|
From: | john o goyo |
Subject: | Re: [Gm2] Numerical library in Modula-2 |
Date: | Fri, 17 Nov 2017 16:17:52 -0500 |
User-agent: | Mozilla/5.0 (Macintosh; U; PPC Mac OS X 10.5; en-GB; rv:1.9.2.23) Gecko/20110920 Thunderbird/3.1.15 |
Greetings, Andreas: On 02/11/2017 10:13, Fischlin Andreas wrote:
There are many interesting things here but the releases appear to be binary only. In particular, your download page (at http://www.sysecol.ethz.ch/ramses/downloads.html) indicates that the Solaris Sparc releases were compiled with EPC Modula-2. EPC is gone and I do not have the compiler. Was your source released?Dear Michael, Sounds all quite interesting.In terms of the different REAL types, that is quite an important aspect not trivial. Have you looked at our environment? I believe we have there quite nice and highly portable code that can accommodate many needs in terms of numerics as it depends on precision, yet allowing client code to be quite flexible and needing basically almost no change when changing compilers, hardware, or operating systems, despite the many definitions of elementary types for real numbers.
Sincerely, john
Here some links of interest in this context:DMPortab <http://se-server.ethz.ch/RAMSES/Objects/DM/DMPortab.html#DMPortab> DMFloatEnv <http://se-server.ethz.ch/RAMSES/Objects/DM/DMFloatEnv.html#DMFloatEnv> DMMathLib <http://se-server.ethz.ch/RAMSES/Objects/DM/DMMathLib.html#DMMathLib> DMLongMathLib <http://se-server.ethz.ch/RAMSES/Objects/DM/DMLongMathLib.html#DMLongMathLib> DMConversions <http://se-server.ethz.ch/RAMSES/Objects/DM/DMConversions.html#DMConversions>Above ‘Dialog Machine' modules are implemented in a manner that they adjust to the underlying hardware (e.g. little vs. big Endian storage of REAL numbers for satisfying IEEE floating point standards etc.).Perhaps this is of interest? Modules based on this foundation doing similar things your package does are then also available (SciLib <http://www.sysecol.ethz.ch/ramses/ramses-software-layers/science-library-scilib.html>, AuxLib <http://www.sysecol.ethz.ch/ramses/ramses-software-layers/auxiliary-library-auxlib.html>), but that is another topic, I guess less of interest to you.Regards, Andreas Cited References: ------------------------IEEE Std 754-2008, 2008. IEEE standard for binary floating-point arithmetic. ANSI/IEEE Std 754-2008, (IEEE Product No. SH10116-TBR) http://standards.ieee.org/findstds/standard/754-2008.html Ie003ETH Zurich Prof. em. Dr. Andreas Fischlin IPCC Vice-Chair WGII Systems Ecology - Institute of Biogeochemistry and Pollutant Dynamics Universitaetstrasse 16, CHN E 24 8092 Zurich SWITZERLAND address@hidden <mailto:address@hidden> www.sysecol.ethz.ch/Staff/af <http://www.sysecol.ethz.ch/Staff/af> +41 44 633-6090 phone +41 79 595-4050 mobile Make it as simple as possible, but distrust it! ________________________________________________________________________On 30/10/2017, at 19:12, Michael Riedl <address@hidden <mailto:address@hidden>> wrote:Gaius,noticed that my feedback to the new package had been addressed only to your personal mailbox ...All, I developed a library of numerical procedures in Modula-2 for my privateprojects which I would like to share with others. I already was in email conversation about that with Gaius quite some time ago.Now I am "near" ready with the packages and I think it's time to includethe whole community in the communication - otherwise I will always find stuff to be improved before doing the next step ... such a thing has a tendency to never become "ready".The library has, among other, procedures for - Numerical integration & differentiation - Special functions (Gamma, Student, Hankel, ... ) - Fourier transformation - Random number generators - Matrix routine (Matrix times vector, Matrix times Matrix, ...) - Solving linear equations - Eigenvalue- and vectors (real symmetric, real general, complex hermitian, complex general) - Singular value decomposition - Multidimensional function optimization routinesThe whole library is about 25.000 lines of code (not including the def-files)which I would like to put to the public under the GPL/LGPL. Here I am looking for some support for - further testings and cleanups - get some more systematic in the test routines (not that well structured at the moment) - documentation (most of the def files are in German, here we need to have it at least bi-lingual) - other improvements & checks At the moment I still develop this with Excelsior XDS compiler but I think it's time to change that ... but I suppose that this will not be that big deal at all - maybe the different sizes of REAL types need to be considered, but that's something I would like to discuss. Furthermore I wrote a complete IO library and some applications based on the libraries mostly for chemistry related topics (spectra analysis with principal component analysis and PLS including a library for genetic algorithms, small library for Fourier transform spectroscopy, a smallquantum mechanical calculation routine including an experimental interfaceto OpenMPI).Also would like to share all of the in one of the next steps, but reviewing the about 40.000 lines of code may take some time.At the moment I am still doing some clean-ups but if someone in the audience is already interested in one of the topics drop me a mail and I can send copies via email. Or we already find a nice place to share it for all who might be interested. Gajus already kindly offered me with some support concerning configure and makefiles. Any feedback is welcome. Gruß Michael _______________________________________________ gm2 mailing list address@hidden <mailto:address@hidden> https://lists.nongnu.org/mailman/listinfo/gm2_______________________________________________ gm2 mailing list address@hidden https://lists.nongnu.org/mailman/listinfo/gm2
[Prev in Thread] | Current Thread | [Next in Thread] |