NMatrix is an experimental linear algebra library for Ruby, written mostly in C and C++. It can be used with or without SciRuby, but is part of the SciRuby project.
NMatrix was inspired by and based heavily upon NArray, by Masahiro Tanaka.
<img src=https://www.pledgie.com/campaigns/15783.png?skin_name=chrome>
Please be aware that SciRuby and NMatrix are ALPHA status. If you’re thinking of using SciRuby/NMatrix to write mission-critical code, such as for driving a car or flying a space shuttle, you may wish to choose other software (for now).
You should also be aware that NMatrix and NArray are incompatible with one another; you should not try to require both at the same time. Unfortunately, that causes problems with Ruby/GSL, which currently depends upon NArray. As such, we are working on a patch for Ruby/GSL.
The following features exist in the version of NMatrix (0.0.2) currently under development, and may or may not exist in 0.0.1 (the current gem):
-
Matrix storage containers: dense, yale, list (more to come)
-
Data types: uint8, int8, int16, int32, int64, float32, float64, complex64, complex128, rational64, rational128 (incomplete)
-
Conversion between storage and data types (except from-complex, and from-float-to-rational)
-
Element-wise operations and comparisons
-
Matrix-matrix multiplication for dense (using ATLAS) and yale
-
Matrix-vector multiplication for dense (using ATLAS)
-
Dense and list matrix slicing and referencing
-
Matlab .MAT v5 file input
-
C and C++ API
-
Level 3 BLAS internal implementations (no library) and ATLAS (with library) access:
-
xGEMM
-
xTRSM
-
-
LAPACK ATLAS access:
-
xGETRF (experimental)
-
These features will exist in our second alpha release, 0.0.2 (incomplete, thus not yet available as a gem):
-
Yale matrix slicing and referencing
-
LU decomposition
-
LAPACK internal implementations (no library needed):
-
xGETRF
-
The estimated release date for 0.0.2 is the end of August 2012.
These are features planned for NMatrix 0.1.0, our first beta.
-
calculation of determinant, trace, and eigenvalues (characteristic polynomial) (0.1.0)
-
Gaussian elimination
-
exponentials and square roots
-
matrix inversions
-
matrix decomposition/factorization
-
calculation of norms
-
tensor products
-
principal component analysis (PCA)
-
improved file I/O
-
operation scheduling
-
parallelization of some types of operations
-
optimization of non-BLAS data types on BLAS-like operations (e.g., matrix multiplication for rational numbers)
-
Ruby/GSL interoperability
For full instructions, please see sciruby.com/nmatrix. Generally, you should be able to do
gem install nmatrix
However, you will need to install ATLAS with CBLAS first. Those directions can be found at our website.
NOTE: The above command will not give you 0.0.2, which is not yet complete. To obtain 0.0.2, you should use the following sequence of commands:
git clone https://github.com/SciRuby/nmatrix.git cd nmatrix/ rake compile rake repackage gem install pkg/nmatrix-0.0.2.gem
If you get errors about clapack.h or cblas.h, figure out where your ATLAS headers are using:
locate clapack.h
Then, tell your system:
export C_INCLUDE_PATH=/usr/local/atlas/include
Finally, try compiling again.
-
ATLAS and LAPACK, probably
-
GCC 4.2
-
Ruby 1.9
See Synopsis (above) for now. Gem coming later.
More detailed installation instructions are available at here
Code in the master branch of SciRuby/nmatrix on github should compile and link, but is not necessarily stable. You might also check out the dev branch if master hasn’t been updated in some time.
git clone https://github.com/mohawkjohn/nmatrix.git
Before commiting any code, you MUST read our Contributor Agreement.
Copyright © 2012, The Ruby Science Foundation.
All rights reserved.
NMatrix, along with SciRuby, is licensed under the BSD 2-clause license. See LICENSE.txt for details.
Support a SciRuby Fellow:
<img src=https://www.pledgie.com/campaigns/15783.png?skin_name=chrome>