[Scilab-users] Scilab 6.0.0, read()

Clément David clement.david at scilab-enterprises.com
Wed Feb 22 21:26:05 CET 2017


Hi Stefan, Lester,

I confirmed the issue (0.02s in 5.5.2 vs 6.90 in 6.0.0) and on a local build I am able to trace the
issue using the Linux perf tool:

+   58,69%    58,68%  libsciast.so.6.0.0              [.] types::getIndexesWithDims              
+   22,09%    22,08%  libsciast.so.6.0.0              [.] types::ArrayOf<double>::resize         
+   12,18%    12,18%  libsciast.so.6.0.0              [.] types::getIndexWithDims                
+    4,04%     4,04%  libsciast.so.6.0.0              [.] types::ArrayOf<double>::getIndexes     

So it seems that too much resizing occurs :/ . Could you post a bug with your test case please ? 

Thanks,

--
Clément

Le mercredi 22 février 2017 à 18:30 +0100, Stefan Du Rietz a écrit :
> Here you are, Lester! A script that generates a comparatively small 
> data file and then reads it is attached.
> 
> In a small laptop computer with Windows 10 and 32-bit versions of Scilab:
> 
> Scilab-5.5.2:
> 
> size(mat) = [10000  7]
> mat took 0.08 sec to read.
> size(mat) = [10000  7]
> 
> Scilab-6.0.0:
> 
> size(mat) = [10000  7]
> mat took 39.14 sec to read.
> size(mat) = [10000  7]
> 
> Regards
> Stefan
> 
> 
> On 2017-02-22 11:43, Lester Anderson wrote:
> > Haven't got as far as testing the functions, but just starting version
> > 6 compared to 5.5.2 is slower; not sure what is going on. Testing on
> > Windows 8.1 (4 Gb RAM).
> > 
> > Do you have a test script to try, see if the issue can be reproduced?
> > 
> > Cheers
> > Lester
> > 
> > On 22 February 2017 at 10:15, Stefan Du Rietz <sdr at durietz.se> wrote:
> > > Hello,
> > > I got Scilab 6 to work in a Windows 10 laptop. However, my loading of data
> > > with read() took almost half an hour.
> > > From toc() and mprintf():
> > > Loading of data took 26 min 23 sec
> > > 
> > > In Scilab 5.5.2 it took just over a second.
> > > From toc() and mprintf():
> > > Loading of data took 1.219 sec.
> > > 
> > > What happened?
> > > 
> > > Regards
> > > Stefan
> > > _______________________________________________
> > > users mailing list
> > > users at lists.scilab.org
> > > http://lists.scilab.org/mailman/listinfo/users
> > 
> > _______________________________________________
> > users mailing list
> > users at lists.scilab.org
> > http://lists.scilab.org/mailman/listinfo/users
> > 
> 
> _______________________________________________
> users mailing list
> users at lists.scilab.org
> http://lists.scilab.org/mailman/listinfo/users



More information about the users mailing list