welcome: please sign in


book: Create a structured PDF document with headings, chapters, etc.
webpage: Specifies that the HTML sources are unstructured (plain web pages.) A page break is inserted between each file or URL in the output.
continuous: Specifies that the HTML sources are unstructured (plain web pages.) No page breaks are inserted between each file or URL in the output.
Title of the document for the front page.
Extract the first heading of the document and use it as title. If checked the title field has no effect.
The title image or HTML page. These file has to be an attachments!
Specify document version to be displayed on the title page.
Intellectual property owner of this document.
Copyright notice for this document.
Information about who and when modified the document are applied at the end.


Specifies the output format.
Grayscale document  Title page
Compression :   JPEG big images 


User defined page size 
Choose one of the predefined standard sizes or select user defined.
Specifies the page size using a standard name or in points (no suffix or ##x##pt), inches (##x##in), centimeters (##x##cm), or millimeters (##x##mm).
Set the target browser width in pixels (400-1200). This determines the page scaling of images.
   2-Sided   Landscape
Specifies the margin size using points (no suffix or ##x##pt), inches (##x##in), centimeters (##x##cm), or millimeters (##x##mm). Keep empty for default value.
Sets the page header to use on body pages.
Sets the page footer to use on body pages.


Sets the number of levels in the table-of-contents. Empty for unlimited levels.
   Numbered headings Check to number all of the headings in the document.
Sets the title for the table-of-contents. Empty for default title.
Sets the page header to use on table-of-contents pages.
Sets the page footer to use on table-of-contents pages.


Enter the HTML color for the body (background).
Enter the image file for the body (background). These file has to be an attachments!
Enter the HTML color for the text.
Sets the color of links.
Enables generation of links in PDF files.


Set the default size of text.
Set the spacing between lines of text.
Choose the default typeface (font) of text.
Choose the default typeface (font) of headings.
Set the size of header and footer text.
Choose the font for header and footer text.
Change the encoding of the text in document.
Check to embed font in the output file.


Controls the initial viewing mode for the document.
Document: Displays only the docuemnt pages.
Outline: Display the table-of-contents outline as well as the document pages.
Full-screen: Displays pages on the whole screen; this mode is used primarily for presentations.
Controls the initial layout of document pages on the screen.
Single: Displays a single page at a time.
One column: Displays a single column of pages at a time.
Two column left/right: Display two columns of pages at a time; the first page is displayed in the left or right column as selected.
Choose the initial page that will be shown.


Check to number all of the headings in the document.
 Print   Modify
 Copy   Annotate
Specifies the document permissions.
Specifies the user password to restrict viewing permissions on this PDF document. Empty for no encryption.
Specifies the owner password to control who can change document permissions etc. If this field is left blank, a random 32-character password is generated so that no one can change the document.


Specify language to use for date and time format.
Shrink code blocks on page.
Show line numbers for code blocks.
Make spaces visable by dots (·) instead of white spaces.
Make line breaks visable by a extra character (¶) at the end.
Enable this feature if you searching for problems or intent to report a bug report


Version 2.4.2 (MoinMoin 1.9.9)

MoinMoin - Generate PDF document using HTMLDOC

This action script generate PDF documents from a Wiki site using
the HTMLDOC (http://www.htmldoc.org) software packages which has
to be preinstalled first.

Copy this script in your's MoinMoin action script plugin directory.

Thanks goes to Pascal Bauermeister who initiated the implementaion.
Lot of things changes since then but the idear using HTMLDOC is the
main concept of this implementation.

Please visit the homepage for further informations:

@copyright: (C) 2006 Pascal Bauermeister
@copyright: (C) 2006-2010 Raphael Bossek <raphael.bossek@solutions4linux.de>
@license: GNU GPL, see COPYING for details


location: Middleware / MpiStart / RequirementsAndDesign

MPI-Start requirements and design


The development of MPI-Start comes from the necessity of having a single interface to upper layers of middleware and users to run a MPI job in a heterogeneous environment like the grid.


With MPI-Start, new MPI implementations and new batch schedulers should be easily supported without modifications in the Grid middleware or the user job descriptions.

Apart from starting MPI jobs, MPI-Start should support "simple" file distribution for sites without a shared filesystem and provide support for the users to help manage their data and applications.


Design Goals

MPI-Start has the following design goals:

. The program must be able to run under any supported operating system
Modular and extensible
. New MPI implementations and batch systems should be easy to add.
. Must be independent of absolute path, to adapt to different site configurations and allow remote "injection" of mpi-start along with the job
"Remote" debugging features
. Must provide extensive debugging information for resolving problems.


MPI-Start is a set of scripts that ease the execution of MPI programs by using a unique and stable interface to the middleware. The scripts are written in a modular way: there is a core and around it there are different plug-ins as shown in following figure. Three main frameworks are defined in the MPI-Start architecture:


The hooks framework opens the possibility of customizing the behavior of MPI-Start. There are four classes of hooks: local-site, compiler, end-user, and file distribution. Local-site hooks allow the system administrator to set any configuration specific to the site that are not set automatically by MPI-Start. In order to obtain good performance and to assure that the binaries will fit the available resources, MPI jobs should be compiled with the local MPI implementation at each site. MPI-Start compiler support in the hooks framework detects the system architecture and sets the adequate compiler flags to build the binaries. Users can compile their application using the end-user hook, this is a shell script where the exact command line for building the application can be specified. Any input data preparation can be also done in that script.

One of the most important features in MPI-Start is the file distribution hook. Most MPI implementations need application binaries to exist in all the machines involved in the execution. This hook is executed once all the previous hooks have finished correctly and assures that all the needed files are available for the job at the different hosts. Prior to file distribution, a test detects the availability of a network file system. If this kind of file system is not found, MPI-Start will copy the files using the most appropriate method according to the site configuration. In the current implementation, the following methods are supported: OSC Mpiexec, ssh, mpi-mt (a MPI tool for copying files with MPI-IO) and copying the files to a shared area defined by the administrator.

The next figure shows the flow of execution of MPI-Start. The first step is the detection of the scheduler plug-in and the creation of a machinefile with the list of hosts in a default format. This is done automatically by checking the execution environment of the job.


The next step is the selection of the execution plug-in according to the value specified by the user. The plug-in will activate the MPI implementation and prepare the correct command line to start the job. MPI-Start is able to detect any special tools available in the site, such as OSC Mpiexec, that may be used for a better startup of the application. If any of the two first steps fail, the program will exit and dump the execution environment to aid in debugging the problem.

Once the environment is ready to execute the MPI type selected, the hooks framework is executed in the following order: compiler, local-site, end-user and file distribution. If any of the hooks is not found it will be skipped. The following step is the actual execution of the application using the appropriate command line created by the execution plug-in and potentially modified by the hooks. The last step is the post-run hooks that allow fetching results or output processing. The post-run hooks executed are local-site, end-user and file distribution.

MPI-Start is implemented in a way that is easy to use even if it is not installed with special privileges. It can be included with the input files of the job and be executed from any location of the file system. The use of shell scripts for its implementation gives to MPI-Start good portability across systems. It also includes exhaustive debugging features. With different levels of debugging verbosity, MPI-Start can give from a totally silent output to a complete trace of all the steps performed during its execution.