Installation of NanoDCAL with bash scripts

Vahid

I installed meson-0.52.1 and ninja-1.11.1 in $HOME/.nanoacademic/NANODCALPLUS on Beluga cluster of Compute Canada by loading gcc, openmpi and python/3.8 modules. I installed the dependencies (except zofu, which did not install). When installing the fortran part, the installation crashed complaining that meson version should be >=57. So I tried again with meson-0.57.0, installed the dependencies and tried the fortran part again. This time, it crashed complaining that meson-0.52.1 was not found.

What version of meson should I use? The package is nanodcalplus-2023B.

Thanks,
Vahid

nanoadmin

Hi Vahid,

I recently compiled NanoDCAL+ on Beluga. I will email you the information on where to get it to save you the compiling.

For your original question. Python packages that are installed outside of a virtual environment at the user level are put into $HOME/.local/bin. You need to make sure that it is on the $PATH and has priority (is first) over the system install. You can do this as so.

export PATH=$HOME/.local/bin:$PATH

A good way of debugging this issue is with the which command with the -a flag. You need to make sure the meson you expect is first in the list.

Guillaume

Vahid

Hi Guillaume,

Thank you for your timely response. I did install meson-0.52.1 with pip using the --prefix option in the $HOME/.nanoacademic/NANODCALPLUS folder. But the fortran compilation part complained that I need >=0.57 meson version. I did not have any meson in my $HOME/.local/bin folder.

I am actually helping another user in our group (Mohammad) to install this package. He has allocation on Narval and I was hoping to show him how to install the NanoDCAL package (on any CC system) but the installation seems to be very tricky, system dependent and dependency-version dependent.

Thanks,
Vahid