feyncalc / loopscalla

Symbolic Loop-Healer

LoopScalla

LoopScalla is a software framework that uses FORM, Mathematica and FeynCalc to calculate multiloop diagrams.

In its current stage it is merely a preview not ready for productive usage.

QGRAF conventions

• incoming fields receive odd indices (-1, -3, -5, ...)
• outging fields receive even indices (-2, -4, -6, ...)

Loopscalla conventions

• p1, p2, ... denote incoming momenta

• q1, q2, ... denote outgoing momenta

• k1, k2, ... denote loop momenta

• lsclMu1, lsclMu2, ... denote internal Lorentz indices

• lsclNu1, lsclNu2, ... denote external Lorentz indices

• lsclCFi1, lsclCFi2, ... denote internal fundamental color indices

• lsclCFj1, lsclCFj2, ... denote external fundamental color indices

• lsclCAi1, lsclCAi2, ... denote internal adjoint color indices

• lsclCAj1, lsclCAj2, ... denote external adjoint color indices

• lsclDi1, lsclDi2, ... denote internal Dirac indices

• lsclDj1, lsclDj2, ... denote external Dirac indices

Available shell scripts

Evaluation stages

• Stage 0 - Dirac and color algebra, expansions, kinematic simplifications. Projectors or tensor reduction. The output should be ready for the topology identification step

• Stage 1 - Application of loop momenta shifts and topology mapping, rewriting of scalar propoducts in terms of inverse propagators. The output should be free any explicit loop momenta.

• Stage 2 - Insertion of IBP reduction tables.

Command line parameters

Can add -d LSCLVERBOSITY=1 to the bash scripts to have a more verbose output

Creating new project

./ShellScripts/lsclAddNewProject.sh MyProject MyModel MyProcess1 MyProcess2 ...

Script naming

Most scripts will have following suffixes attached to their names:

• All: meaning that all diagrams will be processed
• FromTo: only the given range of diagrams will be processed
• Force: force the recalculation of diagrams that have already been processed

Running jobs on a cluster where we reserve complete nodes

• The requested memory should be the product of average memory requirement per single task times the number of simultaneous GNU parallel jobs. For example, for 4GB per task and 8 jobs we need at least 32 GB of RAM.
• The number slices specifies the total number of jobs to be processed by a node. In general, the total number of single tasks divivded by the number of requested nodes should be a good number