Scotch Brand 5.1.10 User Manual

Page 39

Advertising
background image

When the geometry of the graph is available, this mapping file may be
processed by program gout to display the vertex separators and super-
variable amalgamations that have been computed.

-ostrat

Apply ordering strategy strat. The format of ordering strategies is defined
in section 7.3.4.

-toutput tree file

Write to output tree file the structure of the separator tree. The data
that is written resembles much the one of a mapping file: after a first
line that contains the number of lines to follow, there are that many lines
of mapping pairs, which associate an integer number with every graph
vertex index. This integer number is the number of the column block
which is the parent of the column block to which the vertex belongs,
or −1 if the column block to which the vertex belongs is a root of the
separator tree (there can be several roots, if the graph is disconnected).
Combined to the column block mapping data produced by option -m, the
tree structure allows one to rebuild the separator tree.

-V

Print the program version and copyright.

-vverb

Set verbose mode to verb, which may contain several of the following
switches.

s

Strategy information. This parameter displays the ordering strategy
which will be used by gord.

t

Timing information.

6.3.11

gotst

Synopsis

gotst

[input graph file [input ordering file [output data file]]] options

Description

The program gotst is the ordering tester. It gives some statistics on orderings,
including the number of non-zeros and the operation count of the factored
matrix, as well as statistics regarding the elimination tree. Since it performs
the factorization of the reordered matrix, it can take a very long time and
consume a large amount of memory when applied to large graphs.
The first two statistics lines deal with the elimination tree. The first one
displays the number of leaves, while the second shows the minimum height of
the tree (that is, the length of the shortest path from any leaf to the –or a–
root node), its maximum height, its average height, and the variance of the
heights with respect to the average. The third line displays the number of non-
zero terms in the factored matrix, the amount of index data that is necessary
to maintain the block structure of the factored matrix, and the number of
operations required to factor the matrix by means of Cholesky factorization.

Options

-h

Display the program synopsis.

-V

Print the program version and copyright.

39

Advertising