6
$\begingroup$

I am trying to load IGraphM on a remote cluster at a university. I can load the program, then install the package, which is successful

I then run <<IGraphM, and get

enter image description here

Any advice? The package usually loads fine, it may be the Linux version is incompatible.

|improve this question
$\endgroup$
  • 3
    $\begingroup$ This may be a better place to ask this. The developer might need information about the cluster's operating system, which version of Mathematica you are using, etc. $\endgroup$ – Jason B. Nov 2 '18 at 13:43
  • $\begingroup$ Thank you, I have contacted the developer with the details. $\endgroup$ – Alexander Kartun-Giles Nov 2 '18 at 14:06
  • 2
    $\begingroup$ Yes, the version of Linux you are using may be incompatible. For interested onlookers: this will be sorted out behind the scenes, and answered here at a later point. $\endgroup$ – Szabolcs Nov 2 '18 at 14:13
  • $\begingroup$ Ok thank you. Is there some way of loading the new proximity graphs in manually? Without loading IGraphM? That might be a workaround. $\endgroup$ – Alexander Kartun-Giles Nov 2 '18 at 14:15
6
$\begingroup$

I created a new release, v0.3.102, which improves compatibility with various Linux distributions.

With Mathematica 11.2 or later, the easiest way to install it is

PacletInstall["https://github.com/szhorvat/IGraphM/releases/download/v0.3.102/IGraphM-0.3.102.paclet"]

This one-liner is convenient to run on a remote machine. For earlier Mathematica versions, use the update script to avoid the hassle of manual downloading.

I believe that any Intel x86_64 Linux distro should work for as long as it has glibc 2.14 or later.

The glibc version can be checked by running

/lib/libc.so.6

Your libc.so.6 may be in another location. Try locate libc.so.6 if necessary.

|improve this answer
$\endgroup$
  • $\begingroup$ Thanks. This works well. I was able to install this on the cluster using the usual method. $\endgroup$ – Alexander Kartun-Giles Nov 6 '18 at 16:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.