The Partensor library provides efficient algorithms for processing tensors of very large dimensions, their optimal implementation in parallel environments through the development of an integrated software toolbox, and their use in major special applications, such as fMRI.
More information about the library can be found in our site.
The project is funded by the European Regional Development Fund under the Operational Program "Competitiveness – Entrepreneurship – Innovation" (NSRF 2014-2020). It is implemented under the action "Research-Develop-Innovate", and running under the Special Management and Implementation Authority for Research, Technological Development and Innovation Actions by the Ministry of Education, Research and Religious Affairs (MIA-RTDI).
The title of this project is "Parallel algorithms and implementations for large scale tensors", with code T1EDK - 03360.
- Technical University of Crete (TUC)
- Liavas P. Athanasios
- Karakasis Paris
- Kolomvakis Christos
- Papagiannakos John
- Siaminou Nina
- Neurocom SA
- Tsalidis Christos
- Lourakis Georgios
- Lykoudis George
In order to use PARTENSOR library, first clone this repository and then follow the instructions.
A detailed manual can be found in the documentation directory. From there open index.html file in the browser of your preference. Otherwise, a pdf of the manual can be found and downloaded in Repository page of our site. Both the html document and pdf describe in detail the project's API. Moreover, it includes many useful examples for the supporting functionality.
As mentioned in the manual, some libraries are necessary for the library to work. After downloading them, the user can then test or use the toolbox in other projects. Also, the user can compile and run the examples that exist in test directory.
Before using the toolbox make sure to specify the location of these libraries from your system to the corresponding variables in CMakeLists.txt.