On 2010-03-04 16:25-0800 David MacMahon wrote:

> I get the feeling that I shouldn't have any numbers to the right of the 
> colons.  It also seems like this build some of the tests, so I think I should 
> run it several times, discard the times from the first run and average the 
> rest.

P.S. to my answer to this comment before.

"make all" builds all libraries and examples

"make test_noninteractive" builds all libraries and examples (if that has
not been done), and also runs them for all devices that output files (if
that has not been done), and finally does a comparison between C Postscript
results and the PostScript results from other bindings.

Thus, if you run make test_noninteractive twice, the second time nothing
should be built or run.  The only action that should occur is the differences
between PostScript results (I think....  I may generate a fake file in
that case so even that comparison is not re-run unless some source file
has been touched).  Anyhow, the second "make test_noninteractive" will be
much shorter than the first, and will not reflect the run times
for the examples, i.e., the results from "time ..." will be bogus for
the second and subsequent runs of this target.

"make test_interactive is different.  Like the test_noninteractive it does
have proper dependencies to insure everything required to run the
interactive tests has been built.  However, it always runs every interactive
test (since there is no file output).  Thus, times measured with it should
be consistent every time it is run (if "make all" is run first to build
every required example and subject to how quickly and consistently you can
exit from GUIs that need some mouse clicking or keyboarding to do that.)

All of this is by way of explanation.  Probably, the whole idea of timing
these noninteractive and interactive test results is a poor one so I
would forget that aspect and instead concentrate on whether the results
of the tests show no obvious errors (segfaults, hangs, etc.)

Alan

__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); PLplot scientific plotting software
package (plplot.org); the libLASi project (unifont.org/lasi); the Loads of
Linux Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Plplot-devel mailing list
Plplot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/plplot-devel

Reply via email to