Announcement

Collapse
No announcement yet.

GNU Octave 4.2 Advances As MATLAB Alternative

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • GNU Octave 4.2 Advances As MATLAB Alternative

    Phoronix: GNU Octave 4.2 Advances As MATLAB Alternative

    GNU Octave 4.2 is now available as a major update to this scientific programming language that remains largely compatible with the widely-used MATLAB...

    http://www.phoronix.com/scan.php?pag...GNU-Octave-4.2

  • #2
    Nice to see this project advancing. Last time I used it, it still lacked some functionality found in the Matlab Signals Processing Toolkit, but I wouldn't be all that surprised if it's added the few missing things since then.

    Comment


    • #3
      Not my favorite. A lot of packages should be installed and loaded by default. It's also slow and uses python on symbolic calculations (very slow). The biggest thing I miss in octave and python is a valid symulink alternative and a way to write level 2 functions.
      The thing that I don't like in Matlab is the ~= instead of !=, why Matlab?
      There is no perfect solution, especially when using gradient base optimization. Thanks Matlab for separating gradient base optimizations and genetics optimizations in two different packages, and the slow lambda functions,,,

      Comment


      • #4
        For a symulink alternative you should check out openmodelica, https://openmodelica.org/.

        Comment


        • #5
          MATLAB alternative or just a (cheaper) Clint? Hell, they even replicate the color scheme. I see tons of innovation in this project (sarcasm).

          Comment


          • #6
            Clint? No, clone. It's annoying not be able to modify messages...

            Comment


            • #7
              Summary of important user-visible changes for version 4.2: --------------------------------------------------------- ** The parser has been extended to accept, but ignore, underscore characters in numbers. This facilitates writing more legible code by using '_' as a thousands separator or to group nibbles into bytes in hex constants. Examples: 1_000_000 == 1e6 or 0xDE_AD_BE_EF ** The parser has been extended to understand binary numbers which begin with the prefix '0b' or '0B'. The value returned is Octave's default numeric class of double, not at unsigned integer class. Therefore numbers greater than flintmax, i.e., 2^53, will lose some precision. Examples: 0b101 == 5 or 0B1100_0001 == 0xC1 ** gnuplot 4.4 is now the minimum version supported by Octave.
              Already obsolete! Alright!

              Comment


              • #8
                Originally posted by boffo View Post
                The thing that I don't like in Matlab is the ~= instead of !=, why Matlab?
                With all the problems MATLAB has, that is what sticks out at you?

                Comment


                • #9
                  Originally posted by Marc Driftmeyer View Post

                  Already obsolete! Alright!
                  Seriously, what's the problem with raising a minimum version?
                  It's like saying "OK, this software will now require Linux 3.2 instead of Linux 2.6". No, really, your reaction here is inappropriate.

                  The problem I have here is that they seem to drop a lot of functionality in the name of Matlab compatibility. Octave has a lot of merit on its own, and I tend to prefer using it over Matlab. Here, quite a lot of changes are breaking (not that hard to fix, though). I understand why they are doing it, though. "Octave is shit, my script didn't run on it" is a sentence I heard too often, while the only required changes were to import a toolkit, change a function name, or to sum up, do a little bit of research. I don't know if trying to play catch up with Matlab is the right solution, but that's certainly one. The other would be to add some flags to enable a "Matlab compatibility mode", but that's at the expense of seeing the language splitting in two -- Matlab and Octave, so that's not really feasible. And of course, a lot of scripts that use undocumented Matlab features or Java bindings won't work anyway.

                  Some performance improvements on long computations could be useful, though. Octave is faster in some scenarios, but slower when it comes to highly vectorized code. Looks like multithreading optimizations are not enabled by default.

                  Comment


                  • #10
                    Originally posted by TheBlackCat View Post

                    With all the problems MATLAB has, that is what sticks out at you?
                    Because I don't have an US keyboard.

                    Comment

                    Working...
                    X