HERWIG
The latest version of HERWIG available from CERNLIB was obtained
from VXCERN (see below).
For the convenience of the CERNLIB installation procedures,
the file was converted to a PATCHY CAR file as follows:
- A TITLE was added, indicating the version, date and time of release
- A pilot patchy *HERWIG was added
- A PATCH doc, containing the .DOC file below
- The INCLUDE lines in the Fortran were converted to +CDE statements
- The INCLUDE file was converted to a PATCHY sequence (HERWIG58)
- The Fortran was split into DECKs using the CMZ FTOC command
- DOUBLE COMPLEX statements are selected IF=-IBM. IF=IBM selects COMPLEX*16
- WRITE (6,'(X)') statements were replaced by WRITE (6,'(1X)'), the
former being non-standard Fortran
- DBLE(complex) statements were converted to DBLE(REAL(complex))
The file is available both as a CMZ binary file and PATCHY car
file in the appropriate CERNLIB directories.
A new version of the Monte Carlo program HERWIG (version 5.8)
is now available for copying from the following Vax directories
at Cambridge (CBHEP is DECnet node 19616), Fermilab and CERN:
CBHEP::DISK$THEORY:[THEORY.HERWIG]HERWIG58.*
FNALV::USR$ROOT2:[BWEBBER.HERWIG]HERWIG58.*
VXCERN::DISK$CR:[WEBBER.HERWIG]HERWIG58.*
The files (total length 1700 blocks, ie 0.85 Mb) are as follows:
HERWIG58.COM (sample VAX command file for HERWIG58.TST, 1 block)
HERWIG58.DOC (documentation, 343 blocks)
HERWIG58.FOR (subroutine and BLOCK DATA source code, 1109 blocks)
HERWIG58.INC (common blocks and type declarations, 14 blocks)
HERWIG58.MSG (this message, 28 blocks)
HERWIG58.SUD (default Sudakov form factor file, 193 blocks)
HERWIG58.TST (test job main program and user routines, 12 blocks)
The PATCHY version of HERWIG has been withdrawn because it takes
a lot of work to make it and its main function was to enable users
to make a single precision version, which is now not recommended.
Users without DECnet access should be able to obtain the program
via the CERN program library (contact HEPLIBSP@CERNVM)
NEW FEATURES OF THIS VERSION
- A hard matrix element correction has been introduced in DIS (IPROC =
90**) This is switched on and off by the logical variable HARDME
(default=T). The method is essentially identical to the e+e-
correction, generating first order matrix-element events in a
phase-space region complementary to that generated by the parton
shower. The e+e- correction is also now turned on and off by HARDME
for consistency.
- Soft matrix element corrections have been introduced in DIS and e+e-
processes. These correct the distribution of emissions within the
parton shower phase-space. It is similar to the method used in
JETSET, except that the HARDEST emission is matched to the leading
order matrix element, not the first as in JETSET. This ensures that
the correction enters into the form factor, and not just the real
emission probability.
- In the backward evolution of initial-state radiation for photons the
anomalous branching q-qbar <-- gamma has been introduced.
- The treatment of forced branching of gluons and sea (anti-)quarks in
backward evolution has been improved, by allowing it to occur at a
random scale between the space-like cutoff QSPAC and the infrared
cutoff, instead of exactly at QSPAC as before.
A new option ISPAC=2 allows the freezing of structure functions at
the scale QSPAC, while evolution continues down to the infrared
cutoff. The default, ISPAC=0 is equivalent to previous versions, in
which perturbative evolution stops at QSPAC.
- It is now possible to completely switch off initial-state radiation,
by setting NOSPAC=.TRUE. Only forced splitting of non-valence
partons is generated. The default is (of course) NOSPAC=.FALSE.
- If PHOMAS is non-zero, the parton distributions for off-shell
photons are damped relative to on-shell photons, according to the
scheme defined in Drees and Godbole, MAD/PH/819. This has an
adjustable parameter PHOMAS which defines the cross-over from the
non-suppressed to the suppressed regimes. Recommended values lie in
the range QCDLAM to 1 GeV. The default, PHOMAS=0, corresponds to no
suppression as in previous versions.
- The interface to PDFLIB has been slightly changed in version 4.
Instead of indicating a PDF set by a unique number, an `author
group' string and set number are required. Version 3 of PDFLIB can
still be used from HERWIG, simply by setting the author group to
'MODE'. It is also now possible to independently set the PDF set
for each of the two beams. For example, if you previously used MRS
D- for the proton and Gordon-Storrow set 1 for the photon, by
setting
MODPDF=47
MODPHO=231
You should now set
AUTPDF(2)='MRS'
MODPDF(2)=28
AUTPDF(1)='GS'
MODPDF(1)=2
Alternatively, if you are still using PDFLIB version 3, you can set
AUTPDF(2)='MODE'
MODPDF(2)=47
AUTPDF(1)='MODE'
MODPDF(1)=231
- In the CLDIR=1 option for cluster decays a new parameter CLSMR
(default = 0.) allows a Gaussian smearing of the direction of the
perturbative quark's momentum. The smearing is actually exponential
in 1-cos(theta) with mean CLSMR. Thus increasing CLSMR decorrelates
the cluster decay from the initial quark direction.
- New subprocess have been added:
- The direct, higher twist, production of light (u,d,s) L=0 mesons
by point-like photons is now available: IPROC = 5500 all Spin =0,1
mesons, = 5510 only S=0 mesons; = 5520 only S=1 mesons.
- High transverse momentum, scalar Higgs production, in association
with a jet, is now available as IPROC =2300. Only the top quark is
included in the loops with IAPHIG controlling the approx. used: =0
zero top mass limit; = 1 exact result; = 2 infinite top mass limit
(default 1). Note the routines: HWHGJ1, HWHGJA, HWHGJB/C/D, HWUCI2
and HWULI2 use (non-standard FORTRAN-77) DOUBLE COMPLEX variables
which may not be accepted by some compilers. Users can change to
OMPLEX variables, however this involves a risk of rounding errors
spoiling numerical cancellations.
- DIS with neutrino beams is now avalable in processes IPROC = 90**.
Note that these must be marked as polarised by hand - neutrino
beams should have EPOLN(3)=-1 and antineutrino beams PPOLN(3)=+1.
- Note the DIS O(alpha_s) jet production processes, IPROC = 92**, have
been withdrawn and are no longer supported.
- A running electromagnetic coupling has been introduced, HWUAEM(Q2).
ALPHEM (now a single variable) sets the Thomson limit (Q2=0) value,
default = 0.0072993 (1/137.0).
- Two new particles have been invented: 'REMG', IDHW=71, IDHEP=9998
and 'REMN', IDHW=72, IDHEP=9999 are remnant photons and nucleons
respectively. They are identical to photons and nucleons, except
that they have gluons labelled as valence partons and, for the
nucleon, valence quark distributions are set to zero. These are
used internally by the JIMMY generator for multiple interactions,
and are not intended for general use.
- An error in setting the scale EMCMF (now called EMSCA) for QCD
decays of colour neutral particles, preventing parton showers, has
been corrected.
- Minor bugs have been corrected in: phi decays to neutral kaons; the
weights for photo-production processes; the value of EVWGT in di-jet
production by point-like photons. Routines HWHIGC and HWHIGF have
been renamed HWDHGC and HWDHGF respectively.
- The transverse momentum cutoff for final-state photon emission from
quarks, VPCUT, now defaults to 0.4 GeV. Previous versions defaulted
to SQRT(S), switching off such emission.
- The default top quark mass has been increased to 170 GeV/c^2
Version 5.1 of HERWIG was described in detail in the preprint DESY
91-048 (Cavendish-HEP-90/26), Comput. Phys. Commun. 67 (1992) 465.
The new features of each version released since then may be found
in the documentation file HERWIG58.DOC.
An up-to-date LaTeX table of available processes IPROC exists.
Mike Seymour & Bryan Webber
Lund/CERN
24 Aug 1994
CERN Program Library