1.5.14 (27/11/13) OM: Add warning about the fact that newprocess_mg5 is going to be remove in MG5_aMC_V2.0.0
OM: Improved cluster submision/re-submition control.
1.5.13 (04/11/13) OM: Implement a function which check if jobs submitted to cluster are correctly runned.
In case of failure, you can re-submitted the failing jobs automatically. The maximal
number of re-submission for a job can be parametrize (default 1) and how long you have to
wait before this resubmission [to avoid slow filesystem problem, i.e. condor](default 300s)
Supported cluster for this function: condor, lsf, pbs
OM: Fix a problem when more than 10k diagrams are present for a given subprocesses.
(tt~+4jets).
BF: Change nmssm model (The couplings orders were not correctly assigned for some triple
Higgs interactions)
OM: use evince by default to open eps file instead of gv.
OM: Fix a problem with the set command for the card edition for the mssm model.
OM: Update EWdim6 paper according to the snowmass paper. (3 more operator)
The default model is restricted in order to exclude those operators. In order
to have those you have to use import model EWdim6-full
OM: Fix bug #1243189, impossible to load v4 model if a local directory has the name of
the models (which is present in the models directory)
OM: Fix a bug in the complex mass scheme in the reading of the param_card (it was clearly stated)
OM: Improve numerical stability of the phase-space point generation. (thanks Z. Surujon)
1.5.12 (21/08/13) OM: Improve phase-space integration for processes with strong MMJJ cut. Cases where
the cross-section were slightly (~4%) under-evaluated due to such strong cut.
OM: Add a command print_results in the madevent interface. This command print the
cross-section/number of events/...
OM: change the way prompt color is handle (no systematic reset). Which provides better
result when the log is printed to a file. (thanks Bae Taegil)
OM: Fix Bug #1199514: Wrong assignment of mass in the lhe events file if the initial
state has one massive and one massless particles. (Thanks Wojciech Kotlarski)
OM: Fix a compilation problem for SLC6 for the installation of pythia-pgs
OM: Fix a crash linked to bug #1209113.
OM: Fix a crash if python is not a valid executation (Bug #1211777)
OM: Fix a bug in the edition of the run_card if some parameters were missing in the cards
(Bug #1183334)
1.5.11 (21/06/13) OM: Fix CRITICAL bug (returning wrong cross-section) for processes with more than
one W decaying leptonically. For such processes the lepton cuts were also used
on the neutrino particle reducing the cross-section. This bug was present only
for group_subprocesses=True (the default)
OM: Fix Bug #1184213: crash in presence of JIL mechanism (occur on some
LINUX computer only)
OM: The compilation of madevent is now performed by the number of core specify
in the configuration file. Same for pythia, ...
OM: Improve support for Read-Only system
OM: Fix a bug with the detection of the compiler when user specifiy a specific
compiler.
OM: Fix a problem that MG5 fails to compute the cross-section/width after that
a first computation fails to integrate due to a wrong mass spectrum.
OM: Fix a wrong output (impossible to compile) for pythia in presence of photon/gluon
propagator (introduce in 1.5.8)
OM: Allow to have UFO model with "goldstone" attribute instead of "GoldstoneBoson", since
FR change convention in order to match the UFO paper.
1.5.10 (16/05/13) OM: Fix Bug #1170417: fix crash for conjugate routine in presence of
massless propagator. (introduce in 1.5.9)
OM: Fix question #226810: checking that patch program exists before
trying to update MG5 code.
OM: Fix Bug #1171049: an error in the order of wavefunctions
making the code to crash (introduce in 1.5.7)
OM: Allow to use an additional syntax for the set command.
set gauge = Feynman is now valid. (Was not valid before due to the '=')
OM: Fix By Arian Abrahantes. Fix SGE cluster which was not working when
running full simulation (PGS/Delphes).
OM: adding txxxxx.cc (Thanks to Aurelijus Rinkevicius for having
written the routine)
OM: Fix Bug #1177442. This crash occurs only for very large model.
None of the model shipped with MG5 are impacted.
OM: Fix Question #228315. On some filesystem, some of the executable
loose the permission to be executable. Recover those errors
automatically.
OM: Modify the diagram enhancement technique. When more diagram have
the same propagator structure we still combine them but we now include
the interference term in the enhancement technique for those diagrams.
This fix a crash for some multi-jet process in presence of non diagonal
ckm matrices.
1.5.9 (01/06/13) JA: Fix bug in identification of symmetric diagrams, which could
give the wrong propagators included in event files for certain
processes (such as p p > z z j, z > l+ l-). Apart from the
propagators (with status 2) in the event file, this bug didn't
affect any other results (such as distributions).
JA: Fix problem in gensym which made some decay chain processes
slower than they should be. Thanks Eric Mintun for reporting.
JA: Fix problem in event clustering (introduced in v. 1.5.8)
which made events from some processes fail Pythia running.
JA: Fixed bug #1156474, Pythia 8 C++ matrix element output for
decay chain processes such as p p > z j, z > j j.
(Bug #1156474)
JA + OM: Automatically remove libpdf and libgeneric before survey,
so everything works automatically when switching between
built-in PDFs and LHAPDF.
JA: Allow syntax / to remove particles in the define command.
Example: define q = p / g
JA: Added fat warning if any decay process in a decay chain
includes a particle decaying to itself (as is the case
if you do w+ > all all, since you include w+ > w+ a).
JA: Forbid running newprocess_mg5 from a process directory
that has already been generated, to avoid confusion.
OM: Fix lxplus server issue (Bug #1159929)
OM: Fix an issue when MG5 directory is on a read only disk
(Bug #1160629)
OM: Fix a bug which prevent to have the pythia matching plot/cross-section
in some particular case.
OM: Support of new UFO convention allowing to define custom propagator.
(Both in MG5 and ALOHA)
OM: Change ALOHA default propagator to have a specific expression for the
massless case allowing to speed up matrix element computation with
photon/gluon.
OM: Correct the default spin 3/2 propagator (wrong incoming/outcoming
definition)
ML (by OM): Adding support of the SLURM cluster. Thanks to
Matthew Low for the implementation.
OM: Fixing the standalone_cpp output for the mssm model. (only model impacted)
Thanks to Silvan S Kuttimalai for reporting.
OM: Fix Bug #1162512: Wrong line splitting in cpp when some name were very long.
(shorten the name + fix the splitting)
1.5.8 (05/03/13) OM: Fix critical bug introduce in 1.5.0. ALOHA was wrongly written
HELAS routine for expression containing expression square.
(like P(-1,1)**2). None of the default model of MG5 (like sm/mssm)
have such type of expression. More information in bug report #1132996
(Thanks Gezim)
OM+JA: install Delphes now installs Delphes 3
[added command install Delphes2 to install Delphes 2]
OM: Add command in MadEvent interface: add_time_of_flight
This command modifies the lhe events file by adding the time of
flight information in the lhe events. To run this you need to do
$> ./bin/madevent
MGME> generate_events --laststep=parton -f
MGME> add_time_of_flight
MGME> pythia [if needed]
OM: Fix bug in pythia8 output for process using decay chains syntax.
See bug #1099790.
CDe+OM: Update EWdim6 model
OM: Fix a bug preventing model customized via the "customize_model"
command to use the automatic width computation.
OM: Change model restriction behavior: a value of 1 for a width is
not treated as a restriction rule.
OM: Fix incomplete restriction of the MSSM model leading to inefficient
process merging (and larger-than-necessary files) for the MSSM.
OM: Correct bug #1107603 (problem with condor cluster for submission
associated to a large number of jobs). Thanks Sanjay.
JA: Fix one part of the problem in bug #1123974: take into
account invariant mass cuts mmXX above the peak range in
s-channel resonances in the phase space integration,
to make sure such channels find events even for narrow
invariant mass cut ranges. Please note the discussion in
that bug report for other types of channels however.
JA: Fixed bug #1139303: matrix elements for identical
decay chain processes with different propagators (such as
p p > w+, w+ > e+ ve and p p > w-, w- > vm~ mu-)
are now no longer combined, to ensure that resonances are
correctly represented in event file.
OM: Support lhapdf set which contains photon (See bug #1131995).
RF+JA: Reuse last two PDF calls also for LHAPDF PDFs, clarify code
for reuse of PDFs in pdg2pdf.f and pdg2pdf_lhapdf.f
OM: Update the default delphes card to latest Delphes version. This
default card is automatically overwritten by the default Delphes
card when running "install Delphes".
JA: Make sure cuts are only checked once per event - this can save
a lot of time for multiparton event generation.
OM: Fix Bug #1142042 (crash in gridpack).
1.5.7 (15/01/13) OM+JA: Fixed crash linked to model_v4 for processes containing wwww or
zzww interactions. (See bug #1095603. Thanks to Tim Lu)
OM: Fix a bug affecting 2>1 process when the final states particles is
(outcoming fermion) introduced in version 1.5.0. (Thanks to
B. Fuks)
OM+JA: Fix a problem of fermion flow for v4 model (thanks to A. Abrahantes)
OM+DBF: Change the automatically the electroweak-scheme when passing to
complex-mass scheme: the mass of the W is the an external parameter
and Gf is an internal parameter fixed by LO gauge relation.
OM+DBF: Remove the model sm_mw of the model database.
OM: Fix problem in the ./bin/mg5 file command when some question are
present in the file.
OM: Extend support for ~ and ${vars} in path.
OM: Fix a crash in multi_run for more than 300 successive runs.
(Thanks to Diptimoy)
OM: Allow to choose the center of mass energy for the check command.
OM: small change in the pbs cluster submission (see question #218824)
OM: Adding possibility to check gauge/lorentz/...for 2>1 processes.
1.5.6 (20/12/12) JA: Replaced error with warning when there are decay processes
without corresponding core processes final state (see
Question #216037). If you get this warning, please check
carefully the process list and diagrams to make sure you
have the processes you were expecting.
JA: Included option to set the highest flavor for alpha_s reweighting
(useful for 4-flavor matching with massive b:s). Note that
this does not affect the choice of factorization scale.
JA: Fixed Bug #1089199, where decay processes with symmetric
diagrams were missing a symmetry factor.
Note that this only affects decay processes (A > B C ..)
with multiple identical particles in the final state and
some propagators not able to go on the mass shell.
JA: Updated the restriction cards for the sm model to set
Yukawa couplings equal to the corresponding masses
(in order to avoid stupid gauge check failures).
1.5.5 (18/11/12) JA: Fixed Bug #1078168, giving asymmetries in X+gamma generation
(e.g. Z+gamma) when ickkw=1 and pdfwgt=T. Thanks Irakli!
JA: Ensure that t-channel single top gives non-zero cross section
even if maxjetflavor=4 (note that if run with matching,
maxjetflavor=5 is necessary for correct PDF reweighting).
OM: Fixed Bug #1077877. Aloha crashing for pseudo-scalar, 3 bosons
interactions (introduces in 1.5.4)
OM: Fix Bug for the command "check gauge". The test of comparing
results between the two gauge (unitary and Feynman) was not
changing the gauge correctly.
OM: Improvment in LSF cluster support (see bug #1071765) Thanks to
Brian Dorney.
1.5.4 (11/11/12) JA: Fixed bug in combine_runs.py (introduced in v. 1.5.0) for
processes with 5 final-state particles, which prevented
matching to Pythia to work properly (thanks Priscila).
OM: Fix Bug #1076043, error in kinematics for e- p collisions,
thanks to Uta Klein (introduced in 1.5.0).
JA: Fix Bug #1075525, combination of decay processes for
particle and antiparticle (e.g. w+ > all all and
w- > all all), thanks Pierre.
OM: Fix a compilation crash due to aloha (thanks Tim T)
JA: Fixed dynamical scale settings for e- p collisions.
JA: Fixed running LHAPDF on a cluster with cluster_temp_path.
JA: Ensure that the seed is stored in the banner even when Pythia
is run (this was broken in v. 1.5.0).
JA: Improved and clarified phase space presetting for processes
with competing BWs.
1.5.3 (01/11/12) OM: Fix a crash in the gridpack mode (Thanks Baris Altunkaynak)
OM: Fix a crash occuring on cluster with no central disk (only
condor by default) for some complicated process.
OM: If launch command is typed before any output command,
"output madevent" is run automatically.
OM: Fix bug preventing to set width to Auto in the mssm model.
OM: Allow "set width PID VALUE" as an additional possibility to
answer edit card function.
OM: Improve ME5_debug file (include now the content of the
proc_card as well).
1.5.2 (11/10/12) OM: Fix bug for mssm model. The param_card was not read properly
for this model. (introduce in 1.5.0)
OM: If the code is run with an input file (./bin/mg5 cmd.cmd),
all question not answered in the file will be answered by the
default value. Note that piped input still can be used.
OM: Fixing a bug preventing to use the "set MH 125" command in a
script file.
JA: Fixed a bug in format of results.dat file for impossible
configurations in processes with conflicting BWs.
OM: Adding command "launch" in madevent interface which is the
exact equivalent to the launch command in the MG5 interface
in madevent output.
OM: Secure the auto-update, since we receive some report of incomplete
version file information.
1.5.1 (06/10/12) JA: Fixed symmetry factors in non-grouped MadEvent mode
(bug introduced in v. 1.5.0).
JA: Fixed phase space integration problem with multibody
decay processes (thanks Kentarou for finding this!).
OM: Fix that standalone output was not reading correctly the param_card
(introduce in 1.5.0)
OM: Fix a crash when trying to load heft
OM: Fix the case when the UFO model contains one mass which
has the same name as another parameter up to the case.
OM: Fix a bug for result lower than 1e-100 those one are now
consider as zero.
OM: Fix a bug present in the param_card create by width computation
computation where the qnumbers data were written as a float
(makes Pythia 6 crash).
1.5.0 (29/09/12) OM: Allow MG5 to run in complex mass scheme mode
(mg5> set complex_mass True)
OM: Allow MG5 to run in feynman Gauge
(mg5> set gauge Feynman)
OM: Add a new command: 'customize_model' which allow (for a
selection of model) to fine tune the model to your need.
FR team: add a file decays.py in the UFO format, this files
contains the analytical expression for one to two decays
OM: implement a function for computing the 1 to 2 width on
the fly. (requires MG5 installed on the computer, not only
the process directory)
OM: The question asking for the edition of the param_card/run_card
now accepts a command "set" to change values in those cards
without opening an editor. This allow simple implemetation
of scanning. (Thanks G. Durieux to have push me to do it)
OM: Support UFO model with spin 3/2
OM + CDe: Support four fermion interactions. Fermion flow
violation/Majorana are not yet allowed in four fermion
interactions.
OM + PdA: Allow Large Extra Dimension Model (LED) to run in the
MG5 framework.
OM: Add auto-detection if MG5 is up-to-date and propose to
apply a patch if not.
OM: MadEvent changes automatically the compiler according to
the value present in the configuration file.
OM: Aloha modifications: faster to create routines and more
optimized routines (up to 40% faster than previous version).
OM: Aloha now supports Lorentz expression with denominator.
OM: Improve error message when Source didn't compile properly.
OM: The numerical evaluation of the matrix element requires now
less memory than before (madevent and standalone output)
OM: Fix a series of bugs with the madevent command 'remove' and
'run_banner'
JA: Ensure identical results for identical seeds also with
multiple runs in the same directory. Note that identical runs
with previous versions can't be guaranteed (but different
seeds are guaranteed to give statistically independent runs).
Save the results.dat files from all runs.
JA: Amended kinematics to correctly deal with the case of
massive beams, as well as fixed-target proton collisions.
JA: Changed default in the run_card.dat to use -1 as "no cut"
for the max-type cuts (etaj, ptjmax, etc.).
JA: Added support for negative weights in matrix elements
(as required for interference-only terms) and PDFs.
JA: Avoid creating directories for integration channels
that can not give events based on BW settings
(further improvement compared to v. 1.4.8).
JA: Optimize phase space integration when there are resonances
with mass above ECM.
JA: Fixed issue in replace.pl script with more than 9 particles
in an event.
JA+OM: Allow cluster run to run everything on a local (node) disk.
This is done fully automatically for condor cluster.
For the other clusters, the user should set the variable
"cluster_temp_path" pointing to a directory (usefull only if
the directory is on the node filesystem). This still requires
access to central disk for copying, event combination,
running Pythia/PGS/Delphes etc.
OM: Replace fortran script combine_runs by a python script.
This script allows to be more stable when running on clusters
with slow filesystem response (bugs #1050269 and #1028844)
JA: Ensure that process mirroring is turned off for decay
processes of type A > B C...
