[Blindmath] Octave testing so far.

Doug and Molly Miron mndmrn at hbci.com
Mon Feb 6 20:23:35 UTC 2017


Good day All,

Generally speaking, Octave uses the same symbols for operations as Matlab  does.  Arithmetic operators and matrix formation seem to be identical.  However, there are little differences.  I ran a numerical integration program I’d written and run under my 20-year old 32-bit Matlab in Octave and got a strange warning message and the wrong result.  Like many error messages, it didn’t make sense but did identify the program line that triggered it.  It turns out that looping constructs which end in Matlab with a “end” statement terminate in Octave with the name of the construct, e.g. “endfor” in my case.  After I fixed that problem, the function ran correctly.  A second caution is execution time.  For the case I ran the time in Matlab was 0.61 sec, while the Octave time was 3.58 sec, or nearly six times longer.  Given that the electromagnetic problem I’ve been running takes about 52 min with the old Matlab, I hesitate to run in Octave.  On the other hand it was gratifying to be able to define a 16 GB matrix on the 64-bit machine.  As mentioned in a previous post, the Octave manual is present on the web.  I found its index useful in tracking down my mystery warning.

Regards,
Doug Miron


Regards,
Doug Miron



More information about the BlindMath mailing list