- Numeric - Haskell.org

Announce: comfort-array, lapack
by Henning Thielemann 29 May '19

29 May '19

I like two announce two of my packages: 1. comfort-array http://hackage.haskell.org/package/comfort-array It provides Boxed and Storable arrays with very liberal shape definitions. You may use ranges of indices like in 'array' or zero-based indexing like in 'repa', but you can also use arbitrary index Sets, enumerations, concatenation of arrays and more. E.g. an array with (Shape.Enumeration Ordering) has three elements with indices LT, EQ, GT. 2. lapack http://hackage.haskell.org/package/lapack A high-level interface to the numerical Linear Algebra package LAPACK. It is based on comfort-array, which means that a vector can have a rich structure. E.g. matrix vector multiplication has the signature: (#*|) :: Matrix.General height width a -> Vector width a -> Vector height a Thus, a 'Matrix.General (Shape.Enumeration Ordering) (h,w) a' would map a rectangular array with dimensions h and w to a three-element vector with indices LT, EQ, GT. The LAPACK interface provides solvers for simultaneous linear equations, linear least-squares problems, eigenvalue and singular value problems for square, triangular, symmetric, Hermitian, banded and banded Hermitian matrices, as LAPACK supports them. For a motivating example refer to: http://code.henning-thielemann.de/bob2019/main.pdf

3 3

Re: [Numeric] Numeric Digest, Vol 10, Issue 1
by Marco Zocca 02 May '18

02 May '18

Thank you Dominic for sharing this work :) An industrial-strength ODE integrator suite like Sundials is certainly a welcome addition for many people. Regarding the `haskell-numerics` github organization ; such a general name sounds like it can accommodate more projects than just HMatrix. I would be happy to be a part of it, and I'm thinking to perhaps move the development of my old PETSc bindings there. Marco Zocca On Mon, Apr 23, 2018 at 2:00 PM, <numeric-request(a)haskell.org> wrote: > Send Numeric mailing list submissions to > numeric(a)haskell.org > > To subscribe or unsubscribe via the World Wide Web, visit > http://mail.haskell.org/cgi-bin/mailman/listinfo/numeric > or, via email, send a message with subject or body 'help' to > numeric-request(a)haskell.org > > You can reach the person managing the list at > numeric-owner(a)haskell.org > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Numeric digest..." > > > Today's Topics: > > 1. New Version of hmatrix (more ODE solvers) (dominic(a)steinitz.org) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Sun, 22 Apr 2018 15:25:18 +0100 > From: dominic(a)steinitz.org > To: Haskell cafe <haskell-cafe(a)haskell.org>, haskell(a)haskell.org, > numeric(a)haskell.org > Subject: [Numeric] New Version of hmatrix (more ODE solvers) > Message-ID: <2AC8F7D0-05AA-4DB7-8311-E7B664A893DF(a)steinitz.org> > Content-Type: text/plain; charset="utf-8" > > I am pleased to announce a new version of hmatrix: 0.19.0.0. > > This is not intended to be a breaking change but a lot of modules have been modified to ensure that continuous integration (which has now been set up) is green. > Support for SUNDIALS <https://computation.llnl.gov/projects/sundials> has been added. It should be possible to replace Numeric.GSL.ODE with Numeric.Sundials.ARKode.ODE and have your program work as before bearing in mind that the methods and error control might differ (even for those with the same names!). > The packages that comprise hmatrix are: > > hmatrix > hmatrix-gsl > hmatrix-special > hmatrix-sparse > hmatrix-glpk > hmatrix-tests > hmatrix-sundials > I am planning to split these up into separate projects to make it easier for contributors. To this end, I have created an hmatrix <https://github.com/haskell-numerics> organisation in github. Please contact me if you wish to become a member. > > I strongly recommend anyone solving ODEs to look as hmatrix-sundials as the underlying package provides many more methods, more documentation and more diagnostic information (great when your 200 variable model fails). It is a truism that any sufficiently sophisticated ODE solver package is a domain specific language (I include the ODE solving part of GSL here). The current interface both to GSL and to SUNDIALS is somewhat rudimentary but will hopefully evovle to a nicer DSL (something at which Haskell excels). > > Dominic Steinitz > dominic(a)steinitz.org > http://idontgetoutmuch.wordpress.com > Twitter: @idontgetoutmuch > > > >

2 4

New Version of hmatrix (more ODE solvers)
by dominic＠steinitz.org 22 Apr '18

22 Apr '18

I am pleased to announce a new version of hmatrix: 0.19.0.0. This is not intended to be a breaking change but a lot of modules have been modified to ensure that continuous integration (which has now been set up) is green. Support for SUNDIALS <https://computation.llnl.gov/projects/sundials> has been added. It should be possible to replace Numeric.GSL.ODE with Numeric.Sundials.ARKode.ODE and have your program work as before bearing in mind that the methods and error control might differ (even for those with the same names!). The packages that comprise hmatrix are: hmatrix hmatrix-gsl hmatrix-special hmatrix-sparse hmatrix-glpk hmatrix-tests hmatrix-sundials I am planning to split these up into separate projects to make it easier for contributors. To this end, I have created an hmatrix <https://github.com/haskell-numerics> organisation in github. Please contact me if you wish to become a member. I strongly recommend anyone solving ODEs to look as hmatrix-sundials as the underlying package provides many more methods, more documentation and more diagnostic information (great when your 200 variable model fails). It is a truism that any sufficiently sophisticated ODE solver package is a domain specific language (I include the ODE solving part of GSL here). The current interface both to GSL and to SUNDIALS is somewhat rudimentary but will hopefully evovle to a nicer DSL (something at which Haskell excels). Dominic Steinitz dominic(a)steinitz.org http://idontgetoutmuch.wordpress.com Twitter: @idontgetoutmuch

1 0

combinatorics of battleship game
by Henning Thielemann 06 Feb '18

06 Feb '18

Hi friends, how many possibilities are there to place ships on a board according to the rules of the battleship game? Years ago I answered that question using a Haskell program. It needed some more years to motivate myself to finish a paper to describe my approach. Now, I managed at least to write it in German. Anyone like to proof-read? http://code.henning-thielemann.de/battleship-combinatorics/paper.pdf Best, Henning

1 0

Re: [Numeric] Numeric Digest, Vol 8, Issue 1
by Marco Zocca 16 Dec '17

16 Dec '17

Very interesting! This is a great foundation for higher-level libraries. I also very much appreciate the use of bare FFI calls as a way of future-proofing this library, since the various binding DSLs are quite impervious to Haskell newcomers. Would you see `vector` as more fitting than `CArray` ? Also, I'm guessing that the original comments would help bring up to speed other developers on the internals. Last thing, would you be interested in jointly developing bindings to the sparse netlib functionality ( http://www.netlib.org/sparse/) ? Thanks, Marco > I have just uploaded these packages: > > https://hackage.haskell.org/package/netlib-ffi-0.0 > https://hackage.haskell.org/package/netlib-carray-0.0 > https://hackage.haskell.org/package/blas-ffi-0.0 > https://hackage.haskell.org/package/blas-carray-0.0 > https://hackage.haskell.org/package/lapack-ffi-0.0 > https://hackage.haskell.org/package/lapack-carray-0.0 > https://hackage.haskell.org/package/lapack-ffi-tools-0.0 > > The netlib packages provide definitions used by the other packages. The > "ffi" packages provide foreign imports of all functions from the > corresponding netlib package. The "carray" packages contain mid-level > functions that use CArrays for vectors and matrices and try to get input > and output values right according to the FORTRAN comments. CArrays might > not be really appropriate since we always use zero-based arrays but this > cannot be expressed in the type. A zero-based variant of CArrays might be > better in the future. I use closed-world classes for generalizing > functions that are generalizable. This way, I can add "methods" without > extending the class. > > The packages use very basic techniques. No C code, no HSC, no C2HS, no > template Haskell, only plain Haskell 2010 with ForeignFunctionInterface. > > I still have no good idea how to cope with Characters. They are used like > enumeration elements, thus the enumerations deserve custom newtype > wrappers. But not all enumeration elements are allowed everywhere. I am > also uncertain whether to use these wrappers already in the "ffi" packages > or only in "carray". > > I have also not included the FORTRAN comments in the Haskell packages. I > have only included links to online docs where available. >

2 1

autogenerated BLAS and LAPACK interfaces
by Henning Thielemann 16 Dec '17

16 Dec '17

I have just uploaded these packages: https://hackage.haskell.org/package/netlib-ffi-0.0 https://hackage.haskell.org/package/netlib-carray-0.0 https://hackage.haskell.org/package/blas-ffi-0.0 https://hackage.haskell.org/package/blas-carray-0.0 https://hackage.haskell.org/package/lapack-ffi-0.0 https://hackage.haskell.org/package/lapack-carray-0.0 https://hackage.haskell.org/package/lapack-ffi-tools-0.0 The netlib packages provide definitions used by the other packages. The "ffi" packages provide foreign imports of all functions from the corresponding netlib package. The "carray" packages contain mid-level functions that use CArrays for vectors and matrices and try to get input and output values right according to the FORTRAN comments. CArrays might not be really appropriate since we always use zero-based arrays but this cannot be expressed in the type. A zero-based variant of CArrays might be better in the future. I use closed-world classes for generalizing functions that are generalizable. This way, I can add "methods" without extending the class. The packages use very basic techniques. No C code, no HSC, no C2HS, no template Haskell, only plain Haskell 2010 with ForeignFunctionInterface. I still have no good idea how to cope with Characters. They are used like enumeration elements, thus the enumerations deserve custom newtype wrappers. But not all enumeration elements are allowed everywhere. I am also uncertain whether to use these wrappers already in the "ffi" packages or only in "carray". I have also not included the FORTRAN comments in the Haskell packages. I have only included links to online docs where available.

1 0

Banded Matrices
by Dominic Steinitz 13 Apr '17

13 Apr '17

Hi Henning, I just noticed https://hackage.haskell.org/package/hmatrix-banded. I have just made a PR to hmatrix to add a tridiagonal solver: https://github.com/albertoruiz/hmatrix/pull/224. I don’t think there is anything to do but I thought it worth sharing. Dominic Steinitz dominic(a)steinitz.org http://idontgetoutmuch.wordpress.com

1 0

New Haddock
by Dominic Steinitz 06 Jun '16

06 Jun '16

I couldn’t find an announcement of the new version of Haddock. You can now document your libraries even more beautifully. For example, > -- \[ > -- f(k;n,p) = \Pr(X = k) = \binom n k p^k(1-p)^{n-k} > -- \] will render as https://hackage.haskell.org/package/random-fu-0.2.7.0/docs/Data-Random-Dist… <https://hackage.haskell.org/package/random-fu-0.2.7.0/docs/Data-Random-Dist…> Also don’t forget you can put really nice diagrams in your documentation e.g. http://hackage.haskell.org/package/diagrams-contrib-1.3.0.7/docs/Diagrams-E… <http://hackage.haskell.org/package/diagrams-contrib-1.3.0.7/docs/Diagrams-E…> and http://hackage.haskell.org/package/diagrams-contrib-1.3.0.7/docs/Diagrams-T… <http://hackage.haskell.org/package/diagrams-contrib-1.3.0.7/docs/Diagrams-T…>. Dominic Steinitz dominic(a)steinitz.org http://idontgetoutmuch.wordpress.com

1 0

ANN: linear-circuit, comfort-graph, resistor-cube
by Henning Thielemann 26 Aug '15

26 Aug '15

I have uploaded the package linear-circuit that computes the resistance of a network of resistors using linear algebra from hmatrix: https://hackage.haskell.org/package/linear-circuit This package also demonstrates the use of a graph structure that is an alternative to fgl: https://hackage.haskell.org/package/comfort-graph The resistance computation is a generalization of a small example for computing the resistance of a cube of resistors: https://hackage.haskell.org/package/resistor-cube

1 0

ANN: hmm-hmatrix - efficient Hidden Markov Model implementation
by Henning Thielemann 19 Aug '15

19 Aug '15

My HMM implementation is on Hackage and has already proven to be useful: https://hackage.haskell.org/package/hmm-hmatrix It supports Discrete and Gaussian models and can be extended to other moduls using the type classes from Distribution module. The package implements supervised and unsupervised training, as well as a training using a predefined distribution and patterns. It also supports mixing of trainings. Trained models can be read from and written to CSV. The features are demonstrated by three simple examples. https://en.wikipedia.org/wiki/Hidden_Markov_Model

2 2