News
New release version (1.6.0) ↑ Return on top
A new version is available with many features as OpenCL on NVIDIA and ATI.
New release version (1.6-tuto) ↑ Return on top
The CECAM tutorial of October 2011 is finished and the official version of the tutorial is available for download. The exercices are available in the tar file, and on-line also. The talks are available on-line.
OpenCL support for ATI ↑ Return on top
The version 1.6 will have full support OpenCL for NVIDIA and ATI graphics cards (see performances).
New release version (1.5.2) ↑ Return on top
The version 1.5.2 has minor changes mainly when we combine MPI, openMP and openCL.
New production version (1.5.1) ↑ Return on top
The version 1.5.1 has many new features as metallic material treatment, time-dependent DFT, full support of OpenCL (speedup of 10 with k-points). Developer documentation has been improved by means of Doxygen.
Browse the src source (1.5.1).
Daubechies Wavelets in Electronic Structure Calculation: BigDFT Code Tutorial ↑ Return on top
The BigDFT software will be presented at the CECAM tutorial in Grenoble, October 19th to October 21st 2011.
Yarmouk school for computational condensed matter and nano systems ↑ Return on top
Beside DFT introduction lectures and more specialised lectures on PAW, XC, GW, BigDFT were presented at the Yarmouk school in Jordan, Sunday October 31st to Thursday November 4th.
New production version (1.4.0) ↑ Return on top
The version 1.4.0 has many new features as hybrid functionals, XANES calculations and can use OpenCL library for GPGPU. Documentation has been improved with the inclusion of a manual and tutorials on the website.
Pittsburgh tutorial ↑ Return on top
Two introducing tutorials (basics and surface calculations) have been written for the tutorial to be held in Pittsburgh the 1st of July.
Bull - Joseph Fourier Price ↑ Return on top
Luigi Genovese won the first Bull - Joseph Fourier prize for the code BigDFT!
GPU version available ↑ Return on top
Thanks to Matthieu Ospici and Luigi Genovese, a GPU version is available based on CUDA library. A speedup between 6 and 10 is reached. See two publications.
First result with GPU (Graphics Processor Units) ↑ Return on top
Thanks to Matthieu Ospici and Luigi Genovese, it is possible to use NVIDIA Graphics processor units (GPU) for more than 50% of the program (short convolutions). The speedup factor is roughly 30 for double precision float numbers.
Integration inside ABINIT v5.6 ↑ Return on top
As in the version 5.5, BigDFT is a part of ABINIT. All binaries of this version has the capabilities to use wavelets.
The BigDFT project is over: Welcome to BigDFT! ↑ Return on top
The European project is over but the BigDFT works and is already incorporated in ABINIT v5.5
See this page for a long description of the project
The European project BigDFT is started! ↑ Return on top
BigDFT is a FP6-STREP-NEST European project (n. 511815). The goal is to develop a code based on wavelet functions (systematic multiresolution basis set) which scales only linearly with respect to the system size.
See this page for a long description of the project