GNU Octave

The chosen platform for the experimentation herein has been the GNU Octave platform. As a numerical computation platform, it is mostly compatible with comparable platforms, such as MATLAB; however, as GNU Octave is released under a GNU GPLv3 license, the source code was modified for the experiments conducted herein, which resulted in a Modified GNU Octave (M-GNU-O) platform that can better leverage certain accelerants.

In particular, GNU Octave nicely accommodates the need for treatment of matrices as a first class object. By way of example, "it is easy to define a matrix of values in Octave. The size of the matrix is determined automatically" (i.e., N-dimensional array objects are supported).

Source: https://octave.org/doc/v4.0.1/Matrices.html

In contrast, in Python, matrices are a special case (i.e., perhaps, not a first class object); hence, when working with matrices in Python (a.k.a. Python matrices), the NumPy package (a scientific computing package that supports N-dimensional array objects) is often utilized. Likewise, in R, the function matrix() needs to be invoked and the arguments (e.g., m number of rows, n number of columns) need to be specified (i.e., perhaps, not a first class object).

In GNU Octave, ' is matrix transposition; this syntax is quite straightforward and intuitive. In R, the easiest way to effectuate matrix transposition is, perhaps, t(). In Python 3, assuming theArray is defined, the easiest way to effectuate matrix transposition is, perhaps, [*zip(*theArray)]. '

For convenience, the Python version releases are provided in Table 1 below:

Table 1: Python Release Version Information

Python Release Version	Version Release Date
Python 3.9.1	8 December 2020.
Python 3.9.0	5 October 2020.
Python 3.8.7	21 December 2020.
Python 3.8.6	23 September 2020.
Python 3.8.5	20 July 2020.
Python 3.8.4	13 July 2020.
Python 3.8.3	13 May 2020.
Python 3.8.2	24 February 2020.
Python 3.8.1	18 December 2019.
Python 3.8.0	14 October 2019.
Python 3.7.9	17 August 2020.
Python 3.7.8	27 June 2020.
Python 3.7.7	10 March 2020.
Python 3.7.6	18 December 2019.
Python 3.7.5	15 October 2019.
Python 3.7.4	08 July 2019.
Python 3.7.3	25 March 2019.
Python 3.7.2	24 December 2018.
Python 3.7.1	20 October 2018.
Python 3.7.0	27 June 2018.
Python 3.6.12	17 August 2020.
Python 3.6.11	27 June 2020.
Python 3.6.10	18 December 2019.
Python 3.6.9	02 July 2019.
Python 3.6.8	24 December 2018.
Python 3.6.7	20 October 2018.
Python 3.6.6	27 June 2018.
Python 3.6.5	28 March 2018.
Python 3.6.4	19 December 2017.
Python 3.6.3	03 October 2017.
Python 3.6.2	17 July 2017.
Python 3.6.1	21 March 2017.
Python 3.6.0	23 December 2016.
Python 3.5.10	5 September 2020.
Python 3.5.8	1 November 2019.
Python 3.5.7	18 March 2019.
Python 3.5.6	8 August 2018.
Python 3.5.5	4 February 2018.
Python 3.5.4	25 July 2017.
Python 3.5.3	17 January 2017.
Python 3.5.2	27 June 2016.
Python 3.5.1	07 December 2015.
Python 3.5.0	13 September 2015.
Python 3.4.10	18 March 2019.
Python 3.4.9	8 August 2018.
Python 3.4.8	4 February 2018.
Python 3.4.7	25 July 2017.
Python 3.4.6	17 January 2017.
Python 3.4.5	26 June 2016.
Python 3.4.4	06 December 2015.
Python 3.4.3	25 February 2015.
Python 3.4.2	4 October 2014.
Python 3.4.1	18 May 2014.
Python 3.4.0	16 March 2014.
Python 3.3.7	19 September 2017.
Python 3.3.6	12 October 2014.
Python 3.3.5	9 March 2014.
Python 3.3.4	9 February 2014.
Python 3.3.3	17 November 2013.
Python 3.3.2	15 May 2013.
Python 3.3.1	7 April 2013.
Python 3.3.0	29 September 2012.
Python 3.2.6	11 October 2014.
Python 3.2.5	15 May 2013.
Python 3.2.4	7 April 2013.
Python 3.2.3	10 April 2012.
Python 3.2.2	4 September 2011.
Python 3.2.1	10 July 2011.
Python 3.2	20 February 2011.
Python 3.1.5	9 April 2012.
Python 3.1.4	11 June 2011.
Python 3.1.3	27 November 2010.
Python 3.1.2	21 March 2010.
Python 3.1.1	17 August 2009.
Python 3.1	27 June 2009.
Python 3.0.1	13 February 2009.
Python 3.0	3 December 2008.
Python 2.7.18	20 April 2020
Python 2.7.17	19 October 2019
Python 2.7.16	02 March 2019
Python 2.7.15	30 April 2018
Python 2.7.14	16 September 2017
Python 2.7.13	17 December 2016
Python 2.7.12	26 June 2016.
Python 2.7.11	5 December 2015.
Python 2.7.10	23 May 2015.
Python 2.7.9	10 December 2014.
Python 2.7.8	1 July 2014.
Python 2.7.7	31 May 2014.
Python 2.7.6	10 November 2013.
Python 2.7.5	15 May 2013.
Python 2.7.4	6 April 2013.
Python 2.7.3	9 April 2012.
Python 2.7.2	11 June 2011.
Python 2.7.1	27 November 2010.
Python 2.7	4 July 2010.
Python 2.6.9	29 October 2013.
Python 2.6.8	10 April 2012.
Python 2.6.7	3 June 2011.
Python 2.6.6	24 August 2010.
Python 2.6.5	19 March 2010.
Python 2.6.4	25 October 2009.
Python 2.6.3	2 October 2009.
Python 2.6.2	14 April 2009.
Python 2.6.1	4 December 2008.
Python 2.6	1 October 2008.
Python 2.5.4	23 December 2008.
Python 2.5.3	19 December 2008.
Python 2.5.2	21 February 2008.
Python 2.5.1	18 April 2007.
Python 2.5	19 September 2006.
Python 2.4.4	18 October 2006.
Python 2.4.3	29 March 2006.
Python 2.4.2	28 September 2005.
Python 2.4.1	30 March 2005.
Python 2.4	30 November 2004.
Python 2.3.5	8 February 2005.
Python 2.3.4	27 May 2004.
Python 2.3.3	19 December 2003.
Python 2.3.2	3 October 2003.
Python 2.3.1	23 September 2003.
Python 2.3	29 July 2003.
Python 2.2.3	30 May 2003.
Python 2.2.2	14 October 2002.
Python 2.2.1	10 April 2002.
Python 2.2p1	29 March 2002.
Python 2.2	21 December 2001.
Python 2.1.3	8 April 2002.
Python 2.1.2	16 January 2002.
Python 2.1.1	20 July 2001.
Python 2.1	15 April 2001.
Python 2.0.1	22 June 2001.
Python 2.0	16 October 2000.
Python 1.6	5 September 2000.
Python 1.5.2p2	22 March 2000.
Python 1.5.2p1	6 July 1999.
Python 1.5.2	30 April 1999.
Python 1.5.1p1	6 August 1998.
Python 1.5.1	14 April 1998.
Python 1.5	17 February 1998.
Python 1.4	25 October 1996.

Source: https://www.python.org/doc/versions/

The most current Python release version is 3.9.1.

Previously, GNU Octave-related experimentation was performed using Docker (e.g., https://github.com/epfl-sti/octave-x11-novnc-docker). Other repositories that provide a Docker image for GNU Octave include: https://hub.docker.com/r/mtmiller/octave.

Contemporary approaches for building container images, as Docker is deprecated as of Kubernetes v1.20, is to use container tools, such as Kaniko (an open source tool for building container images from a Dockerfile); It has been noted that Kaniko is more focused on building images in Kubernetes. Other container tools include Podman, Buildah, and Buildkit.

Source: https://towardsdatascience.com/its-time-to-say-goodbye-to-docker-5cfec8eff833

For convenience, GNU Octave release version information is provided in Table 2 below:

Table 2: GNU Octave Release Version Information

GNU Octave Release Version	Release Date
GNU Octave 6.1.0	26-Nov-20
GNU Octave 5.2.0	31-Jan-20
GNU Octave 5.1.0	1-Mar-19
GNU Octave 4.4.1	9-Aug-18
GNU Octave 4.4.0	30-Apr-18
GNU Octave 4.2.2	13-Mar-18
GNU Octave 4.2.1	24-Feb-17
GNU Octave 4.2.0	14-Nov-16
GNU Octave 4.0.3	2-Jul-16
GNU Octave 4.0.1	23-Mar-16
GNU Octave 4.0.0	29-May-15
GNU Octave 3.8.1	4-Mar-14
GNU Octave 3.8.0	31-Dec-13
GNU Octave 3.6.4	21-Feb-13

Source: https://www.gnu.org/software/octave/news.html

The current release version for GNU Octave is 6.1.0.

There are also several opportunities for drop-in acceleration of GNU Octave (e.g., https://developer.nvidia.com/blog/drop-in-acceleration-gnu-octave/).

Interim Findings

Perhaps, GNU Octave can be construed as being more syntactically intuitive for ML. Moreover, the GNU Octave documentation is very nicely oriented for sparse matrices (starting on page 635 of the online document: https://octave.org/octave.pdf).