PARDISO

PARDISO 7.2 Solver Project (December 2020)

The package PARDISO is a thread-safe, high-performance, robust, memory efficient and easy to use software for solving large sparse symmetric and unsymmetric linear systems of equations on shared-memory and distributed-memory multiprocessors.

Features of PARDISO 7.2

PARDISO 7.2 was released in December 2020. It contains full support of multi-threaded Schur-complement computations, parallel selected inversion, incremental updates, and many others accelerations. Some of the new features have been used, e.g., in the following papers:

Additional publications related to PARDISO are available on the webpage of the research group of Prof. Olaf Schenk, Advanced Computing Laboratory, Institute of Computing, USI Lugano, Switzerland.

Important: Please note that the Intel MKL version of PARDISO is based on our version from 2006 and that a lot of new features and improvements of PARDISO are not available in the Intel MKL library.

Benchmarks of PARDISO 7.2

MKL PARDISO vs PARDISO 7.2 Performance Comparison, NXP Semiconductors

MKL PARDISO
PARDISO 7.2

PARDISO 7.2 is benchmarked against Intel MKL PARDISO version 2020 on up to 16 cores on an Intel Xeon E7-4880. The matrices are from NXP's industrial circuit simulator Mica. The plots shows the performance acceleration for the factorization and solution phase against MKL PARDISO.



MKL PARDISO vs PARDISO 7.2 Performance Comparison, Silvaco Circuit Simulator

MKL PARDISO
PARDISO 7.2

PARDISO 7.2 is benchmarked against Intel MKL PARDISO version 2020 on up to 32 cores on an Intel Xeon Intel E5-2650 with 2.30GHz architecture. The matrices are from Silvaco industrial circuit simulator. The plots shows the performance acceleration for the factorization and solution phase against MKL PARDISO.



PARDISO 7.2 k-rank Update Analysis, NXP Semiconductor

1 core; y=0.33k
8 cores; y=0.21k
Full factorization

Regression analysis on the k-rank update LU factorization in PARDISO. The scatter plot shows the number of k-rank updates and the corresponding factorization time in milliseconds. The regression analysis cleary demonstates a linear trend both for the single and the multiple core version. The dashed line shows the time for the full factorization.

How To Use PARDISO

Examples

Additional Downloads

BLAS

xerbla

In some BLAS libraries the Fortran routine xerbla.f or the C routine xerbla.c are not included. However, the BLAS routine xerbla must link to I/O routines, which are compiler dependent. Thus, we provide the implementation, which you can compile and link in a compiler dependent fashion.

Binaries / Libraries

Current Available Libraries

Version 7.2 (Architecture x86-64, 64-bit)

Compiler Operating System PARDISO Libraries
gcc/gfortran 8.3.1 (GLIBC 2.17) Linux libpardiso720-GNU831-X86-64.so
icc/ifort 19.1 (GLIBC 2.26) Linux libpardiso720-INTEL120-X86-64.so

Other libraries can be compiled upon request. Please let us know in case that some of these libraries are not working for you. You can only download the files above if you are a registered user and agreeing to the license conditions.

Matpower Libraries

Ipopt 3.12.9 (extended version), PARDISO 6.0 (threaded version)

You need to set the numbers of threads with, for example, if you would like to use 8 cores you should use export OMP_NUM_THREADS=8 before running ipopt.mexa64 in Matlab.

In addition, you need a PARDISO license file pardiso.lic which can be obtained here.

Architecture Operating System Files
Architecture x86-64, 64-bit, Matlab R2016b Linux ipopt.mexa64, ipopt_auxdata.m, ipopt.opt, examplehs071.m
Architecture x86-64, 64-bit, Matlab R2016b macOS ipopt.mexmaci64, ipopt_auxdata.m, ipopt.opt, examplehs071.m

Download & License

Academic Licenses / Corporate Licensing Program

Here you can dowload a cost-free academic license or a corporate license for the current release of PARDISO. The option corporate is only available for selected corporate or commercial entities. Please choose the desired license type:

License Type Period of Validity License
12 months, user-locked, but host-unlocked license PDF
12 months, user-locked, but host-unlocked license PDF

News

  • December 2020
    Release of Version 7.2 including faster LU updates, full support of AMD orderings, several important bug fixes, improved pivoting for symmetric indefinite matrices. (release notes)
  • April 2019
    Release of Version 6.2 including fast supernodal incremental LU updates, improved factorization routines, and approximate minimum degree orderings (release notes)
  • May 2018
    Release of Version 6.0 for the R-INLA project including supernodal data compression and selected inversion software (release notes)
  • November 2015
    Release of new webpage, including user map and list of successful recent consulting projects.).
  • January 2014
    Release of Version 5.0.0, including new Schur-complement solver and selected inversion software (release notes).
  • April 2011
    Release of Version 4.1.2, including the distributed memory solver (release notes).
  • January 2011
    Release of Version 4.1.0 (release notes).
  • March 2010
    Map of pardiso users
  • October 2009
    Release of Version 4.0.0 (release notes).
  • November 2007
    The solver is now available for both academic and commercial use.
  • July 2007
    The solver PARDISO is now available under www.pardiso-project.org.

References

In case that you are using the PARDISO 7.2 please cite:

  1. C. Alappat, G. Hager, O. Schenk, J. Thies, A. Basermann, A. Bischop, H. Fehske, G. Wellein, A Recursive Algebraic Coloring Technique for Hardware-Efficient Symmetric Sparse Matrix-Vector Multiplication, ACM Transactions on Parallel Computing, Vol. 7, No. 3(19), 2020.
  2. M. Bollhöfer, O. Schenk , R. Janalik, S. Hamm, K. Gullapalli, State-of-The-Art Sparse Direct Solvers, Parallel Algorithms in Computational Science and Engineering. Modeling and Simulation in Science, Engineering and Technology, 1-32, Birkhäuser, 2020.
  3. M. Bollhöfer, A. Eftekhari, S. Scheidegger, and O. Schenk Large-Scale Sparse Inverse Covariance Matrix Estimation, SIAM J. Sci. Comput., 41(1), A380–A401, 2019.

Companies

We provide services in the field of computational mathematics, scientific computing and high-performance computing including:

PARDISO Industrial Customers - Selected Multinational Software Corporations

Major multinational software corporations have selected PARDISO as their software of choice. Over 10,000 industrial and academic users in more than 50 countries around the world trust in PARDISO for fast computational operations and robustness.
  • NXP
    Integration of PARDISO 7.2 into NXP circuit simulators (February 2021).
  • AutoForm Engineering AG
    Integration of PARDISO 7.2 into AutoForm's automobile sheet metal forming tools (January 2021).
  • Dassault Systèmes
    Integration of PARDISO 7.2 into Dassault Systèmes - a Fortune 50 large software company (January 2021).
  • Williams Grand Prix Engineering
    Integration of PARDISO 7.0 into Williams Grand Prix Engineering simulation tools (November 2020).
  • Red Bull Advanced Technology
    Integration of PARDISO 6.2 into Red Bull Advanced Technology simulation tools (February 2020).
  • Siemens Digital Industries
    Integration of PARDISO 6.2 into Siemens Digital Industries Software simulation tools (January 2020).
  • MSC Software Corporation
    Integration of PARDISO 6.2 into MSC Software Corporation simulation tools (July 2019).
  • Cadence
    Integration of PARDISO 6.2 into Cadence circuit simulation tools (August 2019).
  • AniForm Engineering
    Integration of PARDISO 6.2 into AniForm Engineering process simulation tools (April 2019).
  • Hexagon AB
    Training and general solver recommendation for precision measuring technology product simulations.
  • Marathon Oil Company
    Training and general solver recommendation for the Marathon Oil Reservoir Simulation product.
  • CST
    Integration of PARDISO 6.0 into CST's electromagnetic design software.
  • MOSEK
    Integration of PARDISO 5.0.0 into MOSEK's optimization software for solving mathematical optimization problems.
  • Endress+Hauser Flowtec AG, Reinach, Switzerland
    Integration of PARDISO 5.0.0 into finite-element tools at Endress+Hauser Flowtec AG.
  • U.S. Naval Observatory
    Integration of PARDISO 5.0.0 into simulation tools at the U.S. Naval Observatory.
  • SMS Siemag AG
    Integration of PARDISO 5.0.0 into simulation tools for metallurgical plant and rolling mill technology.
  • Frustum
    Integration of PARDISO 5.0.0 into Frustum's Distributed Optimization Service.
  • Silvaco
    Integration of PARDISO 5.0.0 into Silvaco simulation tools.
  • JCM Wave
    Integration of PARDISO 5.0.0 into JCMsuite, a complete Finite Element software package for the computation of electromagnetic waves, elasticity and heat conduction.
  • Saudi Aramco
    Integration of PARDISO 4.1.2 into reservoir simulation tools at Saudi Aramco.

For licensing questions related to PARDISO 7.2, please send an e-mail to
info@pardiso-project.org.

Map of PARDISO Academic and Industrial Users
(in 2020)

Contact

Prof. Olaf Schenk
Advanced Computing Laboratory
Faculty of Informatics, USI, Lugano
Switzerland
Email: info@pardiso-project.org
Tel: +41 58 666 4850
SkypeID: olaf.schenk
USI Lugano, Switzerland