I'm again intrigued by the time this process takes with your setup, since for example it takes 7 seconds here to process 36 names exam with 35 questions (the same source file you posted at RE: AMC performance on postcorrection). But maybe you have thousands of students?
Perhaps you can create a second project with only the names you need to correct, adding to your CSV file (the one which contains the students names) two columns studentnumber and checkcode with the student number and check code (the first and last numbers in the +2/1/58+ format at the top of each page) from the first project. Then you set the student number and check code at the very beginning of the \onecopy command with

\setnumbers{\studentnumber}{\checkcode}

\makeatletter
\def\setnumbers#1#2{\AMCid@etud=#1\AMCid@check=#2\advance\AMCid@check\@ne}
\makeatother

Thank you.

Regarding the time the process is taking to complete: any way to trace it, or get a log file with information that could help to discover the problem?

Perhaps you can post the log file produced building the documents (with less students...) in debugging mode (from Help menu), but I'm afraid it will be difficult to get something from it.

I have noticed the process slows after DOC-catalog.pdf and amc-compiled.pdf appear inside the project folder. If I review the command output details, the last line is /usr/share/texmf/tex/generic/pgf/basiclayer/pgfcorelayers.code.tex. It has been so every time the process has taken over an hour to complete.

Alexis Bienvenüe wrote:

I'm again intrigued by the time this process takes with your setup, since for example it takes 7 seconds here to process 36 names exam with 35 questions (the same source file you posted at RE: AMC performance on postcorrection). But maybe you have thousands of students?

No, I only have about 165 students. This is very strange. Maybe a particular LaTeX default setup on openSUSE? No idea.