Update notes for MadGraph 5 (in reverse time order) 1.4.0 (04/02/12): OM: New user interface for the madevent run. Type: 1) (from madevent output) ./bin/madevent 2) (from MG5 command line) launch [MADEVENT_PATH] -i This interface replaces various script like refine, survey, combine, run_..., rm_run, ... The script generate_events still exists but now calls ./bin/madevent. OM: For MSSM model, convert param_card to SLAH1. This card is converted to SLAH2 during the MadEvent run since the UFO model uses SLAH2. This allows to use Pythia 6, as well as having a coherent definition for the flavor. JA+OM: For decay width computations, the launch command in addition to compute the width, creates a new param_card with the width set to the associated values, and with the Branching ratio associated (usefull for pythia). NOTE: This param_card makes sense for future run ONLY if all relevant decay are generated. EXAMPLE: (after launch bin/mg5): import model sm-full generate t > b w+ define all = p b b~ l+ l- ta+ ta- vl vl~ add process w+ > all all add process z > all all define v = z w+ w- add process h > all all add process h > v v, v > all all output launch OM: change output pythia8 syntax: If a path is specified this is considered as the output directory. OM: Change the path of the madevent output files. This allows to run pythia/pgs/delphes mulitple times for the same set of events (with different pythia/... parameters). OM: Madevent output is now insensitive to the relative path to pythia-pgs, delphes, ... In consequence you don't need anymore to have your directory at the same level as Template directory. OM: MadEvent checks that the param_card is coherent with the restriction used during the model generation. OM: Model restrictions will now also force opposite number to match (helpfull for constraining to rotation matrix). OM: Change the import command. It's now allowed to omit the type of import. The type is guessed automaticaly. This is NOT allowed on the web. OM: Add a check that the fermion flow is coherent with the Lorentz structure associates to the vertex. OM: Add a check that the color representation is coherent. This allow to detect/fix various problem linked to some new models created by FR and SARAH. OM: Change the default fortran compiler to gfortran. OM: Add the possibility to force which fortran compiler will be used, either via the configuration file or via the set command. OM: Add the possibility to bypass the automatic opening of the web browser (via the configuration file: ./input/mg5_configuration.txt ) OM: add 'save options' command to save the current configuration in the configuration file. OM: Change the scheme of questions when running madevent and allow to specify in the command interface if you want to run pythia/pgs/... Allow to put the answers to the questions in the proc_card.dat. OM: Add options for the display command: a) display options: return the current option value. i.e. those set via the set command and/or via the configuration file b) display variable NAME: return the current string representation of NAME and/or self.NAME . c) display coupling_order: return the coupling orders with their associated weight (for automatic order restriction) d) display couplings now returns the list of all couplings with the associated expression e) display interactions [PART1] [PART2] [PART3] ... display all interactions containing the particles set in arguments OM: New Python script for the creation of the various html pages. This Requires less disk access for the generation of the files. OM: Modify error treatment, especially for Invalid commands and Configuration problems. JA: Ensure that we get zero cross section if we have non-parton initial states with proton/antiproton beams OM: Improve cluster support. MadEvent now supports PBS/Condor/SGE Thanks to Arian Abrahantes for the SGE implementation. OM: Improve auto-completion (better output/dealing with multi line/...) OM: Improve the parallel suite and change the release script to run some of the parallel tests. This ensures even higher stability of the code for the future releases. JA: Changed the way gridpacks work: Set granularity to 1 (so randomly select channels only if they should generate less than 1 event), but allowing channels to run down to a single iteration. This removes all old problems with increased variance for small channels in the gridpacks, while giving even faster event generation. Thanks to Johan Alwall, Sho Iwamoto for all the important testing/bug reports. 1.3.33 (01/01/12) JA: Revisited colors for propagators in addmothers.f to ensure that propagators in color flow violating processes get the correct color from initial state particles (thanks to Michele Gabusi for forcing me to do this). 1.3.32 (21/12/11) JA: Fixed a bug in the PDF reweighting routine, which caused skewed eta distributions for matched samples with pdfwgt=T. Thanks to Giulio Lenzi for finding this. 1.3.31 (29/11/11) OM: Fix a bug an overflow in RAMBO (affects standalone output only) PdA (via OM): Change RS model (add a width to the spin2) OM: Fix a bug in the cuts associate to allowed mass of all neutrinos+leptons (thanks to Brock Tweedie for finding it) OM: Remove some limitation in the name for the particles 1.3.30 (18/11/11) OM: Fix a bug for the instalation of pythia-pgs on a 64 bit UNIX machine. OM: If ROOTSYS is define but root in the PATH, add it automatically in create_matching_plots.sh This is require for the UIUC cluster. 1.3.29 (16/11/11) OM: Fixed particle identities in the Feynman diagram drawing JA: Fixed bug in pdf reweighting when external LHAPDF is used. OM+JA: Simplify the compilation of pythia-pgs package. 1.3.28 (14/11/11) OM+JA: Fix special case when Lorentz structure combining two different Majorana particles depends on the incoming/outgoing status of the Majorana particles (needed for MSSM with Goldstino). JA: Fixed problem with colors in addmothers.f for complicated multiparticle vertices and simplified color treatment (thanks to Gauthier Durieux for pointing this out). JA: Further improved gridpack parameters OM: Update the parallel test (now testing against MG5 1.3.3) OM: Include some parallel test in the release script. 1.3.27 (05/11/11) JA: Fix bug in mirrored amplitudes (sometimes amplitudes that should not be flagged as mirrored were flagged as mirrored). Thanks Marco Zaro for reporting this! JA: Fix another problem getting enough events in gridpack mode (it was not completely fixed in v. 1.3.24). Thanks Alexis! JA: Added "!" comments for all parameters in the default run_card, since apparently this is still needed for g77 to correctly read the parameters. 1.3.26 (31/10/11) JA: Fix color setting in MadEvent event file for multiparticle vertices, which was not taken into account in the upgrade in v. 1.3.18 OM: Fixed mmnl cut (inv. mass of all leptons and neutrinos) which was never active. OM: Fix td install in Linux were a chmod was missing 1.3.25 (27/10/11) JA: Ensure that the correct intermediate resonance is always written in the event file, even when we have resonances with identical properties. OM: Fix the bug forcing to quit the web browser in order to have MG5 continuing to run. OM: Change the tutorial in order to allow open index.html after the output command. 1.3.24 (22/10/11) JA: Fix problem with getting enough events in gridpack mode (this was broken in v. 1.3.11 when we moved from events to luminocity in refine). Thanks to Alexis Kalogeropoulos. 1.3.23 (19/10/11) JA: Allow user to set scales using setscales.f again (this was broken in v. 1.3.18). Thanks to Arindam Das. JA: Ensure that the error message is displayed if the "make" command is not installed on the system. 1.3.22 (12/10/11) JA: Fixed another bug (also introduced in 1.3.18), which could give the wrong ordering between the s-channel propagators for certain multiprocess cases (this also lead to a hard stop, so don't worry, if you get your events, the bug doesn't affect you). Sorry about that, this is what happens when you add a lot of new functionality... 1.3.21 (12/10/11) OM: Add a new command: install. This allow to install quite easily different package devellop for Madgraph/MadEvent. The list of available package are pythia-pgs/MadAnalysis/ExRootAnalysis/Delphes OM: Adding TopEffth Model OM: Improve display particles and autocompletion in presence of nonpropagating particles OM: Fix Aloha bug linked to four fermion operator PA: fix the problem of degenerate color basis in the diquark sextet model JA: Fixed bug in cluster.f that created a hard stop, introduced in 1.3.18. 1.3.20 (09/10/11) JA: Fixed bug in myamp.f that created a hard stop error for certain cases with many processes with different propagators in the same subprocess dir. 1.3.19 (06/10/11) JA: Fixed problem with SubProcesses makefile on Linux, introduced in 1.3.18. 1.3.18 (04/10/11) JA: Use model information to determine color of particles for reweighting and propagator color info. JA: Changed the definition of "forbidden s-channels" denoted by "$" to exclude on-shell s-channels while keeping all diagrams (i.e., complemetary to the decay chain formalism). This reduces the problems with gauge invariance compared to previously. "Onshell" is as usual defined by the "bwcutoff" flag in the run_card.dat. JA: Enable proper 4-flavor matching (such as gg>hbb~+jets) Note that you need the Pythia/PGS package v. 2.1.9 or later to use with 4F matching. Changes include: alpha_s reweighting also for b vertices, new scale treatment (mu_F for pp>hbb~ is (pT_b^max*m_Th)), no clustering of gluons to final-state massive particles in MadEvent. JA: Ensure that factorization scale settings and matching works also in singlet t-channel exchange processes like single top and VBF. The dynamic factorization scale is given by the pT of the scattered quark (on each side of the event). Note: You need the Pythia/PGS package v. 2.1.10 or later to use with VBF matching, to ensure that both radiated and scattered partons are treated correctly - scattered partons need to be excluded from the matching, since their pT can be below QCUT. An even better treatment would require to individually shower and match the two sides in Pythia, which is not presently possible. Note: In the matched 4-flavor process p p > t b~ j $ w+ w- t~ + p p > t b~ j j $ w+ w- t~, there is an admixture of t-channel single top (with up to 1 radiated jet) and s-channel single top (with up to 2 radiated jets). In this case, the automatic determination of maximum multiplicity sample doesn't work (since max in the file will be 2 jets, but for t-channel max is 1 jet). So MAXJETS=1 must be specified in the pythia_card.dat. JA: Fixed pdf reweighting for matching, which due to a mistake had never been activated. JA: Improved phase space integration presetting further by taking into account special cuts like xpt, ht etc. JA: Introduce new convention for invariant mass cuts - if max < min, exclude intermediate range (allows to exclude W/Z dijet resonances in VBF processes) 1.3.17 (30/09/11) OM: Fix a crash created by ALOHA when it tries to create the full set of ALOHA routines (pythia8 output only). 1.3.16 (11/09/11) JA: Fixed the problem from 1.3.12. 1.3.15 (09/09/11) OM: remove the fix of 1.3.12 (No events in output for some given processes) 1.3.14 (08/09/11) OM: Fix a bug in the RS model introduced in 1.3.8 1.3.13 (05/09/11) JA: Fixed bug with cut_decays=F which removed cuts also for non-decay products in certain channels if there is a forced decay present. Note that this does not affect xqcut, only pt, minv and eta cuts. JA: If non-zero phase space cutoff, don't use minimum of 1 GeV (this allows to go to e.g. 2m_e invariant mass for \gamma* > e+ e-). 1.3.12 (01/09/11) JA: Fixed problem with decay chains when different decays result in identical final states, such as p p > go go, (go > b1/b1~ b/b~, b1/b1~ > b/b~ n1) (only one of the decay chains was chosen, instead of all 3 combinations (b1,b1), (b1,b1~), (b1~,b1~)) JA: Allow for overall orders also with grouped subprocesses JA: Ensure that only leading color flows are included in event output (so no singlet flows from color octets). JA: Fixed small bug in fermion flow determination for multifermion vertices. 1.3.11 (26/08/11) JA: Improved precision of "survey" by allowing 4th and 5th iteration if accuracy after 3 iterations < 10%. JA: Subdivide BW in phase space integration for conflicting BWs also for forced decays, to improve generation with large bwcutoff in e.g. W+ W- production with decays. JA: Do refine using luminocity instead of number of events, to work with badly determined channels. JA: Don't use BW for shat if mass > sqrt(s). JA: Fixed insertion of colors for octet resonances decaying to octet+singlet (thanks Bogdan for finding this) 1.3.10 (23/08/11) OM: Update ALOHA version OM: increase waiting time for jobs to write physically the results on the disks (in ordre to reduce trouble on the cluster). 1.3.9 (01/08/11) OM: Add a new model DY_SM (arXiv:1107.5830). Thanks to Neil for the generation of the model 1.3.8 (25/07/11) JA: Replace the SM and HEFT models with latest versions using the Wolfenstein parameterization for the CKM matrix. JA: Implemented reading of the new UFO information about coupling orders (order hierarchy and expansion_order). JA: New "coupling order" specification WEIGHTED which checks for sum of coupling orders weighted by their hierarchy. JA: Implemented optimal coupling orders for processes from any model if no coupling orders specified. 1.3.7 (21/07/11) JA: Fix makefiles for some v4 models that were forgotten in v. 1.3.5 1.3.6 (18/07/11) OM: Ensure that the new makefiles work on the web 1.3.5 (14/07/11): JA: New organization of make files, ensure that compilation works for all modes (with/without LHAPDF, static/dynamic, regular/gridpack) for both Linux and Mac OS X (be careful with dynamic libraries on Mac OS X though, since it seems that common blocks might not work properly) JA: Fixed proper error messages and clean stop for compilation errors during MadEvent run. 1.3.4 (05/07/11): OM: More informative error message when a compilation error occurs 1.3.3 (29/06/11): JA: Fixed diagram symmetry for case when there are no 3-vertex-only diagrams JA (by OM): More informative error when trying to generate invalid pythia8 process 1.3.2 (14/06/11): OM: Fix fortran output when a model is case sensitive (Bug if a coupling was depending of a case sensitive parameter) SdV: Remove a annoying print in the new cuts (added in 1.3.0) OM: Fix a compilation problem in the standalone cpp output 1.3.1 (02/06/11): JA: Fixed missing file bug with the introduction of inclusive HT cut 1.3.0 (02/06/11): JA: Allow for grouped subprocesses also for MG4 models JA: Improved multiprocess diagram generation to reuse diagrams for crossed processes JA: Automatic optimization of order of particles in multiparticle labels for optimal multiprocess generation JA: Improved efficiency of identification of identical matrix elements JA: Improved identification of diagrams with identical divergency structure for grouped subprocesses JA: Included more fine-grained run options in the run_card, including helicity summation options, whether or not to set ptj and mjj automatically based on xqcut, etc. JA: Fixed some minor array limit and arithmetics warnings for extreme decay and decay chain processes. SdV: Added cuts on H_T(all jets, light and b) OM: Fixed minor bug related to cluster option in launch 1.2.4 (15/05/11): JA: Fixed long-standing bug in DECAY relating to the cross section info in block, and fixed parameter reading for MG5 SM model. 1.2.3 (11/05/11): JA: Fixed problem with scale choice in processes with mixed QED/QCD orders, e.g. p p > t t~ QED=2. Note that this fix doesn't work for p p > t t~ j j QED=4 which should still be avoided. JA: Added the ptllmin/max options in the default run_card.dat 1.2.2 (09/05/11): OM: fix ALOHA symmetries creating not gauge invariant result for scalar octet 1.2.1 (08/05/11): OM: reduce the quantity of RAM use by matrix.f OM: support speed of psyco if this python module is installed OM: fix a minor bug in the model parsing OM: add the check of valid model.pkl also for v4 model OM: add a check that UpdatesNotes is up-to-date when making a release JA: Fixed problem in phase space generation for s-channel mass > s_tot 1.2.0 (05/05/11): OM: minor fixes on check charge conjugation OM: add a check on the path for the validity of the model.pkl JA: Fixed problem with combine_runs on certain compilers 1.1.2 (03/05/11): OM+JA: Fixed problem for models with multiple interactions for the same set of particles, introduced in v. 1.1.1 1.1.1 (02/05/11): JA: Replaced (slow) diagram symmetry determination by evaluation with fast identification based on diagram tags. JA: Replacing the "p=-p" id=0 vertex produced by diagram generation algorithm already in the diagram generation, simplifying drawing, helas objects and color. JA: Fixed compiler warnings for unary operator. JA: Always set all coupling orders for diagrams (needed for NLO implementations). OM: Improved and more elegant "open" implementation for the user interface. OM: minor fixes related to checking the gauge 1.1.0 (21/04/11): JA: Removed hard limit on number of external particles in MadEvent, allowing for unlimited length decay chains there (up to 14 final state particles successfully integrated). JA: Improved helicity selection and automatic full helicity sum if needed. Optimization of run parameters. JA: New flag in run_card.dat to decide whether basic cuts are applied to decay products or not. OM: Merged ALOHA calls for different lorentz structures with the same color structures, increasing the speed and efficiency of matrix element evaluations. OM: Added new "open" command in command line interface, allowing to open standard file types directly. Automatically open crossx.html at launch. JA: Fixed MadEvent bugs for multiparton processes with conflicting decays and some faulty array limits. JA: Suppressed scary but irrelevant warnings for compiling 2->1 and 1->2 processes in MadEvent. JA: Pythia 8 output further optimized. JA, OM: Several minor fixes relating to user interface etc. 1.0.0 (12/04/11): Official release of MadGraph 5. Some of the features: - Complete FeynRules compatibility through the UFO interface - Automatic writing of HELAS routines for any model in Fortran, C++ or Python through ALOHA - Matrix element output in Fortran, C++ or Python - Output formats: MadEvent, Pythia 8, Standalone (Fortran/C++) - Support for spin 0, 1/2, 1, 2 particles - Support for color 1, 3, 6, 8 - Revamped MadEvent with improved subprocess directory organization and vastly increased speed and stability - Unlimited length decay chains (up to 12 final state particles tested with MadEvent, see v. 1.0.1) - Process checks for new model implementations - ...and much more (see paper "MadGraph 5: Going Beyond")