Installation of boost fails

Asked by Ciara Byers on 2020-11-12

Hi,

I'm trying to install madgraph onto my account on my university's cluster, which I mention only to point out that lots of other people are running madgraph on the same 'machine' as me with no problems, and I'm having difficulty installing the HEPTools for it.

I installed pythia- pgs (I realise this is outdated but I was trying to follow this tutorial " https://twiki.cern.ch/twiki/pub/Main/KUCMStHqAnalysis/MadGraph_Guide.pdf " exactly), this was successful, then MadAnalysis5, again successful. MadAnalysis5 prompted me to update it and install delphes along with it, this too was successful.

Then I tried to install boost, because every previous time I tried I had problems with it, or LHAPDF (which gives errors related to boost). The error I get is the following:

"Enter boost directory and install boost
Building Boost.Build engine with toolset gcc...
Failed to build Boost.Build build engine
Consult 'bootstrap.log' for more details
/scratch/cb27g11/MG5_aMC_v2_6_7/HEPTools/HEPToolsInstallers/installBOOST.sh: line 16: ./b2: No such file or directory
Finished installing boost"

Indeed, I cannot find a file called "b2", should this be being downloaded by the installer, or should it be coming with the madgraph download itself?

Upon trying to install LHAPDF I have the same error:

"Enter boost directory and install boost
Building Boost.Build engine with toolset gcc...
Failed to build Boost.Build build engine
Consult 'bootstrap.log' for more details
/scratch/cb27g11/MG5_aMC_v2_6_7/HEPTools/HEPToolsInstallers/installBOOST.sh: line 16: ./b2: No such file or directory
Finished installing boost

    > Now aborting installation of tool 'lhapdf6'.
Command "install lhapdf6" interrupted with error:
InvalidCmd : Installation of lhapdf6 failed."

The same thing happens if I try to install lhapdf5 instead.

I note the errors say that they are trying to build using gcc, I don't know why this would fail, gcc is definitely installed:

"[cb27g11@cyan01 MG5_aMC_v2_6_7]$ gcc --version
gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-23)"

I can only assume this version is fine since other students have no problem using madgraph on the cluster. Is the path to gcc not meant to be found automatically? Do I need to specify it somewhere?

I'm trying to install MG5v2.6.7 if that is relevant at all. I'm also in an anaconda environment with python=2.7 (this is why I am not installing a later version of madgraph, so that I only need to have one version of python loaded).

Is there a way to install an older version of boost perhaps? I'm quite lost with how to proceed since the log files just have what is printed to the terminal in them and nothing more.

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Solved by:
Ciara Byers
Solved:
2020-11-12
Last query:
2020-11-12
Last reply:

This question was reopened

Ciara Byers (ciarab93) said : #1

Apologies, I thought the fact that "gcc --version" gave me an accurate version meant that the module was usable. It turns out I needed to specifically load it first. So no madgraph problems here!