Calling Fortran LAPACK Library from C#

Introduction

From the LAPACK Website:

LAPACK - Linear Algebra PACKage

LAPACK is written in Fortran 90 and provides routines for solving systems of simultaneous linear equations, least-squares solutions of linear systems of equations, eigenvalue problems, and singular value problems. The associated matrix factorizations (LU, Cholesky, QR, SVD, Schur, generalized Schur) are also provided, as are related computations such as reordering of the Schur factorizations and estimating condition numbers. Dense and banded matrices are handled, but not general sparse matrices. In all areas, similar functionality is provided for real and complex matrices, in both single and double precision.

I have been writing code professionally in C# since around 2001 and I love the language. I also love math and the thought of recreating algorithms in C# that somebody had already coded seemed like a big waste of time.

I stumbled upon LAPACK since I was working on a mapping app that needed to solve thousands of simultaneous equations. I figured why not use code that had been battle tested for years.

One problem it was written in Fortran. So I boned up on Fortran using the gfortran compiler on Linux and used the C/PInvoke feature of C#. (See my first foray at the git below:) to call Fortran routines which use the iso_c_binding so that the subroutines can be called by name from c (Fortran name mangles subroutines by default)

GroupTheorist12/CSharpFortran

iso_c_binding is a standard intrinsic module which defines named constants, types, and procedures for C interoperability.

iso_c_binding

The code in this git exploits this interoperability to call Fortran routines from C# using P/Invoke (Platform Invocation Services). Below is a Fortran subroutine we wish to call:

This example can solve a series of simultaneous equations.

Below is the C# code which will call the Fortran subroutine.

Example P/Invoke C# code to call Fortran Subroutine.

Setting up your environment

The assumption is that you have gfortran, make and the .NET 5.0 SDK installed either on Linux or on Windows in the WSL Linux subsystem. If not, directions for Ubuntu are given below.

Install gfortran

sudo apt-get install gfortran

Install make

sudo apt-get install build-essential

Install .NET 5.0 SDK

Install the .NET SDK

Example C# console project

Let’s walk through the steps to create a C# .NET 5.0 project to call the above Fortran subroutine.

Create console app using dotnet cli.

Open a terminal window and enter the following (I am doing it from my HOME directory):

$ dotnet new console -o CSharpCallingLaPack

Change directories to CSharpCallingLaPack

$ cd CSharpCallingLaPack

Edit Program.cs

With your trusty editor (I am using VSCode) edit the Program.cs and cut and paste the code from below into it:

Copy LAPACK libraries to your usr/local/lib directory

I have provided the libraries on my github site so you can just wget them into your /usr/local/lib directory.

Enter the following into a terminal window.

Create Fortran Source File lapack_module.f90

Change back to the CSharpCallingLaPack directory and with your favorite editor cut and paste the Fortran code below and save it.

Create Fortran Shared Library

Enter the following in your terminal window to create the Fortran shared library.

Run the dotnet build command

Enter the following in your terminal window to build the C# project and copy the Fortran module

$ dotnet build
$ cp lapack_module.so bin/Debug/net5.0

Run the app

Run the following from the terminal window to run the app.

$ dotnet run

You should see the following output:

From fortran
[1.00   3.00]

Final Remarks

I was surprised by how much I enjoyed working with Fortran and the huge amount of mathematical library code that is written in it. The cool thing about Fortran is that optimizing compilers (Intel) can really speed up operations that involve huge arrays and matrices.

The source code for the above project and other projects I am working on can be found at:

[CSharpLaPack​] https://github.com/GroupTheorist12/CSharpLaPack

Thanks for reading!