Code review comment for lp:~maddevelopers/mg5amcnlo/2.1.2-lhapdf6

Revision history for this message
marco zaro (marco-zaro) wrote :

Ciao Olivier,
but are these problems related to the lhapdf6 branch or they are also there in 2.1.2?
The changes i did in lhapdf6 are very minimal…
I’ll have a look anyway.
Can oyu please check that if one does one run with internal pdfs and the next one with lhapdf, the libpdf and libgeneric are correctly recompiled with the lhapdf related wrappers?
It seems not to be the case for madevent runs…
I’ll have a look later.

Cheers,

Marco

On 21 Aug 2014, at 13:49, Olivier Mattelaer <email address hidden> wrote:

> Review: Needs Fixing
>
> Hi,
>
> 2.1.3 is for bug fixing only. If we need to rush a release for political reason.
> So this should be for 2.2.0
>
> Note that I can not test for lhapdf6 since I can't install lhapdf6 (and do not want to change the configuration of the machime for that).
>
> I have test however the lhapdf5 one, and it doesn't work ...
> Note sure what the problem is but he code didn't recognize correctly the lhapdf mode.
>
> If I do the following it works nicely:
>
> generate p p > e+ e-
> launch
> set pdlabel lhapdf
> set lhaid 10772
>
> I'm however worry about some misterious print statement which not always appear multiple times.
> This might indicate a side effect and/or that the pdf change during the run...
>
> Here is the log where the see the line twice.
>
> INFO: Using LHAPDF v5.9.1 interface for PDFs
> write compile file for card: /Users/omatt/Documents/eclipse/2.1.2-lhapdf6/PROC_sm_0/Cards/param_card.dat
> run_card missed argument xmtcentral. Takes default: 0.0
> run_card missed argument d. Takes default: 1.0
> run_card missed argument gridrun. Takes default: False
> run_card missed argument fixed_couplings. Takes default: True
> Using random number seed offset = 21
> INFO: Running Survey
> Creating Jobs
> Working on SubProcesses
> P0_qq_ll
> INFO: Idle: 0, Running: 1, Completed: 0 [ current time: 12h15 ]
> DEBUG: Found too many jobs. Recovering
> INFO: Idle: 0, Running: 0, Completed: 1 [ 5s ]
> INFO: End survey
> refine 10000
> INFO: compile directory
> INFO: Using LHAPDF v5.9.1 interface for PDFs <------ WHY HERE?
>
> I guess the main question is why is there also twice the line:
> INFO: compile directory
>
>
> Now they are for sure a side effect since, If I run the following:
>
> Then the two first run go trough but the latest returns 0 cross-section:
> The log indicate the following line:
> Unimplemented distribution= lhapdf
> Implemented are:
> So clearly the switch between the mode are not working in both direction.
>
>
> For lhapdf6 in itself, I will abstain. My understanding was that Paolo was going to check that part, and since I didn't hear about him that's why I didn't include that part in 2.1.2.
>
> Cheers,
>
> Olivier
>
>
>
> --
> https://code.launchpad.net/~maddevelopers/mg5amcnlo/2.1.2-lhapdf6/+merge/231518
> Your team MadDevelopers is subscribed to branch lp:~maddevelopers/mg5amcnlo/2.1.3.

« Back to merge proposal