw3 --exec-prefix /path/to/ffw3 --enable-shared=yes CPPFLAGS=-fPIC

Obtain NFFT from the webpage: NFFT webpage Unless otherwise needed, it is recomended to configure NFFT library using a command:

./configure --prefix=/path/to/nfft --with-pic --enable-shared=fftw3 --enable-nnfft --with-fftw3 = /path/to/fftw3

Run setup script [OPTIONAL]

The setup script is available at:

Download setup script

go to your Rosetta/source directory and execute a script providing NFFT library location:

python setup.py -n /path/to/nfft -f /path/to/fftw -r /path/to/Rosetta

The script will merge NFFT library with your Rosetta distribution and prepare files for compilation.

Setting up GPU card [OPTIONAL]

We provide GPGPU computations with NVIDIA Cuda library. setup.py script (described above) sets up necessary files. If you have already executed this script you are ready to compile Rosetta (next step).

scons -jX bin mode=release extras=cuda

Compile Rosetta

You can compile Rosetta with fiber diffraction module using the standard command:

scons bin mode=release

where X is the number of processors used for compilation. More details about installation can be found at installation guidelines.