nnpdf / pinecards Goto Github PK

View Code? Open in Web Editor NEW

3.0 4.0 1.0 171.03 MB

Runcards needed to generate PineAPPL grids for NNPDF processes

Shell 4.46% Fortran 95.54%

pinecards's Introduction

Runcards for NNPDF

This repository stores all pinecards needed to generate PineAPPL grids for the processes included in NNPDF.

pinecards's People

Contributors

Stargazers

Watchers

Forkers

trellixvulnteam

pinecards's Issues

Enable complex-mass scheme for top-pair production?

Set scales and parameters of all processes to consistent choices

The following experiments have inconsistent/unfavourable scale choices:

ATLASZHIGHMASS49FB has a static scale set to the Z boson mass, whereas it should be set to a dynamic value, similar to CMSDY2D11, for instance
CMSTTBARTOT5TEV has a static scale set to the top boson mass, which is different from the choice for the 7, 8, and 13 TeV total cross sections (see also #80); it very likely has the top mass set set to 172.5 GeV

Store Monte Carlo runcard in grids

We should add the file DATASET/Events/run_01/run_01_tag_1_banner.txt to the metadata collected in the grids.

Reference processes and applgrids

We need cross-checked process runcards and reference applgrids for pure QCD.
@enocera , @marcozaro could you please take care of this?

bug in applgrids?

Umm, there is something strange with the single bins applgrids, e.g. I am getting this error:

carrazza@dom:/media/apfelcomb (ew)$ ./apfel_comb app 1500 199
 EScaleVar: 1
Error in <TCling::RegisterModule>: cannot find dictionary module TFileStringDict_rdict.pcm
Error in <TCling::RegisterModule>: cannot find dictionary module TFileVectorDict_rdict.pcm
appl::grid() reading grid from file ../applgrids/CMSTTBARTOT/TOPDIFF7TEVTOT.root        version 1.4.56(transformed to 1.4.70)
Error in <TVectorT<double>::operator()>: Request index(12) outside vector range of 0 - 12
terminate called after throwing an instance of 'std::length_error'
  what():  vector::reserv

Check metadata for mistakes

check for missing abs operators (rapidity, pseudorapidity),
add empty xn_unit= for rapidity, pseudorapidity, ...

Two-dimensional distributions are not supported by mg5_aMC, and therefore we linearise the 2D distributions to 1D, but after the grid generation we have to undo the linearisation using pineappl remap ... in a postrun.sh script to properly normalise them. The following analyses are affected by this:

ATLAS_DY3D_8TEV, fixed in commit 739ab7f, but needs implementation in PineAPPL
CMSDY2D11, fixed in commit 8e38d71
~~#43: CMSDY2D12~~
#44: ATLAS_DY2D_8TEV
#67: CMS_2JET_7TEV, fixed in commit be0afb7
#69: ATLAS_2JET_7TEV_R06, fixed in commits be0afb7 and 883f58e
#70: CMS_2JET_3D_8TEV, fixed in commit be0afb7, but needs implementation in PineAPPL
#71: ATLAS_1JET_8TEV_R06, fixed in commit 0cca8ba

Validate CMS_ZPT_13TEV

merge bins bug?

After commit 2422e8a, the ATLASZHIGHMASS merge_bin root file gives me:

Bin 0: 241477
Bin 1: 1.80574e+06
Bin 2: 822028
Bin 3: 624694
Bin 4: 294208
Bin 5: 227265
Bin 6: 527880
Bin 7: 66649.7
Bin 8: 25905.6
Bin 9: 10195.7
Bin 10: 2032.64
Bin 11: 346.109
Bin 12: 66.2553

which is completely different from an individual query, e.g. for amcblast_obs_2.root I get:

Bin 0: 0.0566167

applgrid_ew folder

Can we remove the applgrid_ew folder?

Rename datasets consistently?

The naming scheme of the datasets is inconsistent; some use underscores, some don't. We should also remove NORM from all names as the PineAPPL grids aren't normalised.

applgridphoton

I've realised that I need to reinstall applgridphoton given the recent changes. From a clean pull of the github repo, I obtain the following error:
cd applgridphoton
autoreconf -i
./configure --prefix=<path_to_my_prefix>
./configure: line 15090: syntax error near unexpected token 11,
./configure: line 15090: 'AX_CXX_COMPILE_STDCXX(11, noext, mandatory)'
I thought that the ax_cxx_compile_stdcxx_11.m4 file was added exactly to avoid this?
@scarrazza, @cschwan any hints?

Fix PDG id = 0 to 21

In Madgraph5_aMC@NLO the luminosity function is constructed using a 0 for the gluon, but this isn't a proper PDG id. As far as LHAPDF concerns, this isn't a problem, but other programs (EKO, ...) are stricter.

Add y_unit metadata consistently

We should the units for the differential cross sections with y_unit, for instance to something like pb/GeV or pb/GeV^2.

Set missing parameters

We should check all runcards for missing parameters that should be set to values defined by us, not the default ones of mg5_aMC.

Here's an incomplete list:

maxjetflavor. This parameter is set to 4 by default, although for consistency with the model mg5_aMC it is set to 5. We should explicitly set it to 5 to avoid a warning.
mh. The Higgs mass.
gf. The fermi coupling, which determines alpha.
wt, wz, ww, wh: decay widths for all particles.

lz4 grid not created

Here's something that I don't understand. I'm running the LHCb grids on the stoomboot cluster. Everything seems to go as expected, except for the fact that the .lz4 grid is not produced.

Let's look for instance at LHCB_DY_13TEV_DIELECTRON. I have the launch.log, output.log, pineappl.convolute, pineappl.orders, pineappl.pdf_uncertainty, result.log and time.log files generated (see below). Any idea as of why the .lz4 grid is not generated? Incidentally, how it is generated from the .pineappl files in Events/run01?

output.log
launch.log
results.log
pineappl.pdf_uncertainty.txt
pineappl.orders.txt
pineappl.convolute.txt
time.log

Figure out a sane way to hook into mg5_aMC

The functions that are defined in amcfast are actually never directly called from mg5_aMC. This scenario causes the linker, which is called from mg5_aMC when generating code, to discard this library. To avoid this, the config script of amcfast has the following lines:
https://github.com/scarrazza/applgridew/blob/c86737322a25b0a76b28cca58faa48ef994c0e24/amcblast/amcfast-config.in#L37-L38

This instructs the linker to undefines the symbol corresponding to the function setup_amcfast_. For example, my output of amcfast-config --ldflags is:

-u setup_amcfast_ -L/home/cschwan/prefix/lib64 -lamcfast

The undefined symbol (or any undefined symbol, actually) ultimately causes the linker to load this library, so that the following global variable is initialized,
https://github.com/scarrazza/applgridew/blob/c86737322a25b0a76b28cca58faa48ef994c0e24/amcblast/amcfast_interface.cxx#L57
which is irrelevant but which calls the function setup_amcfast_, which in turn sets the function pointers, which finally are invoked by mg5_aMC.

We should figure out a saner to load the interfacing code, for example by simply writing a source file which implements the functions and placing it in mg5_aMC.

Impact of the EW corrections to PDF fits?

Now that we have EW corrections I think we should start thinking about an application using the new grids. I was thinking of a first step, where we generate EW corrections/photon-induced contributions for a simple process an study its impact and thereby also verifying that we did makes sense. We have to keep in mind that the new applgridphoton produces a slightly different file format (if thats a problem I can work on it). Can we maybe make a more concrete plain?

Higgs

Following our discussion yesterday, I am trying to build pineappl grids for Higgs cross-sections.
I am observing 2 problems, the first unrelated and the other related to madgraph+pineappl:

When using MG5 2.8.2 (no pineappl), the VBF channel generated with:

import model HC_NLO_X0_UFO
generate p p > x0 j j $$ w+ w- z / a [QCD]

when running the NLO fixed order simulation raises:

...
INFO:     P0_ucx_x0ucx_no_a done. 
INFO:     P0_ddx_x0ddx_no_a done. 
INFO: Checking test output: 
INFO: P0_uu_x0uu_no_a 
INFO:  Result for test_ME: 
INFO:    Passed. 
INFO:  Result for check_poles: 
Error detected in "launch auto "
write debug file /home/carrazza/higgs/MG5_aMC_v2_8_2/vbf/run_03_tag_1_debug.log 
If you need help with this issue please contact us on https://answers.launchpad.net/mg5amcnlo
str : Poles do not cancel, run cannot continue

When using the MG5 version for pineappl (3.0.4.beta from bzr, following the README.md in this repo), for:

import model HC_NLO_X0_UFO-heft
generate p p > x0 / t [QCD]

I get:

MG5_aMC>generate p p > x0 / t [QCD]
INFO: Checking for minimal orders which gives processes. 
INFO: Please specify coupling orders to bypass this step. 
INFO: Trying coupling order WEIGHTED<=2: WEIGTHED IS 2*QNP+2*QED+QCD 
Setting the born orders automatically in the process definition to QNP=0 QED<=1 QCD=0.
If this is not what you need, please regenerate with the correct orders.
------------------------------------------------------------------------
 No Born diagrams found. Now switching to the loop-induced mode.        
 Please cite ref. 'arXiv:1507.00020' when using results from this mode. 
------------------------------------------------------------------------
Command "generate p p > x0 / t [QCD]" interrupted with error:
Exception : 
Please report this bug on https://bugs.launchpad.net/mg5amcnlo
More information is found in 'MG5_debug'.
Please attach this file to your report.

@marcozaro, maybe the generation command is incomplete for pineappl?

Add y_label and y_label_tex metadata

Validate CMS_DY_13TEV_28FB

Fix problem with photon-initiated corrections

The NLO corrections of a a > e+ e- [QED] in the APPLgrid are wrong:

-----------------------------------------------------------
   APPLgrid       mg5_aMC   mg5_aMC unc.  sigmas  per cent
-----------------------------------------------------------
 9.747559e-02  3.135564e-02 7.118766e-04   92.88  210.87%
 2.451823e-02  3.116021e-02 1.065792e-03    6.23   21.32%
 1.671845e-03  1.911880e-02 2.963357e-04   58.88   91.26%
-1.262473e-03  1.267250e-02 2.003674e-04   69.55  109.96%
 2.033341e-03  8.953376e-03 1.950746e-04   35.47   77.29%
 7.276895e-04  5.951386e-03 1.573784e-04   33.19   87.77%
-2.628181e-04  4.316486e-03 1.469767e-04   31.16  106.09%
 1.541911e-03  6.594314e-03 1.381404e-04   36.57   76.62%
 1.165692e-03  4.834593e-03 1.264922e-04   29.00   75.89%
 4.076927e-04  1.627656e-03 6.030443e-05   20.23   74.95%
-4.398920e-06  8.556226e-04 4.285031e-05   20.07  100.51%
 9.693089e-05  2.235930e-04 1.012199e-05   12.51   56.65%
 7.533169e-06  4.756511e-05 3.570617e-06   11.21   84.16%

LO is fine in this case.

Get rid of memory leak in amcfast/amcblast

That's the last new still present in amcblast:
https://github.com/scarrazza/applgridew/blob/c86737322a25b0a76b28cca58faa48ef994c0e24/amcblast/amcfast_interface.cxx#L236

It seems that this line is absolutely neccessary, meaning the object must be contructed and never destructed, not even when the library is unloaded. If the destructor is called, mg5_aMC eventually crashes. That doesn't make sense to me.

Address remaining issues of the paper

I think it best if we collect the remaining issues of the paper in this issue here, in one place. Right now they are spread over several emails, which is probably not ideal.

Get rid of amcfast

I just had short chat with @scarrazza and I am summarizing it here to let you know what I am doing on my end.

As the title suggests, the plan is to get rid of amcfast. Instead of amcfast, we would like to have a small library providing an interface (I) between the Monte Carlo (MC) and what basically amounts to the library (which in principle can be the same as the interface) which stores the partonic grid (PG), for example applgrids or fktables:

MC -> I -> PG

The interface should be, ideally after only a few modifications to the MC, easily callable from any MC generator, first for QCD only and later on for any type, be it LO/NLO/NNLO EW, QCD, and mixed predictions/corrections.

The test case for MC generators will first be MG5_aMC, since we already have something with amcfast which is basically working, and later on my own MC (not yet published, but public nevertheless). I think having two different MCs enables us to have two point of views which will certainly be helpful when designing the interface.

Some analyses don't compile

Some analyses don't compile on my machine. It seems that at least all analyses with more than one bin fail. The interesting part for the ATLASTOPDIFF8TEVTRAPNORM analysis with the first bin is:

INFO: Using LHAPDF v6.2.1 interface for PDFs
INFO: Compiling source...
INFO:           ...done, continuing with P* directories
INFO: Compiling directories...
INFO: Compiling on 1 cores
INFO:  Compiling P0_gg_ttx...
STOP 1
INFO:     P0_gg_ttx done.
INFO:  Compiling P0_ga_ttx...
STOP 1
INFO:     P0_ga_ttx done.
INFO:  Compiling P0_uux_ttx...
STOP 1
INFO:     P0_uux_ttx done.
INFO:  Compiling P0_ddx_ttx...
STOP 1
INFO:     P0_ddx_ttx done.
INFO:  Compiling P0_bbx_ttx...
STOP 1
INFO:     P0_bbx_ttx done.
INFO:  Compiling P0_uxu_ttx...
STOP 1
INFO:     P0_uxu_ttx done.
INFO:  Compiling P0_dxd_ttx...
STOP 1
INFO:     P0_dxd_ttx done.
INFO:  Compiling P0_bxb_ttx...
STOP 1
INFO:     P0_bxb_ttx done.
INFO:  Compiling P0_ag_ttx...
STOP 1
INFO:     P0_ag_ttx done.
INFO: Checking test output:
INFO: P0_gg_ttx
INFO:  Result for test_ME:
INFO:    Passed.
DEBUG:
     Soft test  1 PASSED. Fraction of failures: 0.02
Collinear test  1 PASSED. Fraction of failures: 0.00
     Soft test  2 PASSED. Fraction of failures: 0.02
Collinear test  2 PASSED. Fraction of failures: 0.02
     Soft test  3 PASSED. Fraction of failures: 0.00
     Soft test  4 PASSED. Fraction of failures: 0.00
     Soft test  5 PASSED. Fraction of failures: 0.00
Collinear test  5 PASSED. Fraction of failures: 0.01
     Soft test  6 PASSED. Fraction of failures: 0.00
Collinear test  6 PASSED. Fraction of failures: 0.01
     Soft test  7 PASSED. Fraction of failures: 0.00
Collinear test  7 PASSED. Fraction of failures: 0.00
     Soft test  8 PASSED. Fraction of failures: 0.00
Collinear test  8 PASSED. Fraction of failures: 0.00
     Soft test  9 PASSED. Fraction of failures: 0.00
Collinear test  9 PASSED. Fraction of failures: 0.00
     Soft test 10 PASSED. Fraction of failures: 0.00
Collinear test 10 PASSED. Fraction of failures: 0.00
     Soft test 11 PASSED. Fraction of failures: 0.00
Collinear test 11 PASSED. Fraction of failures: 0.02
     Soft test 12 PASSED. Fraction of failures: 0.00
Collinear test 12 PASSED. Fraction of failures: 0.01
     Soft test 13 PASSED. Fraction of failures: 0.00
Collinear test 13 PASSED. Fraction of failures: 0.00
     Soft test 14 PASSED. Fraction of failures: 0.00
Collinear test 14 PASSED. Fraction of failures: 0.01
     Soft test 15 PASSED. Fraction of failures: 0.00
Collinear test 15 PASSED. Fraction of failures: 0.02
     Soft test 16 PASSED. Fraction of failures: 0.00
Collinear test 16 PASSED. Fraction of failures: 0.00
     Soft test 17 PASSED. Fraction of failures: 0.00
     Soft test 18 PASSED. Fraction of failures: 0.00
INFO:  Result for check_poles:
INFO:    Poles successfully cancel for 20 points over 20 (tolerance=1.0e-05)
INFO: P0_ga_ttx
INFO:  Result for test_ME:
INFO:    Passed.
DEBUG:
     Soft test  1 PASSED. Fraction of failures: 0.03
Collinear test  1 PASSED. Fraction of failures: 0.02
     Soft test  2 PASSED. Fraction of failures: 0.00
     Soft test  3 PASSED. Fraction of failures: 0.00
     Soft test  4 PASSED. Fraction of failures: 0.00
Collinear test  4 PASSED. Fraction of failures: 0.00
     Soft test  5 PASSED. Fraction of failures: 0.00
Collinear test  5 PASSED. Fraction of failures: 0.00
     Soft test  6 PASSED. Fraction of failures: 0.00
Collinear test  6 PASSED. Fraction of failures: 0.00
     Soft test  7 PASSED. Fraction of failures: 0.00
Collinear test  7 PASSED. Fraction of failures: 0.00
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.
DEBUG:
     Soft test  1 PASSED. Fraction of failures: 0.00
Collinear test  1 PASSED. Fraction of failures: 0.00
     Soft test  2 PASSED. Fraction of failures: 0.00
Collinear test  2 PASSED. Fraction of failures: 0.02
     Soft test  3 PASSED. Fraction of failures: 0.00
     Soft test  4 PASSED. Fraction of failures: 0.00
     Soft test  5 PASSED. Fraction of failures: 0.00
Collinear test  5 PASSED. Fraction of failures: 0.00
     Soft test  6 PASSED. Fraction of failures: 0.01
Collinear test  6 PASSED. Fraction of failures: 0.01
     Soft test  7 PASSED. Fraction of failures: 0.00
Collinear test  7 PASSED. Fraction of failures: 0.01
     Soft test  8 PASSED. Fraction of failures: 0.00
     Soft test  9 PASSED. Fraction of failures: 0.00
     Soft test 10 PASSED. Fraction of failures: 0.00
Collinear test 10 PASSED. Fraction of failures: 0.00
     Soft test 11 PASSED. Fraction of failures: 0.00
Collinear test 11 PASSED. Fraction of failures: 0.02
     Soft test 12 PASSED. Fraction of failures: 0.00
Collinear test 12 PASSED. Fraction of failures: 0.02
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.
DEBUG:
     Soft test  1 PASSED. Fraction of failures: 0.00
Collinear test  1 PASSED. Fraction of failures: 0.00
     Soft test  2 PASSED. Fraction of failures: 0.01
Collinear test  2 PASSED. Fraction of failures: 0.00
     Soft test  3 PASSED. Fraction of failures: 0.00
     Soft test  4 PASSED. Fraction of failures: 0.00
     Soft test  5 PASSED. Fraction of failures: 0.00
Collinear test  5 PASSED. Fraction of failures: 0.00
     Soft test  6 PASSED. Fraction of failures: 0.00
Collinear test  6 PASSED. Fraction of failures: 0.00
     Soft test  7 PASSED. Fraction of failures: 0.02
Collinear test  7 PASSED. Fraction of failures: 0.02
     Soft test  8 PASSED. Fraction of failures: 0.00
     Soft test  9 PASSED. Fraction of failures: 0.00
     Soft test 10 PASSED. Fraction of failures: 0.00
Collinear test 10 PASSED. Fraction of failures: 0.02
     Soft test 11 PASSED. Fraction of failures: 0.00
Collinear test 11 PASSED. Fraction of failures: 0.00
     Soft test 12 PASSED. Fraction of failures: 0.00
Collinear test 12 PASSED. Fraction of failures: 0.00
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.
DEBUG:
     Soft test  1 PASSED. Fraction of failures: 0.00
Collinear test  1 PASSED. Fraction of failures: 0.00
     Soft test  2 PASSED. Fraction of failures: 0.00
Collinear test  2 PASSED. Fraction of failures: 0.00
     Soft test  3 PASSED. Fraction of failures: 0.00
     Soft test  4 PASSED. Fraction of failures: 0.00
     Soft test  5 PASSED. Fraction of failures: 0.00
Collinear test  5 PASSED. Fraction of failures: 0.00
     Soft test  6 PASSED. Fraction of failures: 0.03
Collinear test  6 PASSED. Fraction of failures: 0.01
     Soft test  7 PASSED. Fraction of failures: 0.02
Collinear test  7 PASSED. Fraction of failures: 0.00
     Soft test  8 PASSED. Fraction of failures: 0.00
     Soft test  9 PASSED. Fraction of failures: 0.00
     Soft test 10 PASSED. Fraction of failures: 0.00
Collinear test 10 PASSED. Fraction of failures: 0.00
     Soft test 11 PASSED. Fraction of failures: 0.00
Collinear test 11 PASSED. Fraction of failures: 0.02
     Soft test 12 PASSED. Fraction of failures: 0.00
Collinear test 12 PASSED. Fraction of failures: 0.01
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.
DEBUG:
     Soft test  1 PASSED. Fraction of failures: 0.01
Collinear test  1 PASSED. Fraction of failures: 0.00
     Soft test  2 PASSED. Fraction of failures: 0.00
Collinear test  2 PASSED. Fraction of failures: 0.00
     Soft test  3 PASSED. Fraction of failures: 0.00
     Soft test  4 PASSED. Fraction of failures: 0.01
     Soft test  5 PASSED. Fraction of failures: 0.00
Collinear test  5 PASSED. Fraction of failures: 0.00
     Soft test  6 PASSED. Fraction of failures: 0.00
Collinear test  6 PASSED. Fraction of failures: 0.00
     Soft test  7 PASSED. Fraction of failures: 0.00
Collinear test  7 PASSED. Fraction of failures: 0.00
     Soft test  8 PASSED. Fraction of failures: 0.00
     Soft test  9 PASSED. Fraction of failures: 0.00
     Soft test 10 PASSED. Fraction of failures: 0.00
Collinear test 10 PASSED. Fraction of failures: 0.00
     Soft test 11 PASSED. Fraction of failures: 0.00
Collinear test 11 PASSED. Fraction of failures: 0.00
     Soft test 12 PASSED. Fraction of failures: 0.00
Collinear test 12 PASSED. Fraction of failures: 0.07
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.
DEBUG:
     Soft test  1 PASSED. Fraction of failures: 0.01
Collinear test  1 PASSED. Fraction of failures: 0.00
     Soft test  2 PASSED. Fraction of failures: 0.00
Collinear test  2 PASSED. Fraction of failures: 0.02
     Soft test  3 PASSED. Fraction of failures: 0.00
     Soft test  4 PASSED. Fraction of failures: 0.00
     Soft test  5 PASSED. Fraction of failures: 0.00
Collinear test  5 PASSED. Fraction of failures: 0.00
     Soft test  6 PASSED. Fraction of failures: 0.01
Collinear test  6 PASSED. Fraction of failures: 0.00
     Soft test  7 PASSED. Fraction of failures: 0.01
Collinear test  7 PASSED. Fraction of failures: 0.02
     Soft test  8 PASSED. Fraction of failures: 0.00
     Soft test  9 PASSED. Fraction of failures: 0.00
     Soft test 10 PASSED. Fraction of failures: 0.00
Collinear test 10 PASSED. Fraction of failures: 0.05
     Soft test 11 PASSED. Fraction of failures: 0.00
Collinear test 11 PASSED. Fraction of failures: 0.00
     Soft test 12 PASSED. Fraction of failures: 0.00
Collinear test 12 PASSED. Fraction of failures: 0.03
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.
DEBUG:
     Soft test  1 PASSED. Fraction of failures: 0.00
Collinear test  1 PASSED. Fraction of failures: 0.02
     Soft test  2 PASSED. Fraction of failures: 0.00
Collinear test  2 PASSED. Fraction of failures: 0.00
     Soft test  3 PASSED. Fraction of failures: 0.00
     Soft test  4 PASSED. Fraction of failures: 0.00
     Soft test  5 PASSED. Fraction of failures: 0.00
Collinear test  5 PASSED. Fraction of failures: 0.02
     Soft test  6 PASSED. Fraction of failures: 0.01
Collinear test  6 PASSED. Fraction of failures: 0.00
     Soft test  7 PASSED. Fraction of failures: 0.04
Collinear test  7 PASSED. Fraction of failures: 0.01
     Soft test  8 PASSED. Fraction of failures: 0.00
     Soft test  9 PASSED. Fraction of failures: 0.00
     Soft test 10 PASSED. Fraction of failures: 0.00
Collinear test 10 PASSED. Fraction of failures: 0.01
     Soft test 11 PASSED. Fraction of failures: 0.00
Collinear test 11 PASSED. Fraction of failures: 0.00
     Soft test 12 PASSED. Fraction of failures: 0.00
Collinear test 12 PASSED. Fraction of failures: 0.04
INFO:  Result for check_poles:
INFO:    Poles successfully cancel for 20 points over 20 (tolerance=1.0e-05)
INFO: P0_ag_ttx
INFO:  Result for test_ME:
INFO:    Passed.
DEBUG:
     Soft test  1 PASSED. Fraction of failures: 0.01
Collinear test  1 PASSED. Fraction of failures: 0.00
     Soft test  2 PASSED. Fraction of failures: 0.00
     Soft test  3 PASSED. Fraction of failures: 0.00
     Soft test  4 PASSED. Fraction of failures: 0.00
Collinear test  4 PASSED. Fraction of failures: 0.00
     Soft test  5 PASSED. Fraction of failures: 0.00
Collinear test  5 PASSED. Fraction of failures: 0.00
     Soft test  6 PASSED. Fraction of failures: 0.00
Collinear test  6 PASSED. Fraction of failures: 0.00
     Soft test  7 PASSED. Fraction of failures: 0.00
Collinear test  7 PASSED. Fraction of failures: 0.00
INFO:  Result for check_poles:
INFO:    Poles successfully cancel for 20 points over 20 (tolerance=1.0e-05)
INFO: Starting run
INFO: Using 1 cores
INFO: Doing fixed order NLO
Command "launch auto " interrupted with error:
ValueError : max() arg is an empty sequence
Please report this bug on https://bugs.launchpad.net/mg5amcnlo
More information is found in '/home/cschwan/projects/applgridew/ATLASTOPDIFF8TEVTRAPNORM_1/run_02_tag_1_debug.log'.
Please attach this file to your report.

(note the additional STOP 1 messages that usually don't appear). @marcozaro, @enocera : Do you have any idea how to debug the problem? The logfile doesn't contain any useful information.

Check impact of wrong binning for ATLAS_1JET_8TEV_R06

Commit f8da697 fixed a bug which caused some but not all events of the last bin for each slice to be filled into the first bin of the next slice. This should have a large impact on the bin which isn't filled and a negligible impact for the bin that is wrongly filled, see also the comparison in #71.

some comments on sect3

Hi,
here i put some commetns on sect3:

processes in which EW corrections are expected to be sizeable: well, not particularly per se (We are always speaking of few %s...) I would maybe write " have an important impact in terms of the prediction accuracy"
in the description of the processes, the sentence about the luminosities:
The process receives contributions from 12 (34) parton luminosities at NLO QCD (NLO QCD+EW), mainly from qq ̄ scattering,
mainly is ambiguos here: does it mean that most of them are qqbar, or that the dominant contribution in term of xsect is qqbar?

I would also write explicitly the coupling combinations (e.g for DY):
The cross section includes contributions at O(\alpha^2) at LO, and O(\alpha^3) and O(\alpha^2 \alphas) at NLO

Check rapidity vs. pseudorapitdy

We shouldn't use rapidity when we actually have to use pseudorapidity. This makes a difference when photons are recombined with leptons, which happens when EW corrections are enabled.

Validate top-pair production grids

Right now the top-pair production total cross sections don't agree with each other (all with NNPDF31_nlo_as_0118, scale set to the top mass):

7 TeV: 1.5627757e2 (PineAPPL) vs. 1.661941e+02 (APPLgrid) - difference is -5.97 %
8 TeV: 2.2420122e2 (PineAPPL) vs. 2.380239e+02 (APPLgrid) - difference is -5.81 %
13 TeV: 7.4323410e2 (PineAPPL) vs. 7.844937e+02 (APPLgrid) - difference is 5.26 %

Problem with high mass DY

I've defined the relevant files for ATLASZHIGMASS39FB. However, when I try to produce the corresponding applgrids (./run.sh ATLASZHIGMASS39FB 1), I obtain the attached error. It seems to me that the output, launch and analysis cards are correct, at least as far as I can tell from a benchmark of a [QCD] run against the NNPDF FK tables. But please check that I didn't do something stupid.
error.txt

Fix scale variations in apfelcomb

Scale variations are very likely to be broken in our 'fixed'-apfelcomb version from the ew branch.

Do not use a non-zero photon PDF for NLO QCD comparisons

Since the APPLgrids usually don't contain photon-photon contributions but the PineAPPL do contain them, for Drell-Yan we shouldn't use a PDF set with a non-zero photon. This is probably OK for Drell-Yan Z around the Z peak, were the relative size of the photon-photon contributions are negligable, but especially for higher invariant masses it makes a noticable difference.

TODOs

Here the TODO list:

Warm-up:

Download 3.0.1 and amcfast 2.0.0
Runcard for the test mll, DY
Generate the APPLgrid

Fluid work:

Reference NLO QCD, NLO QCD + EW + PI
aMCfast and update the methods: fill and reco, and check the output with the ref.

Total mess:

applgridphoton and update the amc_convolute method.
check the results...

Optional:

merging the changes in applgrid 1.56 (NNLO QCD) > DYNNLO + EW

Split jet process runcards to maximise statistics in all bins

Split up jet runcards for

ATLAS_1JET_8TEV_R06,
ATLAS_2JET_7TEV_R06 (commit e7deef0, probably needs the pole cancellation check disabled),
CMS_2JET_7TEV (commits 2bd1c06 and 961490a) and
CMS_2JET_8TEV_3D.

The last one has low priority, since I won't be included in the NNPDF4.0 fit. See also #105, which is used instead.

Each of them has to checked again, see also #71, #69, #67 and #70.

Poles do not cancel

@cschwan , @marcozaro I am getting the following error for ATLASTOPDIFF8TEVRAPNORM_1:

INFO: Checking test output:
INFO: P0_gg_ttx
INFO: Result for test_ME:
INFO: Passed.
DEBUG:
Soft test 1 PASSED. Fraction of failures: 0.03
Collinear test 1 PASSED. Fraction of failures: 0.00
Soft test 2 PASSED. Fraction of failures: 0.00
Collinear test 2 PASSED. Fraction of failures: 0.00
Soft test 3 PASSED. Fraction of failures: 0.00
Soft test 4 PASSED. Fraction of failures: 0.00
Soft test 5 PASSED. Fraction of failures: 0.00
Collinear test 5 PASSED. Fraction of failures: 0.03
Soft test 6 PASSED. Fraction of failures: 0.00
Collinear test 6 PASSED. Fraction of failures: 0.02
Soft test 7 PASSED. Fraction of failures: 0.00
Collinear test 7 PASSED. Fraction of failures: 0.02
Soft test 8 PASSED. Fraction of failures: 0.00
Collinear test 8 PASSED. Fraction of failures: 0.03
Soft test 9 PASSED. Fraction of failures: 0.00
Collinear test 9 PASSED. Fraction of failures: 0.00
Soft test 10 PASSED. Fraction of failures: 0.00
Collinear test 10 PASSED. Fraction of failures: 0.00
Soft test 11 PASSED. Fraction of failures: 0.00
Collinear test 11 PASSED. Fraction of failures: 0.00
Soft test 12 PASSED. Fraction of failures: 0.00
Collinear test 12 PASSED. Fraction of failures: 0.01
Soft test 13 PASSED. Fraction of failures: 0.00
Collinear test 13 PASSED. Fraction of failures: 0.00
Soft test 14 PASSED. Fraction of failures: 0.00
Collinear test 14 PASSED. Fraction of failures: 0.01
Soft test 15 PASSED. Fraction of failures: 0.00
Collinear test 15 PASSED. Fraction of failures: 0.01
Soft test 16 PASSED. Fraction of failures: 0.00
Collinear test 16 PASSED. Fraction of failures: 0.00
Soft test 17 PASSED. Fraction of failures: 0.01
Soft test 18 PASSED. Fraction of failures: 0.00
INFO: Result for check_poles:
Error detected in "launch auto "
write debug file /media/applgridew/ATLASTOPDIFF8TEVTRAPNORM_1/run_02_tag_1_debug.log
If you need help with this issue, please, contact us on https://answers.launchpad.net/mg5amcnlo
aMCatNLOError : Poles do not cancel, run cannot continue
quit
INFO:

Fix problems with distributions having more than one bin

I currently cannot generate applgrids with more than one kinematic bin. For this reason, for the time being, one has to define a different analysis for of them. This is of course not optimal, as it introduces a lot of unnecessary redundance. Can we improve on this?

Implement CMS single-jet inclusive cross section for 8 TeV

Latex compilation problem (missing figure)

! LaTeX Error: File `figures/grid-insertion' not found.

See the LaTeX manual or LaTeX Companion for explanation.
Type H for immediate help.
...

l.253 ...\includegraphics{figures/grid-insertion}

Understanding iamp

@marcozaro could you please tell us which values iamp takes e.g. in pp->jj (interesting setup where there are more LO and NLO contributions)?

Make run script print bin boundaries

For plotting we need the bin boundaries.

Verifying/Testing generated root files

It would be nice if were able regression test the root files generated by amcfast. The problem is that, when I run mg5_aMC+aMCfast+APPLgrid once, store the root file, and run the same commands again, I get a different root file. One difference is due to the automatic seeding of mg5_aMC, but this should be fixed by set run_card iseed 1 (it isn't, I still get a different file).

Intuitively, a run with the same seed should give the same events and therefore the same file. Does anyone know why this is not the case?

Check impact of potentially wrong ptj1 cut for ATLAS_2JET_7TEV_R06

~~After merging into master the agreement is gone:~~ I used a different PDF set for the APPLgrid results. After fixing this problem I get:

 -1.02222430    0.23783085    0.42707336    0.08416723   -0.35890027
  0.53332775   -0.65175490    0.37084229    1.34268569    2.56243214
 -1.15846103   -0.37074841    6.52606247    7.92922876    3.69557847
  8.83242374    5.29342734   -0.70616165   15.38417262    9.88136292
 19.30995302   -0.48186943   -0.60185192    1.15701443   -1.05622221
  0.30655539    1.35661477    0.79734111    0.12810878    1.16299842
  0.47932707    0.39998455    4.34439270    1.30922030    2.58526248
  4.41567310    7.18822509   13.96898531    0.68302036   10.16927331
 13.85612764   11.73701319   -1.33815810    1.76292978   -1.34275879
  1.73296662    0.72089029   -3.36712141    2.73036803    1.23907121
  0.45803406    1.47322575    1.83255856    2.45793111    0.80202801
 16.03063018    4.44055573   -2.72964868   25.40312861    5.43674058
 13.30669472   -6.99308465   -0.32325814    5.81383437    8.53492793
  6.65636401   -3.79326103    6.41711561    1.61106071   -4.87539892
 22.15059217   -1.93323024   38.50962797    8.58001113  -28.05461531
 47.73147603   -7.30697707   37.84938836   10.82953882    9.77293391
-16.17762630   -4.82867701  -11.82768266   26.06950929   34.65325647
 59.89078615  140.35227357 -172.62612498  -53.56737926  182.66990903

This is worse than the previous comparison, but the differences are consistent with the MC uncertainties for the PineAPPL result. Another problem is that the PineAPPL results are missing a factor 2.

Validate ATLAS_DY_8TEV_3D

Compiling afpelcomb against applgridphoton

@scarrazza Suppose that we are able to generate an applgrid with EW corrections. Then in order to produce the FK table, we need to compile apfelcomb against the modified version of applgridphoton. If I do so, I obtain the error pasted below. Before I look into this, are you able to reproduce this error?

noceral@tplxlap211:~/Documents/nnpdfgit/apfelcomb$ make
make[1]: Entering directory /home/noceral/Documents/nnpdfgit/apfelcomb/src/core' g++ -Wall -O3 -D RESULTS_PATH="./results/" -D APPL_PATH="../applgrids/" -D DB_PATH="./db/" -I/home/noceral/Programs/LHAPDF6/include -I/home/noceral/Programs/APPLGRID/include -pthread -std=c++11 -m64 -I/home/noceral/Programs/ROOT/include -I/home/noceral/Programs/APFEL/include -std=c++14 -I/usr/include -I .. -c -o fk_appl.o fk_appl.cc fk_appl.cc: In function ‘double APP::compute_wgt_norm(const appl::grid*, const int&, const double&, const double&, const double&, const double&)’: fk_appl.cc:177:26: error: ‘const class appl::grid’ has no member named ‘leadingOrder’ const double LO = g->leadingOrder(); ^ fk_appl.cc: In member function ‘virtual void APP::SubGrid::Combine(const QCD::qcd_param&, NNPDF::FKGenerator*) const’: fk_appl.cc:283:77: error: ‘const class appl::grid’ has no member named ‘leadingOrder’ const double renscale = (as/(2.0*M_PI))*2.0*M_PI*QCD::beta0()*g->leadingOrder()*log ^ make[1]: *** [fk_appl.o] Error 1 make[1]: Leaving directory /home/noceral/Documents/nnpdfgit/apfelcomb/src/core'
make: *** [core] Error 2

Investigate possible LHCb issue

For all LHCb datasets we assume that we have a detector with 4pi coverage, and correct by dividing by a factor of two. However, I think this slightly incorrect and possibly overcounts lepton pairs, which are in both hemispheres (one lepton has negative and the other one positive rapidity). I think what we have to additionally do is to require both leptons have the same sign in rapidity. In practice this is probably not a problem, however, because of the chosen cuts: if

2 < |y lepton| < 4.5 and
|y Z| > 2

then the worst that can happen is y l- = -2 and y l+ = 4.5 so that y Z = 0.5 * (-2 + 4.5) = 1.25 which will be cut away.

-1.346218521   -0.587107620   -0.703440259   -1.053975521   -3.379005245
-1.661724065   -0.311729776   -2.359492089   -5.519613276   -3.907902083
-1.877671720   -4.735872295   -2.497950713   -2.098728547   -2.056511911
-2.460630695   -3.361925531    1.458201504   -2.432899397   -2.167920361
-3.924735796   -4.645240877   15.580261663   -0.216431151   -3.073440962
-2.379884404   -2.101045419   -4.369990286   -3.692887271   -2.467626584
-2.461568173   -3.749626537   -3.602052399   -3.613410438   -3.738116914
-1.357365122   -3.080457011   -3.212379785   -3.655696135   -3.333956892
-1.410190519   -1.322924655   -5.220695155   -2.637604062   -3.236343359
-6.090585181    1.562830613   16.766377584   -3.678460786   -3.337087238
-3.556328318   -2.988696805   -6.141479640   -3.969023989   -3.579067446
-2.724952330   -3.570205366   -5.247091327   -2.773535556   -3.099697177
-3.240647620   -3.370337754   -2.924081432   -1.199551235   -3.385903092
-1.807980434   -4.787281917   -2.998331936    2.884574172   -9.382087689
 9.591731946   -0.670773020    0.581238706    0.711796325    1.067924603
 0.773593840    0.005127222   -0.136720056   -0.427642523    0.106449737
 0.322410105    0.774264705   -0.216042433    0.040182027   -1.163421221
 0.591663582    1.292666962    1.539679171    1.795249695   -0.039904795
-0.099665584   -0.412194743   -1.301399526   -0.289479469    0.565422722
19.968499781   -1.172734766   -0.869580931    0.725626885    0.749196524
 3.411316027   -0.449890943    2.075293898   -0.623596915    2.420866151
 3.215761533   -0.268035679   -1.404507025   -1.769583095   -3.252481801
 3.765636081    8.706539149   -0.024636235   -1.833241153    3.471259395
-6.318277671    0.362691779   13.020224660    1.764048818  -52.667644523
 2.616668918   -0.016427303    0.455678396    0.148082189   -3.584033900
-2.162188623   -6.586867248    6.691720816   32.350654235  -13.098561694
-0.398430135 -150.202819597

Rapidity issue for ttbar processes

Some ttbar processes calculate rapidities without taking the absolute value. This potentially wastes 50% of the phase space events, which should be fixed.