MathLib

This is my hobby C++ header-only math library including Analysis, Linear Algebra, Signal Processing. This library is utilized and tested in my several private projects.

1. Prerequisites

1.1. Essential

C++14 or above compiler

1.2. Recommended

Unit test relies on several build tools. If you just want to use this library, you do not have to run test cases and setup the following tools.

CMake
Google Test

Following pages and shell-scripts will help you install Google Test framework.
- Google Testを導入してみた
- Setup GoogleTest on Windows.md

2. Setup

You just have to copy header files into your workspace and include them.

3. Build & Run test

On Linux

./scripts/doDebugBuild.sh
./scripts/doTest.sh

On Windows Power Shell

./scripts/doDebugBuild.bat
./scripts/doTest.bat

You can clean output files by running doClean.sh / doClean.bat, or simply deleting ./build directory.

4. files

file	description
include/common.hpp	common settings switch for bug hunting mode (default is OFF)
include/analysis.hpp	Analysis library fast version of `sin`, `cos`, `atan2` with 5,9-coefficients polynomial approximation fast square-root using table lookup (for embedded special processor)
include/linAlg.hpp	Linear Algebra library real and complex number type are supported matrix and vector scaling, addition, multiplication, transpose, conjugate, shaping, etc. vector inner/outer product LDL decomposition linear equation solver (using Gaussian elimination, LDL decomposition) functions to print real/complex vector/matrix on console. `printRealVec`, `printComplexVec`, `printRealMat`, `printComplexMat` are useful in debug.
include/sigProc.hpp	Signal Processing library convolution exponential smoothing filter stable recursive moving average
include/RlsFilter.hpp	Recursive Least Square filter

5. Assumptions

It is assumed that the entries of any matrix are aligned on memory in row-oriented order (so called C-style array).

6. Performance option

include/common.hpp has a macro constant MATH_LIB_INLINE_AGGRESSIVELY, which is defined as true in default. When this macro is defined as true, some relatively-small functions are FORCIBLY in-line expanded. You can find which functions are expanded by searching #if MATH_LIB_INLINE_AGGRESSIVELY in MathLib include directory. You may have to switch this macro to false when you have to keep the program size small due to lack of memory space.

7. Canary Mode

include/common.hpp has a macro constant MATH_LIB_ENABLE_CANARY_MODE, which is defined as false in default. When this macro is defined as true, some functions in MathLib performs costly parameter validations such as range check, and when encounter invalid parameters, the functions print error message to std::cerr and exit with EXIT_FAILURE.

When the validity of parameters are guaranteed, parameter validations are just waste of time. It is recommended to enable Canary Mode in experimental stage, and disable when the application becomes stable.