problems in run.sh (ver 1.0)
tmisawa opened this issue · 2 comments
I am now trying to use SpM ver 1.0 in the supercomputer at ISSP (sekirei).
In the last part of fermion/run.sh (maybe part for plotting ?),
I found the error messages as shown in the followings.
Is it necessary to use "epstopdf" for generating the pdf files ?
Is there another way to plot the results without "epstopdf" ?
Thank you in advance for your cooperation.
================================
'find_lambda_opt.dat'
*** optimal lambda = 1.585e-02 (l=19)
'spectrum.dat'
'lambda_dep.dat'
plotting...
plot "SV.dat"u 1:2 title"" w p ls 99,
^
"../../plt/SV.plt", line 22: function to plot expected
sh: epstopdf: command not found
plot "iter.dat"u 1:2 title"diff(x, x_{old})" w lp ls 1, ""u 1:3 title"(for L1 regular.) primary residual" w lp ls 2, ""u 1:4 title"dual residual" w lp ls 3, ""u 1:5 title"(for non-negative) primary residual" w lp ls 4, ""u 1:6 title"dual residual" w lp ls 5,
sh: epstopdf: command not found
^
"../../../plt/lambda_fix/iter.plt", line 27: function to plot expected
================================
@tmisawa
Thank you for your feedback.
Yes, as you say, epstopdf is used to generate PDFs in run.sh.
It converts EPS files generated by gnuplot to PDFs.
Since epstopdf is not always installed in the system and EPS would be enough for test run, we are going to make this optional.
The script run.sh will generate EPS files by default, or plotting itself will become optional.
We changed the default graph format from pdf to eps.
We confirmed that it now works on sekirei.
This update was released as v1.0.1.
https://github.com/SpM-lab/SpM/releases/tag/v1.0.1
Thank you again for your report.