Awesome

YCMatrix

A flexible Matrix library for Objective-C and Swift, that interfaces BLAS, LAPACK and vDSP functions via the Accelerate Framework.

YCMatrix is available for MacOS (10.7+), as well as iOS (8.0+).

Getting started

Adding the framework to your project

Import the project in your workspace, or open the framework project in XCode, build and import the product.

.. or alternatively, just drag+drop the files (.h and .m) that you want to your project, presto.

Importing

YCMatrix defines a module so in newer versions of XCode, you can simply import the framework with:

@include YCMatrix;

Alternatively, you can do:

#import "YCMatrix/YCMatrix.h"

Dependencies

YCMatrix has no dependencies other than system frameworks (namely Foundation.framework and Accelerate.framework).

Importing/Using with CocoaPods (no longer supported)

Unfortunately Cocoapods is no longer supported as of 0.6.8. The decision to end Cocoapods support is related to the increasing time needed for debugging related to Cocoapods support (e.g. Xcode/OS updates, addition of C++ etc.). Dropping Cocoapods support will allow for more time spent to developing the project. We recommend Carthage for managing your project's dependencies. You may easily include YCMatrix in your cartfile as follows:

github "yconst/YCMatrix"

In addition, YCMatrix can be easily included manually into an Xcode project, by copying header and implementation files (and optionally creating a new target).

Usage

Naming Conventions

Methods that result in a new Matrix instance are usually prefixed with "matrixFrom", e.g. -matrixFromSubtraction: . Methods that change the receiver in-place do not have the prefix. In some cases, such as matrix-matrix multiplication, the prefix is not included in the method name, even though it is obvious that the result of the operation is a new Matrix instance.

Example

The snippet below is a basic example of matrix multiplication. It also shows how you can easily create matrices with predefined values.

@include YCMatrix;

Matrix *I = [Matrix identityOfRows:3 columns:3]; // 3x3 Identity
Matrix *C = [Matrix matrixOfRows:3 columns:3 value:2]; // 3x3 filled with 2s
Matrix *S = [I matrixByAddition:C]; // Outputs a new matrix
NSLog(@"Result:\n%@", S);

// Result:
// 3.0  2.0  2.0
// 2.0  3.0  2.0
// 2.0  2.0  3.0

What's in there?

The Framework functionality is split into four files: The base class definition, and three categories:

Matrix.h : YCMatrix class definition and basic operations.
Matrix+Advanced.h : Interface to more advanced LAPACK functions.
Matrix+Manipulate.h : Functions for manipulating rows/columns etc.
Matrix+Map.h : Functions for linearly mapping matrices.

In addition, there is a file that implements functionality related to NSArrays containing Matrices.

Please refer to the docs for a complete overview of the functionality contained in each of the categories.

In addition, YCMatrix comes with many unit tests included. Tests are divided in six files:

YCMatrixTests : General Matrix-related tests.
YCMatrixAdvancedTests : Tests related to higher-level operations (decompositions, inverses etc.).
YCMatrixManipulateTests : Tests related to matrix manipulation operations.
YCMatrixMapTests : Tests related to linear mapping operations.
YCMatrixNSArrayTests : Tests related to NSArray categories.
YCMatrixPerformanceTests: Tests for measuring performance of various operations.

Performance

YCMatrix has been built from the ground up with performance in mind. vDSP, BLAS and LAPACK are being used extensively throughout the framework. In addition, care has been given in memory allocation during initialization and copying.

The test file YCMatrixPerformanceTests.m contains some benchmarks using different addition/subtraction implementations (vDSP, BLAS, and pure C).

License

YCMatrix

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

Acknowledgments

This software relies on the following copyrighted material, the use of which is hereby acknowledged.

soboldata.h and sobolseq.c Part of the NLOpt Software

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

haltondata.cpp

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

Routine for computing the Singular Value Decomposition of column-major matrix A using LAPACK

Author: Luke Lonergan Date: 5/31/08 License: Use pfreely