| February 21, 2010 11:00 PM PST | |
| What If Home | Product Overview | Intel® TM ABI specification | Technical Requirements FAQ | Primary Technology Contacts | Discussion Forum | Blog |
October 2011: This WhatIf project has been retired, but will remain here for reference.
The Intel® Ordinary Differential Equation Solver Library (Intel® ODE Solver Library) is a powerful, cross-platform tool set for solving initial value problems for Ordinary Differential Equations. It offers explicit, implicit, and mixed solvers for non-stiff, stiff, and ODE problems with variable stiffness. The solvers have a set of configurable parameters that allow experienced users to tune the algorithms for both better accuracy and performance. The Intel® ODE Solver Library is available for both Windows* and Linux*.
The Intel® ODE Solver Library offers:
- Explicit ODE solver – Multi-stage solver with extended stability domains for non-stiff and middle-stiff ODE problems
- Implicit ODE solvers – L-stable solvers for stiff ODE problems supporting both automatic and user-defined Jacobi matrix computations
- Hybrid ODE solvers – Efficient solvers for the ODE problems with unknown or variable stiffness supporting both automatic and user-defined Jacobi matrix computations as well as automatic choice of the appropriate integration method
- Universal ODE solver – Powerful interface to all three above mentioned types of solvers that permits experienced users and researchers tune the solver to fit their needs for performance.
Our solvers explore some new approaches for solving ODE problems with an advanced stability control in explicit schemes. The main goals for releasing these solvers to the public are to provide our customers with powerful ODE solvers and to get feedback on what you do or don’t like about them. We appreciate your help, which will make future versions of these solvers, or other Intel software products, even better.
As the capabilities in the solvers are experimental, we cannot guarantee any level of support for them. Some of the features and interface designs may find their way into released and supported products, some may not.
The current version of the solvers is built as an archive, which can be unpacked and used in your environment.
Please see the Release Notes for more details on technical requirements, including the list of supported processors and operating systems.
1. How do I get started using the Intel® ODE Solver Library?
There is only 1 thing that we recommend you do before starting to use this tool. Make sure that you have reviewed the Installation and Startup Guide. This guide provides a detailed explanation on how to work with Intel® ODE Solver Library.
2. Where can I get support for the use of this utility?
We encourage you to visit our Intel® Ordinary Differential Equations Solver Library forum for support.
3. What are the licensing terms that spell out how exactly I can use this utility?
The licensing terms are listed on the download page.
Alexander Kalinkin (PhD) is a Sr. Software Engineer at the MKL Linear Solvers team in the Developer Products Division. His current role involves support and development of math software for Intel® MKL. His research interests include various aspects of Computational Mathematics including but not limited to Cluster Solvers.
Sergey Gololobov
Sergey Gololobov (PhD) is a Sr. Software Engineer at the Intel Performance Lab in the Developer Products Division. His current role involves support and defining new directions for next generation of math software for Intel® MKL. His research interests include various aspects of Computational Mathematics with the main focus on Iterative Solvers.
For more complete information about compiler optimizations, see our Optimization Notice.
Comments (23) 
| December 25, 2008 11:27 PM PST
SERGEY G (Intel)
|
Thank you, DEDA! We hope you enjoy our software. Sergey G |
| May 13, 2009 2:13 AM PDT
Pat R |
You guys are awesome. (In general, I love Intel software, compilers, libraries - it is of top notch quality.) I thank you for this library and the fact that I won't be stuck using numerical recipes. I remember implicit methods and variable stiffness algorithms when I was learning about numerical mathematics at princeton. It is great to see you guys have implemented this into a library. Very nice work. |
| May 24, 2009 6:57 PM PDT
SERGEY G (Intel)
|
Thanks a lot for the nice words! We really appreciate it! Sergey G |
| August 6, 2009 11:04 AM PDT
Frank Chi |
I download this library to try to solve my system using intel visual fortran compiler. It is easy to use and compile sucessfully. But the speed is slow comparing with my current using package dlsode solver. My system is complicated, it has around 2000 equations, so the size of the dpar is huge. I am wondering if I should be carfule particular when I use the intel ODE solver to deal with such large system. |
| August 21, 2009 2:32 AM PDT
woleakande |
Hello Sirs. Am a new member of your team and here is my credentials(I think you should know me): Am a graduate of the university of Ibadan (Nigeria) and am about to start my Msc. I can write any type of code but i specialise in theoretical (computational) physics. I love Fortran the most but i can also write in java,C++,Basic and VB. I cannot deceive you, I don't have good materials to meet up with your requirements. I will appreciate if I can get some ebooks and softwares to work with.Latest cutting-edge softwares are difficult to come about in this country. Please I need a supervisor to guide me through. Am sorry if all my contributions will have to be mathematial programing. I will start by uploading my undergraduate project for a start. I hope I will impress you. Thanks for the oppurtunity. |
| November 19, 2009 8:21 AM PST
ReenshAves | This is a great library! I was looking for something like that for a very long time. Thanks a lot! |
| March 31, 2010 4:54 PM PDT
enrico | Thank you very much for this package, I am successfully using it on a Windows machine and it works extremely well... Is there by chance a version for OSX? |
| May 11, 2010 10:17 PM PDT
Alexander Kalinkin (Intel)
|
Hi ReenshAves, Thank a lot for using ODE and good comments about it! With best regards, Alexander Kalinkin |
| May 11, 2010 10:22 PM PDT
Alexander Kalinkin (Intel)
|
Hi enrico, Thanks a lot for using ODE and nice word about it. Now ODE package support only Windows and Linux but I will include you request about supporting OSX in next releases. With best regards, Alexander Kalinkin |
| September 10, 2010 2:05 AM PDT
Kirill Osipov |
Hi, Can you please provide the x64 implementation? Thanks in advance |
| December 22, 2010 6:00 PM PST
zmi007
|
Used to solve some stiff ODE's and it just works. Amazing job! THANKS! Спасибо! |
| December 24, 2010 8:15 AM PST
Ray Wood | Really like the solver. I'm using it to solve sets of radioacative decay decay equations with additional input terms and non-zero initial conditions, and it works very well. It is easy to integrate into larger FORTRAN modules and runs quite fast (depending on which routine I use) on my system (quad-core AMD PhenomII, 8GB ram). I see a post above requesting a 64-bit version - I downloaded what looks like a 64-bit version (libiode_intel64.lib). I've only used the 32-bit library so far, but is this not the 64-bit library for Windows? As for updates - in paraticular I'd like to see some info on using this in multi-threaded, multi-core applications or High Performance Computing system. Excellent stuff gentlemen. |
| April 5, 2011 7:11 AM PDT
Alexander Kalinkin (Intel)
|
Kirill, Did you try to use library: intel_ode/lib/intel64/libiode_intel64.a ? With best regards, Alexander Kalinkin |
| April 5, 2011 7:17 AM PDT
Alexander Kalinkin (Intel)
|
Gentelmen, Again, thanks a lot for good words about ODE library. To continue support this library we really need to collect some feedbacks about this library from you. Please send to me some information about your project/ company if you can and what additional functionality do you want to have in this library? With best regards, Alexander Kalinkin |
| April 5, 2011 12:03 PM PDT
compila
|
Passing of 'void* userdata' to rhs function is strongly needed. Now I have to put userdata at the end of y array. Thanks! |
| April 5, 2011 1:00 PM PDT
David |
I downloaded this library awhile ago and finally got time to try it out. Very impressive! Easy to use, and very fast - what not to love!? Like "Compila" I need the ability to pass information into the rhs function. An additional argument (void* userData) would be perfect. Thanks, in advance. |
| April 7, 2011 6:30 AM PDT
Jorge Martinis
| Good work! It would have a positive impact if included to the Intel MKL. As mentioned before, adding a mechanism to pass user data to the rhs function is strongly recommended. |
| April 8, 2011 4:46 PM PDT
Gennadiy Rishkin
| Excellent library. Well-designed, powerful and easy to use. I'll also echo the comments by Jorge above that a mechanism for passing user data to the rhs function is vitally needed. |
| May 4, 2011 6:50 AM PDT
Sören Plönnigs
| Great work guys! Please add the library to the MKL and also add the possibility to exchange userdata to the rhs function. |
| July 27, 2011 11:45 AM PDT
oleglebedev
| Does it work well with complex system of ODE? Must I perform some simplification to reduce Im and Re parts increasing an order of the system by 2, respectively? And so, I solve Im and Re parts of ODE separately. |
| August 8, 2011 3:43 AM PDT
Alexander Kalinkin (Intel)
|
Hi, Current version of Intel ODE doesn't support complex data so to compute complex system you need to reduce it on Re and Im parts. With best regards, Alexander Kalinkin |
| January 8, 2012 4:26 AM PST
Gnomilodon |
Hi, I use Intel ODE library for MOL applications. What are restrictions for a commercial utilisation ? Is it true that Intel ODE developpement is stop (MKL integration) ? Thanks |
Trackbacks (0)
Leave a comment 
| Yuri Laevsky (Intel) | ||
SERGEY G (Intel)
|



DEDA