SMEFT NLO ttbar - Inconsistent integrals

Asked by Pim Verschuuren on 2020-09-08

Dear experts,
My MC generation process remains stuck at a particular line. Here is my setup:

MG v3.0.3-neworders. ->. I pulled this branch with bazaar onto lxplus. The UFO model I use is only compatible with this beta version for now.

After running ./bin/mg5_aMC I put in:

import model SMEFTatNLO-NLO
generate p p > t t~ QCD=2 QED=0 NP=2 [QCD]
output Output/SMEFT_NLO_tt

I keep Herwig6 as parton shower and only reduce the number of events to a 100 before I launch.

INFO: Update the dependent parameter of the param_card.dat
WARNING: Unknow type of information in the card: decay
INFO: Starting run
INFO: Compiling the code
INFO: Using built-in libraries for PDFs
INFO: Compiling source...
INFO: ...done, continuing with P* directories
INFO: Compiling directories...
INFO: Compiling on 10 cores
INFO: Compiling P0_gg_ttx...
INFO: Compiling P0_uux_ttx...
INFO: Compiling P0_ccx_ttx...
INFO: Compiling P0_ddx_ttx...
INFO: Compiling P0_uxu_ttx...
INFO: Compiling P0_cxc_ttx...
INFO: Compiling P0_dxd_ttx...
INFO: Compiling P0_bbx_ttx...
INFO: Compiling P0_bxb_ttx...
INFO: P0_uxu_ttx done.
INFO: P0_dxd_ttx done.
INFO: P0_gg_ttx done.
INFO: P0_bbx_ttx done.
INFO: P0_ccx_ttx done.
INFO: P0_uux_ttx done.
INFO: P0_bxb_ttx done.
INFO: P0_ddx_ttx done.
INFO: P0_cxc_ttx done.
INFO: Checking test output:
INFO: P0_gg_ttx
INFO: Result for test_ME:
INFO: Passed.
INFO: Result for test_MC:
INFO: Passed.
INFO: Result for check_poles:
INFO: Poles successfully cancel for 20 points over 20 (tolerance=1.0e-05)
INFO: P0_uux_ttx
INFO: Result for test_ME:
INFO: Passed.
INFO: Result for test_MC:
INFO: Passed.
INFO: Result for check_poles:
INFO: Poles successfully cancel for 20 points over 20 (tolerance=1.0e-05)
INFO: P0_ccx_ttx
INFO: Result for test_ME:
INFO: Passed.
INFO: Result for test_MC:
INFO: Passed.
INFO: Result for check_poles:
INFO: Poles successfully cancel for 20 points over 20 (tolerance=1.0e-05)
INFO: P0_ddx_ttx
INFO: Result for test_ME:
INFO: Passed.
INFO: Result for test_MC:
INFO: Passed.
INFO: Result for check_poles:
INFO: Poles successfully cancel for 20 points over 20 (tolerance=1.0e-05)
INFO: P0_uxu_ttx
INFO: Result for test_ME:
INFO: Passed.
INFO: Result for test_MC:
INFO: Passed.
INFO: Result for check_poles:
INFO: Poles successfully cancel for 20 points over 20 (tolerance=1.0e-05)
INFO: P0_cxc_ttx
INFO: Result for test_ME:
INFO: Passed.
INFO: Result for test_MC:
INFO: Passed.
INFO: Result for check_poles:
INFO: Poles successfully cancel for 20 points over 20 (tolerance=1.0e-05)
INFO: P0_dxd_ttx
INFO: Result for test_ME:
INFO: Passed.
INFO: Result for test_MC:
INFO: Passed.
INFO: Result for check_poles:
INFO: Poles successfully cancel for 20 points over 20 (tolerance=1.0e-05)
INFO: P0_bbx_ttx
INFO: Result for test_ME:
INFO: Passed.
INFO: Result for test_MC:
INFO: Passed.
INFO: Result for check_poles:
INFO: Poles successfully cancel for 20 points over 20 (tolerance=1.0e-05)
INFO: P0_bxb_ttx
INFO: Result for test_ME:
INFO: Passed.
INFO: Result for test_MC:
INFO: Passed.
INFO: Result for check_poles:
INFO: Poles successfully cancel for 20 points over 20 (tolerance=1.0e-05)
INFO: Starting run
INFO: Using 10 cores
INFO: Cleaning previous results
INFO: Doing NLO matched to parton shower
INFO: Setting up grids
WARNING: program /afs/cern.ch/work/p/pverschu/private/TopXSec/EFT_Truth_Study/MG_v303-neworders/Output/SMEFT_NLO_tt/SubProcesses/P0_gg_ttx/ajob1 4.0 F 0 0 launch ends with non zero status: 1. Stop all computation
INFO: Idle: 14, Running: 2, Completed: 8 [ current time: 15h11 ]
INFO: Idle: 14, Running: 1, Completed: 9 [ 0.02s ]
INFO: Idle: 14, Running: 0, Completed: 10 [ 0.084s ]

The generation only shows two warnings and then stays stuck at the last output line. A previous post showed a similar problem and pointed to the log-file SubProcesses/P0_gg_ttx/GF4.0/log.txt:

 ERROR: inconsistent integrals #0 0.39522033881000501 0.39522033881000423 3.9352936258306546 1.9748364191593679E-016
           1 17 1 0.56438190180640890
           1 0.39524078806073626
STOP 1
           2 0 1 0.17487563109042450
           3 0 1 2.3898319008196191E-002
           4 0 1 -3.5783079459863025E-002
           5 0 1 -1.1087507559816811E-002
           6 0 1 -2.0961137280989398E-003
           7 0 1 -1.6511852284625307E-003
           8 0 1 -2.8768138241017040E-004
           9 0 1 0.13784451761993766
          10 0 1 4.0435365771617708E-002
          11 0 1 5.6651814610137984E-003
          12 6 1 -6.7337382545329990E-004
           1 -2.0449250732018484E-005
          13 0 1 -3.9352936258306546
          14 0 1 -2.0864714532662563E-004
          15 0 1 -1.2193639699299188
          16 0 1 -2.7986481099079889E-005
          17 0 1 -0.16355702659588642
          18 0 1 6.5663341947430002E-004
          19 0 1 3.5558931863983996
          20 0 1 2.0532678214541452E-004
          21 0 1 1.1119143253484136
          22 0 1 2.7597999527272327E-005
          23 0 1 0.14945254927143536
Time in seconds: 1

Not really sure what this inconsistency means and how to fix it. Let me know if more information is needed.

Kind regards,

Pim

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
marco zaro Edit question
Solved by:
marco zaro
Solved:
2020-10-14
Last query:
2020-10-14
Last reply:
2020-10-14
marco zaro (marco-zaro) said : #1

Dear Pim,
we are aware of this bug, we are fixing it in a development branch. As soon as possible we will let you know how to correct the code behaviour.

Cheers,

Marco

Pim Verschuuren (pverschu) said : #2

Dear Marco,
Thats great to hear. Hope to hear from you soon then.
Cheers
Pim

Pim Verschuuren (pverschu) said : #3

Dear Marco,
Just checking if there has been any progress on the matter. Any indication on when this might be fixed?
Kind regards,
Pim

marco zaro (marco-zaro) said : #4

Dear Pim,
apologies for not having updated you. You can try this branch
https://code.launchpad.net/~maddevelopers/mg5amcnlo/3.0.3-betternlops
which includes the new sintax from 3.0.3-neworders, and fixes the behaviour at NLOPS.

Just a question, how do you shower the events? with the built-in analyses/tools or within a separate framework?

Let me know

Best wishes,

Marco

Pim Verschuuren (pverschu) said : #5

Hi Marco,
Thanks! I will try the new branch then. I will start with the built-in tools and then look later on to other shower frameworks. Will this branch be merged with an official release anytime soon?
Cheers,
Pim

marco zaro (marco-zaro) said : #6

Hi Pim,
the plans are to merge this branch in the next version (3.0.4).
If you start with the built-in tools (and I suppose you are going to use pythia8), there are few extra changes needed (I did not implement them yet, because I would like to have them implemented in a more flexible way):
Basically, you need to change the length of some string from 50 to 80. Below are the changes to be done.

Best,

Marco

=======

Template/MCatNLO/include/LHEFRead.h

< void lhef_read_wgtsinfo_(int &cwgtinfo_nn, char (cwgtinfo_weights_info[1024][50])) {
---
> void lhef_read_wgtsinfo_(int &cwgtinfo_nn, char (cwgtinfo_weights_info[1024][80])) {
 include/LHEFRead.h:
< void lhef_read_wgtsinfo_(int &cwgtinfo_nn, char (cwgtinfo_weights_info[1024][50])) {
—--
> void lhef_read_wgtsinfo_(int &cwgtinfo_nn, char (cwgtinfo_weights_info[1024][80])) {

< sprintf(cwgtinfo_weights_info[0], "%50s","central value");
--—
> sprintf(cwgtinfo_weights_info[0], “%80s","central value");

<
sprintf(cwgtinfo_weights_info[cwgtinfo_nn],"%50s",sRWGT.c_str());
--—
sprintf(cwgtinfo_weights_info[cwgtinfo_nn],”%80s",sRWGT.c_str());

- If you use the HwU analyses we ship within MG5_aMC:
diff ../Template/NLO/FixedOrderAnalysis/HwU.f FixedOrderAnalysis/
22c22
< character(len=50), allocatable :: wgts_info(:)
---
> character(len=80), allocatable :: wgts_info(:)

 diff ../Template/NLO/MCatNLO/PY8Analyzer/py8an_HwU_rates.f MCatNLO/PY8Analyzer/ ### this has to be done in the actual analysis you use
44c44
< character*50 weights_info(max_weight_shower)
---
> character*80 weights_info(max_weight_shower)

diff -r ../Template/NLO/MCatNLO/srcPythia8/Pythia8.cc <http://pythia8.cc/> MCatNLO/srcPythia8/Pythia8.cc <http://pythia8.cc/>
24c24
< void pyabeg_(int&,char(*)[50]);
---
> void pyabeg_(int&,char(*)[80]);
33c33
< char cwgtinfo_weights_info[1024][50];
---
> char cwgtinfo_weights_info[1024][80];
diff -r ../Template/NLO/MCatNLO/srcPythia8/Pythia82.cc <http://pythia82.cc/> MCatNLO/srcPythia8/Pythia82.cc <http://pythia82.cc/>
25c25
< void pyabeg_(int&,char(*)[50]);
---
> void pyabeg_(int&,char(*)[80]);
34c34
< char cwgtinfo_weights_info[1024][50];
---
> char cwgtinfo_weights_info[1024][80];

> On 8 Oct 2020, at 10:55, Pim Verschuuren <email address hidden> wrote:
>
> Question #692808 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/692808
>
> Pim Verschuuren posted a new comment:
> Hi Marco,
> Thanks! I will try the new branch then. I will start with the built-in tools and then look later on to other shower frameworks. Will this branch be merged with an official release anytime soon?
> Cheers,
> Pim
>
> --
> You received this question notification because you are assigned to this
> question.

Pim Verschuuren (pverschu) said : #7

Hi Marco,
The NLO+PS generation was working but with the latest commit 979 to branch betternlops it seems to break down again in the following way:

MG5_aMC>import model SMEFTatNLO-NLO
INFO: Restrict model SMEFTatNLO with file ../3.0.3-betternlops/models/SMEFTatNLO/restrict_NLO.dat .
INFO: Change particles name to pass to MG5 convention
Pass the definition of 'j' and 'p' to 5 flavour scheme.
Kept definitions of multiparticles l- / vl / l+ / vl~ unchanged
Defined multiparticle all = g ghg ghg~ u c d s b u~ c~ d~ s~ b~ a gha gha~ ve vm vt e- mu- ta- ve~ vm~ vt~ e+ mu+ ta+ t t~ z w+ ghz ghwp ghwm h w- ghz~ ghwp~ ghwm~
MG5_aMC>generate p p > t t~ QCD=2 QED=0 NP=2 [QCD]
Command "generate p p > t t~ QCD=2 QED=0 NP=2 [QCD]" interrupted with error:
TypeError : cannot concatenate 'str' and 'list' objects
Please report this bug on https://bugs.launchpad.net/mg5amcnlo
More information is found in 'MG5_debug'.
Please attach this file to your report.

The problem seems to be the [QCD] argument which when removed removes the error message.

Cheers
Pim

marco zaro (marco-zaro) said : #8

Hi Pim,
I cannot reproduce your error, I can generate the process you tried and
output it.
Note that changes in rev 979 do not affect code generation, only the nlo+PS
analyses..
In case, please include the content of MG5_debug

Cheers,

Marco

On Tue, 13 Oct 2020, 18:45 Pim Verschuuren, <
<email address hidden>> wrote:

> Question #692808 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/692808
>
> Pim Verschuuren posted a new comment:
> Hi Marco,
> The NLO+PS generation was working but with the latest commit 979 to branch
> betternlops it seems to break down again in the following way:
>
> MG5_aMC>import model SMEFTatNLO-NLO
> INFO: Restrict model SMEFTatNLO with file
> ../3.0.3-betternlops/models/SMEFTatNLO/restrict_NLO.dat .
> INFO: Change particles name to pass to MG5 convention
> Pass the definition of 'j' and 'p' to 5 flavour scheme.
> Kept definitions of multiparticles l- / vl / l+ / vl~ unchanged
> Defined multiparticle all = g ghg ghg~ u c d s b u~ c~ d~ s~ b~ a gha gha~
> ve vm vt e- mu- ta- ve~ vm~ vt~ e+ mu+ ta+ t t~ z w+ ghz ghwp ghwm h w-
> ghz~ ghwp~ ghwm~
> MG5_aMC>generate p p > t t~ QCD=2 QED=0 NP=2 [QCD]
> Command "generate p p > t t~ QCD=2 QED=0 NP=2 [QCD]" interrupted with
> error:
> TypeError : cannot concatenate 'str' and 'list' objects
> Please report this bug on https://bugs.launchpad.net/mg5amcnlo
> More information is found in 'MG5_debug'.
> Please attach this file to your report.
>
>
> The problem seems to be the [QCD] argument which when removed removes the
> error message.
>
> Cheers
> Pim
>
> --
> You received this question notification because you are assigned to this
> question.
>

Pim Verschuuren (pverschu) said : #9

Hi Marco,
Strangely enough it seems to occur when I run the above on Scientific Linux CERN SLC release 6.10. When I run on CentOs Linux 7 it works fine.
Cheers,
Pim

marco zaro (marco-zaro) said : #10

Hi Pim,
what are the python versions in the two OSs?
Thanks,

Marco

> On 13 Oct 2020, at 21:20, Pim Verschuuren <email address hidden> wrote:
>
> Question #692808 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/692808
>
> Pim Verschuuren posted a new comment:
> Hi Marco,
> Strangely enough it seems to occur when I run the above on Scientific Linux CERN SLC release 6.10. When I run on CentOs Linux 7 it works fine.
> Cheers,
> Pim
>
> --
> You received this question notification because you are assigned to this
> question.

Pim Verschuuren (pverschu) said : #11

Hi Marco,
Thanks for the hint. I changed from Python 2.7.4 GCC 48 to 2.7.10 GCC 62 and the code runs now.
Cheers,
Pim

Best marco zaro (marco-zaro) said : #12

Perfect.
Please confirm that the issue is solved, once you verify the results.

Best wishes,

Marco

> On 14 Oct 2020, at 11:30, Pim Verschuuren <email address hidden> wrote:
>
> Question #692808 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/692808
>
> Pim Verschuuren posted a new comment:
> Hi Marco,
> Thanks for the hint. I changed from Python 2.7.4 GCC 48 to 2.7.10 GCC 62 and the code runs now.
> Cheers,
> Pim
>
> --
> You received this question notification because you are assigned to this
> question.

Pim Verschuuren (pverschu) said : #13

Thanks marco zaro, that solved my question.