GMT 4.5.9 faster than 6.4?

Q&A
1

Hi everyone,

I’m working on updating some code from an older version to 6.4, and I noticed that when I run the code with the new version it is slower, not something huge, but it’s like 4 secs slower than before. Is this the usual behavior? I just changed what I had to on syntax, nothing more, the code is pretty much the same.

Thanks in advance :slight_smile:

2

Hi @VEXEnzo – do you have some sample code so I can try to reproduce the problem?

3

Hi, thanks for the help. I will leave here a bit of the code and explain what is the objective as well leaving an image that is the expected result.

gmt psbasemap -R$region -JX15/5 $ba -X3.0 -Y7.0   -K >! $outfile

#POINTS

if ($nobs == 0) then

  cat $siterad | awk '{printf("%10.5f %10.5f\n",$1,$2)}' > tempf.$site.$$

  gmt psxy tempf.$site.$$ -R -JX -Sc0.05 -G$cuo -O -K  >> $outfile

endif

#REC opcional

if ($e == 1) then

  if ((-e $siterec) && ($rec == 1)) then

    set REC = `cat $siterec`

    foreach epoch ($REC)

      set valepo = `echo $epoch $xmin $xmax | awk '{if (($1 > $2) && ($1 < $3)) print(1) ; else print (0)}'`

      set rep = 0

      if ($valepo) then

        foreach item ($rec_array_x)

          if ($item == $epoch) then

            set rep = 1

          endif

        end

        if ($rep == 0) then

          set rec_array_x = ($rec_array_x $epoch)

        endif

      endif

    end

    foreach item ($rec_array_x)

      echo $item $ymin >  tempf.$site.$$

      echo $item $ymax >> tempf.$site.$$

      gmt psxy tempf.$site.$$ -R -JX -W1p,orange,. -O -K  >> $outfile

    end

  endif

endif

ACNS GMT6
ACNS GMT6726×913 127 KB

The code I posted is just want produces the points on the graph and the dotted lines. There is ofc more code but again as I said on the original post I only really updated the GMT version and change from “psxy” to “gmt psxy” nothing more.

4

João, you have to understand that we can do nothing with that script. It’s full of $variables. Can you isolate the part that is slower and make a reproducible example?

We have this old issue but the complain then was that GMT5 was slower than GMT4.

5

Given the generic rewrite of i/o in GMT 5 it is no mystery that 5 and 6 are slower at the expense of greater flexibility such as making GMT/MEX, pyGMT and GMT.jl possible.

6

But are you saying that plotting those 3 subplots takes reasonably more than 4 sec such that those extra ~4 sec are perceived as a slow down?

7

It’s not a huge difference, it’s slower, but nothing that I would consider a problem in any sense. Just wanted to know if it was “normal” for this little performance hit to happen.

Thanks for helping out, I got what I wanted :slight_smile:

8

Tell us if GMT.jl version of that script takes more than 2 secs.