gibuu is hosted by Hepforge, IPPP Durham
GiBUU

TABLE OF CONTENTS

/PILIndex [ Modules ]

[ Top ] [ Modules ]

NAME

module PILIndex

PURPOSE

Provide a multi usable list connecting the (unique) particle number with the index number of some additional information in some other array.

While the list in the 'other array' is filled entry by entry, the list provided here is sorted by the index number in increasing order. This is done in order to speed up finding information about a particle from the 'other array' and whether information about this specific particle is actually stored in the array.

This implementation speeds up getting information at the expense of inserting new information.

INPUTS

(none)

NOTES

  • ...maybe one could change the access to PIL_nEntry0
  • "PIL" stands for "PartInfoList"

PILIndex/tIndexList [ Types ]

[ Top ] [ PILIndex ] [ Types ]

NAME

type tIndexList

PURPOSE

store all information of a IndexList

SOURCE

  type, public :: tIndexList
     integer :: nEntry = -999              ! number of used entries
     integer :: nHole  = 0                 ! number of deleted entries
     integer, allocatable :: PartNumber(:) ! array of particle numbers
     integer, allocatable :: Entry(:)      ! array of indices
  end type tIndexList

PILIndex/PIL_nEntry0 [ Global module-variables ]

[ Top ] [ PILIndex ] [ Global module-variables ]

PURPOSE

default size of arrays (at the beginning)

SOURCE

  integer,parameter :: PIL_nEntry0 = 1000

PILIndex/PILIndex_PUT [ Functions ]

[ Top ] [ PILIndex ] [ Functions ]

NAME

integer function PILIndex_PUT(List, number)

PURPOSE

find the place, where some information concerning a particle with "number" should be put into some info vector.

INPUTS

  • type(tIndexList) :: List -- The List to put the entry in
  • integer :: number -- The particle number to insert
  • character(*),intent(in),optional :: CallName -- Name of the List

OUTPUT

  • List is modified
  • return value gives the position where the info to be stored. If this value is negative, some allocation/reallocation of the infovector has to be done, then the info can be stored at the position given by the absolute value of the return value.

PILIndex/PILIndex_FIND [ Functions ]

[ Top ] [ PILIndex ] [ Functions ]

NAME

integer function PILIndex_FIND(List, number)

PURPOSE

find the place, where the "number" is in the list or where it should be stored

INPUTS

OUTPUT

if the return value R is > 0: here the entry "number" was found in the list

if the return value R <= 0: This means, that the entry at position (-R) is smaller than number, while the entry at position (-R)+1 is already larger.

NOTES

the return value R==0 is possible. This means of course, that already the first entry is larger than number. But you should never (!) try to access some entries at position R==0, because this undershoots the lower bound of the arrays.

PILIndex/PILIndex_Print [ Subroutines ]

[ Top ] [ PILIndex ] [ Subroutines ]

NAME

subroutine PILIndex_Print(List, file)

PURPOSE

just print the list to a file stream

INPUTS

  • type(tIndexList) :: List -- The List
  • integer :: file -- The file number

PILIndex/PILIndex_Allocate [ Subroutines ]

[ Top ] [ PILIndex ] [ Subroutines ]

NAME

subroutine PILIndex_Allocate(List)

PURPOSE

Do all allocation and reallocatiomn stuff connected with "List"

INPUTS

OUTPUT

  • if the arrays of "list" were not allocated before, they are allocated with some default size
  • if the arrays were allocated, they are reallocated with a size 1.3 times larger than the original size

NOTES

I think to remeber having read that the factor 1.3 is a good compromise between the effort of copying around all data and how often this happens. Of course this implies that one starts with some reasonable value. Maybe one could improve. (Kai)

For security one should insert here checks, whether the memory allocations failed and stop execution in these cases.

It would be a good idea to check whether the compiler knows the intrinsic MOVE_ALLOC subroutine and use this instead of copying the contents twice.

PILIndex/PILIndex_DeAllocate [ Subroutines ]

[ Top ] [ PILIndex ] [ Subroutines ]

NAME

subroutine PILIndex_DeAllocate(List)

PURPOSE

DeAllocate all memory connected wit this index list.

PILIndex/PILIndex_DELETE [ Functions ]

[ Top ] [ PILIndex ] [ Functions ]

NAME

integer function PILIndex_DELETE(List, number)

PURPOSE

find the place, where the "number" is in the list, delete it and return the index for others list to follow.

INPUTS

OUTPUT

if the return value R is > 0: here the entry "number" was found in the list

if the return value R <= 0: the entry was not found and not deleted

NOTES

Attention, this routine returns iH=List%Entry(i1), i.e. directy the index for ValueList(...). (Contrary to the _FIND routine, where you have to calculate the index by looking it up in %Entry(...). But here this entry is already deleted!)