Millepede-II  V04-08-02
Functions/Subroutines
pede.f90 File Reference

Millepede II program, subroutines. More...

Go to the source code of this file.

Functions/Subroutines

program mptwo
 Millepede II main program Pede. More...
 
subroutine solglo (ivgbi)
 Error for single global parameter from MINRES. More...
 
subroutine solgloqlp (ivgbi)
 Error for single global parameter from MINRES-QLP. More...
 
subroutine addcst
 Add constraint information to matrix and vector. More...
 
subroutine prpcon
 Prepare constraints. More...
 
subroutine feasma
 Matrix for feasible solution. More...
 
subroutine feasib (concut, iact)
 Make parameters feasible. More...
 
subroutine peread (more)
 Read (block of) records from binary files. More...
 
subroutine peprep (mode)
 Prepare records. More...
 
subroutine pechk (ibuf, nerr)
 Check Millepede record. More...
 
subroutine pepgrp
 Parameter group info update. More...
 
subroutine pargrp (inds, inde)
 Parameter group info update for block of parameters. More...
 
subroutine isjajb (nst, is, ja, jb, jsp)
 Decode Millepede record. More...
 
subroutine loopn
 Loop with fits and sums. More...
 
subroutine ploopa (lunp)
 Print title for iteration. More...
 
subroutine ploopb (lunp)
 Print iteration line. More...
 
subroutine ploopc (lunp)
 Print sub-iteration line. More...
 
subroutine ploopd (lunp)
 Print solution line. More...
 
subroutine explfc (lunit)
 Print explanation of iteration table. More...
 
subroutine mupdat (i, j, add)
 Update element of global matrix. More...
 
subroutine mgupdt (i, j1, j2, il, jl, sub)
 Update global matrix for parameter group. More...
 
subroutine loopbf (nrej, ndfs, sndf, dchi2s, numfil, naccf, chi2f, ndff)
 Loop over records in read buffer (block), fits and sums. More...
 
subroutine prtglo
 Print final log file. More...
 
subroutine prtstat
 Print input statistic. More...
 
subroutine avprds (n, l, x, b, is)
 Product symmetric (sub block) matrix times sparse vector. More...
 
subroutine avprd0 (n, l, x, b)
 Product symmetric (sub block) matrix times vector. More...
 
subroutine anasps
 Analyse sparsity structure. More...
 
subroutine avprod (n, x, b)
 Product symmetric matrix times vector. More...
 
subroutine ijpgrp (itema, itemb, ij, lr, iprc)
 Index (region length and precision) for sparse storage of parameter groups. More...
 
integer(mpi) function ijprec (itema, itemb)
 Precision for storage of parameter groups. More...
 
integer(mpl) function ijadd (itema, itemb)
 Index for sparse storage. More...
 
real(mpd) function matij (itema, itemb)
 Get matrix element at (i,j). More...
 
subroutine mhalf2
 Fill 2nd half of matrix for extended storage. More...
 
subroutine sechms (deltat, nhour, minut, secnd)
 Time conversion. More...
 
integer(mpi) function inone (item)
 Translate labels to indices (for global parameters). More...
 
subroutine upone
 Update, redefine hash indices. More...
 
subroutine useone
 Make usable (sort items and redefine hash indices). More...
 
integer(mpi) function iprime (n)
 largest prime number < N. More...
 
subroutine loop1
 First data loop (get global labels). More...
 
subroutine loop1i
 Iteration of first data loop. More...
 
subroutine loop2
 Second data loop (number of derivatives, global label pairs). More...
 
subroutine monres
 Monitor input residuals. More...
 
subroutine vmprep (msize)
 Prepare storage for vectors and matrices. More...
 
subroutine minver
 Solution by matrix inversion. More...
 
subroutine mchdec
 Solution by Cholesky decomposition. More...
 
subroutine mdiags
 Solution by diagonalization. More...
 
subroutine zdiags
 Covariance matrix for diagonalization (,correction of eigenvectors). More...
 
subroutine mminrs
 Solution with MINRES. More...
 
subroutine mminrsqlp
 Solution with MINRES-QLP. More...
 
subroutine mcsolv (n, x, y)
 Solution for zero band width preconditioner. More...
 
subroutine mvsolv (n, x, y)
 Solution for finite band width preconditioner. More...
 
subroutine xloopn
 Standard solution algorithm. More...
 
subroutine filetc
 Interprete command line option, steering file. More...
 
subroutine filetx
 Interprete text files. More...
 
integer(mpi) function nufile (fname)
 Inquire on file. More...
 
subroutine intext (text, nline)
 Interprete text. More...
 
subroutine additem (length, list, label, value)
 add item to list More...
 
subroutine mstart (text)
 Start of 'module' printout. More...
 
subroutine mend
 End of 'module' printout. More...
 
subroutine mvopen (lun, fname)
 Open file. More...
 
subroutine petime
 Print times. More...
 
subroutine peend (icode, cmessage)
 Print exit code. More...
 
subroutine binopn (kfile, ithr, ierr)
 Open binary file. More...
 
subroutine bincls (kfile, ithr)
 Close binary file. More...
 
subroutine binrwd (kfile)
 Rewind binary file. More...
 
subroutine addsum (add)
 Accurate summation. More...
 
subroutine getsum (asum)
 Get accurate sum. More...
 

Detailed Description

Millepede II program, subroutines.

Author
Volker Blobel, University Hamburg, 2005-2009 (initial Fortran77 version)
Gero Flucke, University Hamburg (support of C-type binary files)
Claus Kleinwort, DESY (maintenance and developement)

Definition in file pede.f90.

Function/Subroutine Documentation

◆ addcst()

subroutine addcst

◆ additem()

subroutine additem ( integer(mpi), intent(inout)  length,
type(listitem), dimension(:), intent(inout), allocatable  list,
integer(mpi), intent(in)  label,
real(mpd), intent(in)  value 
)

add item to list

Parameters
[in,out]lengthlength of list
[in,out]listlist of items
[in]labelitem label
[in]valueitem value

Definition at line 10289 of file pede.f90.

References mpdef::mpl.

Referenced by intext().

◆ addsum()

subroutine addsum ( real(mpd)  add)

Accurate summation.

Parameters
[in]addsummand

Definition at line 10623 of file pede.f90.

References mpmod::accuratedsum, mpmod::accuratenexp, mpmod::accuratensum, and mpmod::nexp20.

Referenced by loopn().

◆ anasps()

subroutine anasps

Analyse sparsity structure.

Definition at line 5174 of file pede.f90.

References mpmod::globalallindexgroups, mpmod::matsto, mpmod::napgrp, mpmod::nspc, mpmod::sparsematrixcolumns, and mpmod::sparsematrixoffsets.

Referenced by loop2().

◆ avprd0()

subroutine avprd0 ( integer(mpi), intent(in)  n,
integer(mpl), intent(in)  l,
real(mpd), dimension(n), intent(in)  x,
real(mpd), dimension(n), intent(out)  b 
)

Product symmetric (sub block) matrix times vector.

A(sym) * X => B. Used by MINRES method (Is most CPU intensive part). The matrix A is the global matrix in full symmetric or (compressed) sparse storage. In full symmetric storage it could be block diagonal (MATSTO=3) and only a single block is used in the product.

Parameters
[in]nsize of (sub block) matrix
[in]loffset of (sub block) matrix
[in]xvector X
[out]bresult vector B

Definition at line 4990 of file pede.f90.

References mpmod::globalallindexgroups, mpmod::globalmatd, mpmod::globalmatf, mpmod::matsto, mpmod::mextnd, mpmod::mthrd, mpmod::napgrp, mpmod::nspc, peend(), mpmod::sparsematrixcolumns, and mpmod::sparsematrixoffsets.

Referenced by avprod(), mchdec(), mdiags(), minver(), mminrs(), mminrsqlp(), and xloopn().

◆ avprds()

subroutine avprds ( integer(mpi), intent(in)  n,
integer(mpl), intent(in)  l,
real(mpd), dimension(n), intent(in)  x,
real(mpd), dimension(n), intent(out)  b,
integer(mpi), dimension(*), intent(in)  is 
)

Product symmetric (sub block) matrix times sparse vector.

A(sym) * X => B. Used by MINRES method (Is most CPU intensive part). The matrix A is the global matrix in full symmetric or (compressed) sparse storage. In full symmetric storage it could be block diagonal (MATSTO=3) and only a single block is used in the product.

Parameters
[in]nsize of (sub block) matrix
[in]loffset of (sub block) matrix
[in]xvector X
[out]bresult vector B
[in]issparsity structure of x (number of non-zero regions, {region start, end})

Definition at line 4759 of file pede.f90.

References mpmod::globalallindexgroups, mpmod::globalmatd, mpmod::globalmatf, mpmod::matsto, mpmod::mextnd, mpmod::mthrd, mpmod::napgrp, mpmod::nspc, peend(), mpmod::sparsematrixcolumns, and mpmod::sparsematrixoffsets.

Referenced by mminrs(), and mminrsqlp().

◆ avprod()

subroutine avprod ( integer(mpi), intent(in)  n,
real(mpd), dimension(n), intent(in)  x,
real(mpd), dimension(n), intent(out)  b 
)

Product symmetric matrix times vector.

A(sym) * X => B. Used by MINRES method (Is most CPU intensive part). The matrix A is the global matrix in full symmetric or (compressed) sparse storage. Allows for size of X and smaller than size of matrix in case of solution with constriants by elimination.

Parameters
[in]nsize of matrix ( <= size of global matrix)
[in]xvector X
[in]bresult vector B

Definition at line 5267 of file pede.f90.

References avprd0(), mpmod::nagb, peend(), qlmlq(), mpmod::vecbav, and mpmod::vecxav.

Referenced by mminrs(), mminrsqlp(), solglo(), and solgloqlp().

◆ bincls()

subroutine bincls ( integer(mpi), intent(in)  kfile,
integer(mpi), intent(in)  ithr 
)

Close binary file.

Parameters
[in]kfilefile number
[in]ithrthread number ([1..nthrd] - close and reopen) for C files

Definition at line 10569 of file pede.f90.

References mpmod::nfilf.

Referenced by filetc(), and peread().

◆ binopn()

subroutine binopn ( integer(mpi), intent(in)  kfile,
integer(mpi), intent(in)  ithr,
integer(mpi), intent(out)  ierr 
)

Open binary file.

Parameters
[in]kfilefile number
[in]ithrthread number ([1..nthrd] - close and reopen) or 0 (next file - keep open) for C files
[out]ierrerror flag

Definition at line 10486 of file pede.f90.

References mpmod::nfilf, peend(), mpmod::sfd, mpmod::tfd, and mpmod::yfd.

Referenced by filetc(), and peread().

◆ binrwd()

subroutine binrwd ( integer(mpi), intent(in)  kfile)

Rewind binary file.

Parameters
[in]kfilefile number

Definition at line 10595 of file pede.f90.

References mpmod::nfilf.

Referenced by peread().

◆ explfc()

subroutine explfc ( integer(mpi)  lunit)

Print explanation of iteration table.

Definition at line 3189 of file pede.f90.

References mpmod::metsol.

Referenced by mptwo().

◆ feasib()

subroutine feasib ( real(mps), intent(in)  concut,
integer(mpi), intent(out)  iact 
)

Make parameters feasible.

Correct for constraint equation discrepancies.

Parameters
[in]concutcut for discrepancies
[out]iact=1 if correction needed, else =0

Definition at line 1630 of file pede.f90.

References dbsvx(), mpmod::globalparameter, mpmod::globalparlabelindex, mpmod::globalparvartototal, mpmod::lenconstraints, mpmod::listconstraints, mpmod::matconsblocks, mpmod::matconsproduct, mpmod::matconssort, mpmod::ncblck, mpmod::ncgb, mpmod::nvgb, mpmod::vecconsresiduals, mpmod::vecconssolution, and mpmod::vecconsstart.

Referenced by xloopn().

◆ feasma()

subroutine feasma

◆ filetc()

subroutine filetc

Interprete command line option, steering file.

Fetch and interprete command line options, if steering file specified, check file existence (calling NUFILE)

If no steering file specified, check default steering file.

Create test files for command line option '-t'.

Read steering file, print some lines, detect names of text and binary files, check file existence, store all file names.

Open all binary files.

Definition at line 8828 of file pede.f90.

References bincls(), binopn(), mpmod::cfd, mpmod::dfd, mpmod::filnam, mpmod::icheck, mpmod::ictest, mpmod::ifd, mpmod::ifile, mpmod::iforce, intext(), mpmod::isubit, mpmod::jfd, mpmod::keepopen, mpmod::kfd, mpmod::lfd, mpmod::mdebug, mpmod::mfd, mpmod::mprint, mptest(), mptst2(), mstart(), mpmod::mthrdr, mpmod::nfd, mpmod::nfilb, mpmod::nfilc, mpmod::nfiles, mpmod::nfilf, mpmod::nfilw, mpmod::nfnam, mpmod::ofd, peend(), ratext(), rltext(), mpmod::sfd, mpmod::tfd, mpmod::times, mpmod::wfd, mpmod::xfd, and mpmod::yfd.

Referenced by mptwo().

◆ filetx()

subroutine filetx

Interprete text files.

Reset flags and read steering and all other text files. Print some lines from each file.

Store parameter values, constraints and measurements.

Check flags METSOL (method of solution) and MATSTO (matrix storage mode). Set default values for flags, which are undefined.

Parameter values, format:

  1  label
  2  (initial) parameter value
  3  pre-sigma
  4  label
  5  (initial) parameter value
  6  pre-sigma
  7  label
...  ...
(number of words is multiple of 3)

Constraint data, format:

  1  0                 ! constraint header of four words:
  2  right-hand-side   ! 0 and -1 ...
  3  -1; -2            ! ... indicate (weighting) ...
  4  sigma             ! ... header
  5  label
  6  factor
  7  label
  8  factor
  9  ...
...  ...
(number of words is multiple of 2, at least 6)

Measured data, format:

  1  0                 ! constraint header of four words:
  2  right-hand-side   ! 0 and -1 ...
  3  -1                ! ... indicate ...
  4  sigma             ! ... header
  5  label
  6  factor
  7  label
  8  factor
  9  ...
...  ...
(number of words is multiple of 2, at least 6)

Definition at line 9288 of file pede.f90.

References mpmod::chicut, mpmod::chirem, mpmod::dwcut, mpmod::filnam, intext(), mpmod::lfd, mpmod::lhuber, mpmod::lsearch, mpmod::lunkno, mpmod::matsto, mpmod::mbandw, mend(), mpmod::metsol, mpmod::mfd, mpmod::mitera, mpmod::mprint, mpmod::nfd, mpmod::nfiles, mpmod::nfnam, mpmod::nummeasurements, peend(), rltext(), and mpmod::tfd.

Referenced by mptwo().

◆ getsum()

subroutine getsum ( real(mpd), intent(out)  asum)

Get accurate sum.

Parameters
[out]asumaccurate sum

Definition at line 10648 of file pede.f90.

References mpmod::accuratedsum, mpmod::accuratenexp, mpmod::accuratensum, and mpmod::nexp20.

Referenced by loopn().

◆ ijadd()

integer(mpl) function ijadd ( integer(mpi), intent(in)  itema,
integer(mpi), intent(in)  itemb 
)

Index for sparse storage.

In case of (compressed) sparse storage calculate index for off-diagonal matrix element.

Parameters
[in]itemarow number
[in]itembcolumn number
Returns
index (>(<) 0: double(single) precision element, =0: not existing)

Definition at line 5434 of file pede.f90.

References mpmod::globalallindexgroups, mpmod::globalallpartogroup, ijpgrp(), and mpmod::nagb.

Referenced by matij().

◆ ijpgrp()

subroutine ijpgrp ( integer(mpi), intent(in)  itema,
integer(mpi), intent(in)  itemb,
integer(mpl), intent(out)  ij,
integer(mpi), intent(out)  lr,
integer(mpi), intent(out)  iprc 
)

Index (region length and precision) for sparse storage of parameter groups.

Calculate index for parameter group block matrix (region of continous groups)

Parameters
[in]itemarow number
[in]itembcolumn number
[out]ijindex of first element (>(<) 0: double(single) precision element, =0: not existing)
[out]lrlength of region (2nd row in group has index ij+lr)
[out]iprcprecision (1: REAL(mpd), 2: REAL(mps))

Definition at line 5307 of file pede.f90.

References mpmod::globalallindexgroups, mpmod::napgrp, mpmod::nspc, mpmod::sparsematrixcolumns, and mpmod::sparsematrixoffsets.

Referenced by ijadd(), ijprec(), and mgupdt().

◆ ijprec()

integer(mpi) function ijprec ( integer(mpi), intent(in)  itema,
integer(mpi), intent(in)  itemb 
)

Precision for storage of parameter groups.

Parameters
[in]itemarow number
[in]itembcolumn number
Returns
precision (1: REAL(mpd), 2: REAL(mps))

Definition at line 5405 of file pede.f90.

References ijpgrp(), mpmod::matsto, and mpmod::nspc.

◆ inone()

integer(mpi) function inone ( integer(mpi), intent(in)  item)

Translate labels to indices (for global parameters).

Functions INONE and subroutine UPONE are used to collect items, i.e. labels, and to order and translate them.

In the first phase items are collected and stored by calling IRES=INONE(ITEM).

At the first entry the two sub-arrays "a" (globalParLabelIndex) and "b" (globalParHashTable) of length 2N are generated with a start length for N=128 entries. In array "a" two words are reserved for each item: (ITEM, count). The function INONE(ITEM) returns the number of the item. At each entry the argument is compared with the already stored items, new items are stored. Search for entries is done using hash-indices, stored in sub-array "b". The initial hash-index is

   j = 1 + mod(ITEM, n_prime) + N

where n_prime is the largest prime number less than N. At each entry the count is increased by one. If N items are stored, the size of the sub-arrays is increased by calling CALL UPONE.

Parameters
[in]itemlabel
Returns
index

Definition at line 5663 of file pede.f90.

References mpmod::globalparhashtable, mpmod::globalparheader, mpmod::globalparlabelindex, iprime(), mpmod::lunlog, mpmod::lvllog, and upone().

◆ intext()

subroutine intext ( character (len=*), intent(in)  text,
integer(mpi), intent(in)  nline 
)

◆ iprime()

integer(mpi) function iprime ( integer(mpi), intent(in)  n)

largest prime number < N.

Parameters
[in]nN
Returns
largest prime number < N

Definition at line 5825 of file pede.f90.

References mpdef::mpi, and mpdef::mps.

Referenced by inone().

◆ isjajb()

subroutine isjajb ( integer(mpi), intent(in)  nst,
integer(mpi), intent(inout)  is,
integer(mpi), intent(out)  ja,
integer(mpi), intent(out)  jb,
integer(mpi), intent(out)  jsp 
)

Decode Millepede record.

Get indices JA, JB, IS for next measurement within record:

  • Measurement is: GLDER(JA)
  • Local derivatives are: (INDER(JA+J),GLDER(JA+J),J=1,JB-JA-1) i.e. JB-JA-1 derivatives
  • Standard deviation is: GLDER(JB)
  • Global derivatives are: (INDER(JB+J),GLDER(JB+J),J=1,IS-JB) i.e. IST-JB derivatives

End_of_data is indicated by returned values JA=0 and JB=0 Special data are ignored. At end_of_data the info to the special data is returned: IS = pointer to special data; number of words is NSP=-GLDER(IS).

Parameters
[in]nstindex of last word of record
[in,out]isindex of last global derivative (index of first word of record at the first call)
[out]jaindex of measured value (=0 at end), = pointer to local derivatives
[out]jbindex of standard deviation (=0 at end), = pointer to global derivatives
[out]jspindex to special data

Definition at line 2552 of file pede.f90.

References mpmod::readbufferdatad, and mpmod::readbufferdatai.

Referenced by loop1i(), loop2(), loopbf(), pepgrp(), and peprep().

◆ loop1()

subroutine loop1

◆ loop1i()

subroutine loop1i

Iteration of first data loop.

Read all data files again skipping measurements with any parameter below the entries cut to update the number of entries.

Redefine variable and fixed global parameters (depending on updated entries).

Definition at line 6198 of file pede.f90.

References mpmod::globalparcounts, mpmod::globalparheader, mpmod::globalparlabelindex, isjajb(), mpmod::iteren, mpmod::lunlog, mpmod::mreqenf, mpmod::mthrdr, mpmod::ncache, mpmod::ndimbuf, mpmod::ntgb, mpmod::numreadbuffer, mpmod::nvgb, peprep(), peread(), mpmod::readbufferdatad, mpmod::readbufferdataf, mpmod::readbufferdatai, and mpmod::readbufferpointer.

Referenced by loop1().

◆ loop2()

subroutine loop2

Second data loop (number of derivatives, global label pairs).

Calculate maximum number of local, global derivatives and equations per record.

For sparse storage count index pairs with bit (field) counters to construct sparsity structure (row offsets, (compressed) column lists).

Determine read/write cache splitting from average record values (length, global par. vector/matrix).

Check constraints for rank deficit.

Definition at line 6318 of file pede.f90.

References anasps(), mpmod::appearancecounter, mpmod::backindexusage, ckbits(), clbits(), clbmap(), mpmod::dflim, mpmod::fcache, feasma(), mpmod::globalallindexgroups, mpmod::globalallpartogroup, mpmod::globalindexranges, mpmod::globalindexusage, mpmod::globalparlabelindex, mpmod::globalparvartototal, mpmod::globaltotindexgroups, gpbmap(), hmpdef(), mpmod::icelim, mpmod::icheck, mpmod::ifd, mpmod::imonit, inbits(), inbmap(), isjajb(), mpmod::lenmeasurements, mpmod::listconstraints, mpmod::listmeasurements, mpmod::localglobalmap, mpmod::localglobalmatrix, mpmod::localglobalstructure, mpmod::lunlog, mpmod::lunmon, mpmod::matconsblocks, mpmod::matconssort, mpmod::matmon, mpmod::matparblockoffsets, mpmod::matsto, mpmod::mbandw, mpmod::mdebg2, mpmod::mdebug, mpmod::measbins, mpmod::meashists, mpmod::measindex, mpmod::measres, mend(), mpmod::metsol, mpmod::mextnd, mpmod::mhispe, mpmod::mitera, mpmod::mprint, mpmod::mreqpe, mpmod::mrestl, mpmod::msngpe, mstart(), mpmod::mthrd, mpmod::mthrdr, mvopen(), mpmod::mxrec, mpmod::naeqn, mpmod::nagb, mpmod::nagbn, mpmod::nalcn, mpmod::napgrp, mpmod::ncache, mpmod::ncblck, mpmod::ncgb, ndbits(), mpmod::ndimbuf, mpmod::nfgb, mpmod::nhistp, mpmod::npblck, mpmod::nprecond, mpmod::nrec, mpmod::nspc, mpmod::ntgb, mpmod::ntpgrp, mpmod::nummeas, mpmod::numreadbuffer, mpmod::nvgb, mpmod::nvpgrp, mpmod::paircounter, peend(), peprep(), peread(), prpcon(), mpmod::readbufferdatad, mpmod::readbufferdataf, mpmod::readbufferdatai, mpmod::readbufferpointer, sort1k(), mpmod::sparsematrixcolumns, mpmod::sparsematrixoffsets, spbits(), mpmod::vecconsstart, mpmod::vecparblockoffsets, vmprep(), mpmod::wolfc1, mpmod::wolfc2, mpmod::writebufferdata, mpmod::writebufferheader, mpmod::writebufferindices, mpmod::writebufferinfo, and mpmod::writebufferupdates.

Referenced by mptwo().

◆ loopbf()

subroutine loopbf ( integer(mpi), dimension(0:3), intent(inout)  nrej,
integer(mpi), intent(inout)  ndfs,
real(mpd), intent(inout)  sndf,
real(mpd), intent(inout)  dchi2s,
integer(mpi), intent(in)  numfil,
integer(mpi), dimension(numfil), intent(inout)  naccf,
real(mps), dimension(numfil), intent(inout)  chi2f,
integer(mpi), dimension(numfil), intent(inout)  ndff 
)

Loop over records in read buffer (block), fits and sums.

Loop over records in current read buffer block (with multiple threads). Perform local fits (optionally with outlier downweigthing) to calculate Chi2, ndf and r.h.s. 'b' of linear equation system A*x=b. In first iteration(s) fill global matrix A.

For the filling of the global matrix each thread creates from his share of local fits (small) udpdate matrices ( $\Vek{\D C}_1+\Vek{\D C}_2$ from equations (15), (16)) stored in a write buffer. After all events in the read buffer block have been processed the global matrix is being updated from the matrices in the write buffer in parallel (each row by different thread).

The matrices of the local fits are checked for bordered band structure. For border size b and band width m all elements (i,j) are zero for min(i,j)>b and abs(i-j)>m. For sufficient small (b,m) a solution by root free Cholesky decomposition and forward/backward substitution of the band part is much faster compared to inversion (see broken lines in references). Based on the expected computing cost the faster solution method is selected.

Parameters
[in,out]nrejnumber of rejected records
[in,out]ndfssum(ndf)
[in,out]sndfsum(weighted ndf)
[in,out]dchi2ssum(weighted chi2)
[in]numfilnumber of binary files
[in,out]naccfnumber of accepted records per binary file
[in,out]chi2fsum(chi2/ndf) per binary file
[in,out]ndffsum(ndf) per binary file

Definition at line 3476 of file pede.f90.

References mpmod::aux, mpmod::backindexusage, mpmod::blvec, mpmod::chhuge, mpmod::chicut, mpmod::clmat, dbavat(), dbavats(), mpmod::dwcut, mpmod::globalallindexgroups, mpmod::globalallpartogroup, mpmod::globalcounter, mpmod::globalindexusage, mpmod::globalparameter, mpmod::globalparlabelindex, mpmod::globalvector, mpmod::ibandh, mpmod::icalcm, mpmod::imonit, mpmod::imonmd, isjajb(), mpmod::iterat, mpmod::lfitbb, mpmod::lfitnp, mpmod::lhuber, mpmod::localcorrections, mpmod::localequations, mpmod::localglobalmap, mpmod::localglobalmatrix, mpmod::localglobalstructure, mpmod::lunlog, mpmod::measbins, mpmod::measbinsize, mpmod::meashists, mpmod::measindex, mpmod::measres, mgupdt(), mpmod::monpg1, mpmod::mthrd, mpmod::naeqn, mpmod::nagb, mpmod::nagbn, mpmod::nbdrx, mpmod::nbndr, mpmod::nbndx, mpmod::newite, mpmod::nloopn, mpmod::nrec1, mpmod::nrec2, mpmod::nrec3, mpmod::nrecer, mpmod::nrecp2, mpmod::nrecpr, mpmod::nummeas, mpmod::numreadbuffer, mpmod::nvgb, mpmod::readbufferdatad, mpmod::readbufferdatai, mpmod::readbufferpointer, mpmod::scdiag, mpmod::scflag, sort1k(), sqmibb(), sqmibb2(), sqminv(), mpmod::value1, mpmod::value2, mpmod::vbdr, mpmod::vbk, mpmod::vbnd, mpmod::vzru, mpmod::writebufferdata, mpmod::writebufferheader, mpmod::writebufferindices, mpmod::writebufferinfo, and mpmod::writebufferupdates.

Referenced by loopn().

◆ loopn()

subroutine loopn

◆ matij()

real(mpd) function matij ( integer(mpi), intent(in)  itema,
integer(mpi), intent(in)  itemb 
)

Get matrix element at (i,j).

Parameters
[in]itemarow number
[in]itembcolumn number
Returns
value

Definition at line 5481 of file pede.f90.

References mpmod::globalindexranges, mpmod::globalmatd, mpmod::globalmatf, ijadd(), mpmod::matparblockoffsets, mpmod::matsto, mpmod::nagb, and mpmod::vecparblockoffsets.

◆ mchdec()

subroutine mchdec

◆ mcsolv()

subroutine mcsolv ( integer(mpi), intent(in)  n,
real(mpd), dimension(n), intent(in)  x,
real(mpd), dimension(n), intent(out)  y 
)

Solution for zero band width preconditioner.

Used by MINRES.

Parameters
[in]nsize of vectors
[in]xrhs vector
[out]yresult vector

Definition at line 8072 of file pede.f90.

References mpmod::matprecond, mpmod::ncgb, mpmod::nprecond, mpmod::nvgb, and presol().

Referenced by mminrs(), mminrsqlp(), solglo(), and solgloqlp().

◆ mdiags()

subroutine mdiags

◆ mend()

subroutine mend

End of 'module' printout.

Definition at line 10361 of file pede.f90.

References petime(), and mpmod::textl.

Referenced by filetx(), loop1(), loop2(), and xloopn().

◆ mgupdt()

subroutine mgupdt ( integer(mpi), intent(in)  i,
integer(mpi), intent(in)  j1,
integer(mpi), intent(in)  j2,
integer(mpi), intent(in)  il,
integer(mpi), intent(in)  jl,
real(mpd), dimension(*), intent(in)  sub 
)

Update global matrix for parameter group.

Add values -SUB to matrix elements (continous block in smaller index) .

Parameters
[in]ilarger index
[in]j1smaller index first group
[in]j2smaller index last group
[in]ilsubtrahends first row
[in]jlsubtrahends first col
[in]subsubtrahends matrix ('small', size fits in 'mpi')

Definition at line 3355 of file pede.f90.

References mpmod::globalallindexgroups, mpmod::globalindexranges, mpmod::globalmatd, mpmod::globalmatf, ijpgrp(), mpmod::matparblockoffsets, mpmod::matsto, and mpmod::vecparblockoffsets.

Referenced by loopbf().

◆ mhalf2()

subroutine mhalf2

Fill 2nd half of matrix for extended storage.

Definition at line 5531 of file pede.f90.

References mpmod::globalallindexgroups, mpmod::globalmatd, mpmod::globalmatf, mpmod::mthrd, mpmod::napgrp, mpmod::nspc, mpmod::sparsematrixcolumns, and mpmod::sparsematrixoffsets.

Referenced by loopn().

◆ minver()

subroutine minver

◆ mminrs()

subroutine mminrs

◆ mminrsqlp()

subroutine mminrsqlp

◆ monres()

subroutine monres

Monitor input residuals.

Read all data files again to monitor input residuals

Definition at line 7248 of file pede.f90.

References mpmod::globalparlabelindex, mpmod::imonmd, mpmod::lunmon, mpmod::measbins, mpmod::measbinsize, mpmod::meashists, mpmod::measindex, mpmod::measres, mpmod::mthrd, mpmod::nloopn, mpmod::ntgb, and mpmod::nummeas.

Referenced by loopn(), and xloopn().

◆ mptwo()

program mptwo

◆ mstart()

subroutine mstart ( character (len=*), intent(in)  text)

Start of 'module' printout.

Definition at line 10325 of file pede.f90.

References mpmod::textl.

Referenced by filetc(), loop1(), loop2(), and mptwo().

◆ mupdat()

subroutine mupdat ( integer(mpi), intent(in)  i,
integer(mpi), intent(in)  j,
real(mpd), intent(in)  add 
)

Update element of global matrix.

Add value ADD to matrix element (I,J).

Parameters
[in]ifirst index
[in]jsecond index
[in]addsummand

Definition at line 3280 of file pede.f90.

References mpmod::globalindexranges, mpmod::globalmatd, mpmod::globalmatf, mpmod::indprecond, mpmod::matparblockoffsets, mpmod::matprecond, mpmod::matsto, mpmod::mbandw, mpmod::metsol, mpmod::nvgb, peend(), and mpmod::vecparblockoffsets.

Referenced by addcst(), and loopn().

◆ mvopen()

subroutine mvopen ( integer(mpi), intent(in)  lun,
character (len=*), intent(in)  fname 
)

Open file.

Evtl. move existing file to ~ version.

Parameters
[in]lununit number
[in]fnamefile name

Definition at line 10378 of file pede.f90.

References peend().

Referenced by loop2(), mptwo(), peend(), prtglo(), and prtstat().

◆ mvsolv()

subroutine mvsolv ( integer(mpi), intent(in)  n,
real(mpd), dimension(n), intent(in)  x,
real(mpd), dimension(n), intent(out)  y 
)

Solution for finite band width preconditioner.

Used by MINRES.

Parameters
[in]nsize of vectors
[in]xrhs vector
[out]yresult vector

Definition at line 8092 of file pede.f90.

References equslv(), mpmod::indprecond, mpmod::matprecond, and mpmod::nprecond.

Referenced by mminrs(), mminrsqlp(), solglo(), and solgloqlp().

◆ nufile()

integer(mpi) function nufile ( character (len=*), intent(inout)  fname)

Inquire on file.

Result = 1 for existing binary file, =2 for existing text file, else =0, < 0 open error.

Text file names are recognized by the filename extension, which should contain 'xt' or 'tx'.

Parameters
[in,out]fnamefile name, optionaly strip prefix.

Definition at line 9526 of file pede.f90.

References matint().

◆ pargrp()

subroutine pargrp ( integer(mpi), intent(in)  inds,
integer(mpi), intent(in)  inde 
)

Parameter group info update for block of parameters.

Build and split groups (defined by common first parameter).

Parameters
[in]indsindex of first parmeters in read buffer
[in]indeindex of last parmeters in read buffer

Definition at line 2442 of file pede.f90.

References mpmod::globalparheader, mpmod::globalparlabelindex, and mpmod::readbufferdatai.

Referenced by pepgrp().

◆ pechk()

subroutine pechk ( integer(mpi), intent(in)  ibuf,
integer(mpi), intent(out)  nerr 
)

Check Millepede record.

Check integer structure of labels and markers (zeros). Check floats for NaNs.

Parameters
[in]ibufbuffer number
[out]nerrerror flags

Definition at line 2228 of file pede.f90.

References mpmod::readbufferdatad, mpmod::readbufferdatai, and mpmod::readbufferpointer.

Referenced by peprep().

◆ peend()

subroutine peend ( integer(mpi), intent(in)  icode,
character (len=*), intent(in)  cmessage 
)

Print exit code.

Print exit code and message.

Parameters
[in]icodeexit code
[in]cmessageexit massage

Definition at line 10466 of file pede.f90.

References mvopen().

Referenced by avprd0(), avprds(), avprod(), binopn(), devrot(), feasma(), filetc(), filetx(), loop1(), loop2(), mpdalc::mpalloccheck(), mpdalc::mpdealloccheck(), mptwo(), mupdat(), mvopen(), peprep(), peread(), sort1k(), sort22(), sort2i(), sort2k(), and xloopn().

◆ pepgrp()

subroutine pepgrp

Parameter group info update.

Group parameters on level of equations or records (counting in addition).

Definition at line 2308 of file pede.f90.

References mpmod::backindexusage, mpmod::globalparheader, mpmod::globalparlabelindex, isjajb(), mpmod::mcount, mpmod::mthrd, mpmod::numreadbuffer, pargrp(), mpmod::readbufferdatai, mpmod::readbufferpointer, sort1k(), and useone().

Referenced by loop1().

◆ peprep()

subroutine peprep ( integer(mpi), intent(in)  mode)

Prepare records.

For global parameters replace label by index (INONE).

Parameters
[in]mode<=0: build index table (INONE) for global variables;
>0: use index table, can be parallelized, optional scale errors

Definition at line 2131 of file pede.f90.

References mpmod::dscerr, mpmod::iscerr, isjajb(), mpmod::mthrd, mpmod::numreadbuffer, pechk(), peend(), mpmod::readbufferdatad, mpmod::readbufferdatai, and mpmod::readbufferpointer.

Referenced by loop1(), loop1i(), loop2(), and loopn().

◆ peread()

subroutine peread ( integer(mpi), intent(out)  more)

Read (block of) records from binary files.

Optionally using several threads (each file read by single thread). Records larger than the read buffer (ndimbuf) are skipped. In case of skipped events in the first loop over all binary files the buffer size is adapted to the maximum record size (and the initial loop (LOOP1) is repeated). C binary files are handled by readc.c and may be gzipped.

Parameters
[out]moremore records to come

The records consist of parallel integer and real arrays:

    real array              integer array
1   0.0                     error count (this record)
2   RMEAS, measured value   0                            JA
3   local derivative        index of local derivative
4   local derivative        index of local derivative
5    ...
6   SIGMA, error (>0)       0                            JB
    global derivative       label of global derivative
    global derivative       label of global derivative   IST
    RMEAS, measured value   0
    local derivative        index of local derivative
    local derivative        index of local derivative
    ...
    SIGMA, error            0
    global derivative       label of global derivative
    global derivative       label of global derivative
    ...
NR  global derivative       label of global derivative

Definition at line 1791 of file pede.f90.

References bincls(), binopn(), binrwd(), mpmod::icheck, mpmod::ifd, mpmod::ifile, mpmod::ireeof, mpmod::keepopen, mpmod::kfd, mpmod::lunlog, mpmod::maxrecordsinblock, mpmod::minrecordsinblock, mpmod::mprint, mpmod::mthrdr, mpmod::mxrec, mpmod::ncache, mpmod::ndimbuf, mpmod::nfilb, mpmod::nfilf, mpmod::nfilw, mpmod::nloopn, mpmod::nrderr, mpmod::nrec, mpmod::nrecd, mpmod::numblocks, mpmod::numreadbuffer, peend(), mpmod::readbufferdatad, mpmod::readbufferdataf, mpmod::readbufferdatai, mpmod::readbufferinfo, mpmod::readbufferpointer, mpmod::skippedrecords, sort2k(), mpmod::sumrecords, mpmod::wfd, and mpmod::xfd.

Referenced by loop1(), loop1i(), loop2(), and loopn().

◆ petime()

subroutine petime

Print times.

Print the elapsed and total time.

Definition at line 10415 of file pede.f90.

References mpdef::mpi.

Referenced by mend().

◆ ploopa()

subroutine ploopa ( integer(mpi), intent(in)  lunp)

Print title for iteration.

Parameters
[in]lunpunit number

Definition at line 3037 of file pede.f90.

Referenced by xloopn().

◆ ploopb()

subroutine ploopb ( integer(mpi), intent(in)  lunp)

Print iteration line.

Parameters
[in]lunpunit number

Definition at line 3058 of file pede.f90.

References mpmod::angras, mpmod::chicut, mpmod::delfun, mpmod::deltim, mpmod::fvalue, mpmod::iitera, mpmod::istopa, mpmod::iterat, mpmod::lcalcm, mpmod::lsinfo, mpmod::nloopn, mpmod::nrejec, ptlopt(), mpmod::rstart, sechms(), and mpmod::stepl.

Referenced by xloopn().

◆ ploopc()

subroutine ploopc ( integer(mpi), intent(in)  lunp)

Print sub-iteration line.

Parameters
[in]lunpunit number

Definition at line 3115 of file pede.f90.

References mpmod::deltim, mpmod::fvalue, mpmod::lcalcm, mpmod::lsinfo, mpmod::nloopn, mpmod::nrejec, ptlopt(), mpmod::rstart, sechms(), and mpmod::stepl.

Referenced by xloopn().

◆ ploopd()

subroutine ploopd ( integer(mpi), intent(in)  lunp)

Print solution line.

Parameters
[in]lunpunit number

Definition at line 3163 of file pede.f90.

References mpmod::deltim, mpmod::lcalcm, mpmod::rstart, and sechms().

Referenced by xloopn().

◆ prpcon()

subroutine prpcon

◆ prtglo()

subroutine prtglo

Print final log file.

For each global parameter:

  • label (I10)
  • parameter value (G14.5)
  • presigma (G14.5)
  • difference of parameters values (G14.5)
  • difference at last iteration (G14.5)
  • error (standard deviation) (G14.5)
  • global correlation (F8.3), on request only
  • Entries from binary files or during iterations

Definition at line 4504 of file pede.f90.

References mpmod::globalcounter, mpmod::globalindexranges, mpmod::globalmatd, mpmod::globalparameter, mpmod::globalparcounts, mpmod::globalparlabelindex, mpmod::globalparpresigma, mpmod::globalparstart, mpmod::globalparvartototal, mpmod::igcorr, mpmod::ipcntr, mpmod::matparblockoffsets, mpmod::metsol, mvopen(), mpmod::nagb, mpmod::ntgb, mpmod::nvgb, mpmod::vecparblockoffsets, mpmod::workspacediag, mpmod::workspaceeigenvalues, and mpmod::workspaceeigenvectors.

Referenced by xloopn().

◆ prtstat()

subroutine prtstat

Print input statistic.

For each global parameter:

  • label (I10)
  • parameter value (G14.5)
  • presigma (G14.5)
  • difference of parameters values (G14.5), = 0.
  • Entries from binary files

Definition at line 4686 of file pede.f90.

References mpmod::appearancecounter, mpmod::globalparameter, mpmod::globalparcons, mpmod::globalparcounts, mpmod::globalparlabelindex, mpmod::globalparpresigma, mpmod::icheck, mvopen(), mpmod::ntgb, and mpmod::paircounter.

Referenced by mptwo().

◆ sechms()

subroutine sechms ( real(mps), intent(in)  deltat,
integer(mpi), intent(out)  nhour,
integer(mpi), intent(out)  minut,
real(mps), intent(out)  secnd 
)

Time conversion.

Convert from seconds to hours, minues, seconds

Parameters
[in]deltattime in seconds
[out]nhourhours
[out]minutminutes
[out]secndseconds

Definition at line 5618 of file pede.f90.

References mpdef::mpi.

Referenced by mptwo(), ploopb(), ploopc(), and ploopd().

◆ solglo()

subroutine solglo ( integer(mpi), intent(in)  ivgbi)

Error for single global parameter from MINRES.

Calculate single row 'x_i' from inverse matrix by solving A*x_i=b with b=0 except b_i=1.

Parameters
[in]ivgbiindex of variable parameter

Definition at line 1035 of file pede.f90.

References avprod(), mpmod::globalcorrections, mpmod::globalparameter, mpmod::globalparlabelindex, mpmod::globalparpresigma, mpmod::globalparstart, mpmod::globalparvartototal, mpmod::globalvector, mpmod::mbandw, mcsolv(), minresmodule::minres(), mpmod::mrestl, mvsolv(), and mpmod::nagb.

Referenced by xloopn().

◆ solgloqlp()

subroutine solgloqlp ( integer(mpi), intent(in)  ivgbi)

Error for single global parameter from MINRES-QLP.

Calculate single row 'x_i' from inverse matrix by solving A*x_i=b with b=0 except b_i=1.

Parameters
[in]ivgbiindex of variable parameter

Definition at line 1119 of file pede.f90.

References avprod(), mpmod::globalcorrections, mpmod::globalparameter, mpmod::globalparlabelindex, mpmod::globalparpresigma, mpmod::globalparstart, mpmod::globalparvartototal, mpmod::globalvector, mpmod::mbandw, mcsolv(), minresqlpmodule::minresqlp(), mpmod::mrestl, mpmod::mrmode, mpmod::mrtcnd, mvsolv(), and mpmod::nagb.

Referenced by xloopn().

◆ upone()

subroutine upone

Update, redefine hash indices.

Definition at line 5730 of file pede.f90.

References mpmod::globalparhashtable, mpmod::globalparheader, mpmod::globalparlabelindex, mpmod::lunlog, mpmod::lvllog, and sort22().

Referenced by inone(), and loop1().

◆ useone()

subroutine useone

Make usable (sort items and redefine hash indices).

Definition at line 5793 of file pede.f90.

References mpmod::globalparhashtable, mpmod::globalparheader, mpmod::globalparlabelindex, and sort22().

Referenced by pepgrp().

◆ vmprep()

subroutine vmprep ( integer(mpl), dimension(2), intent(in)  msize)

◆ xloopn()

subroutine xloopn

Standard solution algorithm.

Iterative solution. In current iteration calculate:

ICALCM = +1   Matrix, gradient, Function value & solution
ICALCM =  0   gradient, Function value
ICALCM = -1   solution
ICALCM = -2   end

Solution is obtained by selected method and improved by line search.

Definition at line 8125 of file pede.f90.

References addcst(), mpmod::angras, avprd0(), mpmod::chicut, mpmod::chirem, mpmod::delfun, mpmod::deltim, mpmod::dflim, mpmod::dscerr, mpmod::dwcut, feasib(), mpmod::flines, mpmod::fvalue, mpmod::globalcorrections, mpmod::globalmatd, mpmod::globalparameter, mpmod::globalparcopy, mpmod::globalparlabelindex, mpmod::globalparvartototal, mpmod::globalvector, mpmod::icalcm, mpmod::iitera, mpmod::imonit, mpmod::iscerr, mpmod::isubit, mpmod::iterat, mpmod::lbmnrs, mpmod::lcalcm, mpmod::lhuber, loopn(), mpmod::lprecm, mpmod::lsearch, mpmod::lsinfo, mpmod::lunlog, mpmod::lunmon, mpmod::matconsblocks, mpmod::matparblockoffsets, mpmod::matrit, mpmod::matsto, mpmod::mbandw, mchdec(), mdiags(), mend(), mpmod::metsol, minver(), mpmod::mitera, mminrs(), mminrsqlp(), mpmod::mnrsel, monend(), monini(), mpmod::monpg1, mpmod::monpg2, monres(), mpmod::mrmode, mpmod::mrtcnd, mpmod::nagb, mpmod::nalow, mpmod::ncgbe, mpmod::ndefec, mpmod::ndfsum, mpmod::nfgb, mpmod::nfilw, mpmod::nloopn, mpmod::nmiss1, mpmod::nofeas, mpmod::npblck, mpmod::npresg, mpmod::nrderr, mpmod::nrecal, mpmod::nregul, mpmod::nrejec, mpmod::nvgb, peend(), ploopa(), ploopb(), ploopc(), ploopd(), prtglo(), ptldef(), ptline(), ptlprt(), qlssq(), mpmod::regpre, mpmod::regula, mpmod::rstart, solglo(), solgloqlp(), mpmod::stepl, mpmod::sumndf, mpmod::times, mpmod::vecparblockoffsets, mpmod::wolfc2, mpmod::workspacelinesearch, and zdiags().

Referenced by mptwo().

◆ zdiags()

subroutine zdiags

Covariance matrix for diagonalization (,correction of eigenvectors).

Definition at line 7843 of file pede.f90.

References devinv(), mpmod::globalmatd, mpmod::nfgb, mpmod::nvgb, qlmlq(), mpmod::workspaceeigenvalues, and mpmod::workspaceeigenvectors.

Referenced by xloopn().