octave-bug-tracker
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Octave-bug-tracker] [bug #55407] [octave forge] (interval) mpfr_matrix_


From: Mike Miller
Subject: [Octave-bug-tracker] [bug #55407] [octave forge] (interval) mpfr_matrix_mul_d performance test is problematic
Date: Sun, 6 Jan 2019 17:55:27 -0500 (EST)
User-agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/71.0.3578.98 Safari/537.36

URL:
  <https://savannah.gnu.org/bugs/?55407>

                 Summary: [octave forge] (interval) mpfr_matrix_mul_d
performance test is problematic
                 Project: GNU Octave
            Submitted by: mtmiller
            Submitted on: Sun 06 Jan 2019 02:55:26 PM PST
                Category: Octave Forge Package
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: Other
                  Status: None
             Assigned to: None
         Originator Name: 
        Originator Email: 
             Open/Closed: Open
         Discussion Lock: Any
                 Release: dev
        Operating System: GNU/Linux

    _______________________________________________________

Details:

The unit test in mpfr_matrix_mul_d.cc that attempts to test for performance
improvement has a couple of problems, I think it should be removed entirely.

Recently, a unit test in the symbolic package that attempted to measure and
compare run time was demonstrated to fail on a CI system. This is because the
test inherently assumes the system load will be constant while the test runs.
But if other jobs are running at the same time, this can easily fail.

The second problem is that the test relies on the Octave internal function
'__octave_config_info__' which shouldn't be used by any code outside of
Octave.

If there is a legitimate need for user code to know that OpenMP is enabled in
Octave, maybe we can add a documented 'openmp' function or similar to Octave.
But since this test should be removed anyway, that seems unnecessary at this
point.




    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?55407>

_______________________________________________
  Message sent via Savannah
  https://savannah.gnu.org/




reply via email to

[Prev in Thread] Current Thread [Next in Thread]