KnowledgeBase Archive

An Archive of Early Microsoft KnowledgeBase Articles

View on GitHub

Q106731: PRB: Large BLOCK DATA Subprogram Compiles Slowly

Article: Q106731
Product(s): Microsoft Fortran Compiler
Version(s): 1.0,1.0a,4.0
Operating System(s): 
Keyword(s): kbLangFortran
Last Modified: 02-NOV-1999

-------------------------------------------------------------------------------
The information in this article applies to:

- Microsoft FORTRAN PowerStation for MS-DOS, versions 1.0, 1.0a 
- Microsoft Fortran Powerstation 32 for Windows NT, versions 1.0, 4.0 
-------------------------------------------------------------------------------

SYMPTOMS
========

Compilation of a large BLOCK DATA subprogram can take a very long time using the
FORTRAN PowerStation. In tests, a BLOCK DATA subprogram with 32,000 DATA
statements took approximately 45 minutes to compile on a 50 megahertz (MHz) 486
machine.

CAUSE
=====

This slow compilation is caused by initializing a large array in a BLOCK DATA
subprogram with several DATA statements. One DATA statement for each array
element produces the slowest compile. The array elements must also be
initialized in the order that they appear in the array for the slow compilation
to occur.

FORTRAN PowerStation uses the COFF object file format. Because COFF requires that
every element of an array be enumerated exactly once, all the data records
(which are generated for each DATA statement in the BLOCK DATA) are sorted. The
sorting algorithm is least efficient when the records are already sorted and
most efficient when they are in reverse order.

RESOLUTION
==========

If array elements are initialized in the reverse order that they appear in the
array, compilation time will be significantly reduced (for example, reduced from
45 minutes to 3 minutes). Reducing the number of DATA statements by initializing
thousands of array elements in a single DATA statement also reduces compile
time.

MORE INFORMATION
================

The following sample BLOCK DATA subprograms illustrate the problem and one
possible resolution:

Sample Code #1
--------------

  C This sample illustrates the problem by initializing a large array,
  C one element at a time, in the order the elements appear in the array.
        BLOCK DATA INITME
        COMMON /BIGCOM/ ARRAY(40000)
        DATA ARRAY(1)/1.5/ 
        DATA ARRAY(2)/7.4/ 
        .
        .                     ! 39996 DATA statements go here
        .
        DATA ARRAY(39999)/0.4/ 
        DATA ARRAY(40000)/12.0/ 
        END

Sample Code #2
--------------

  C This sample initializes the array elements in reverse order to
  C alleviate the problem.
        BLOCK DATA INITME
        COMMON /BIGCOM/ ARRAY(40000)
        DATA ARRAY(40000)/12.0/ 
        DATA ARRAY(39999)/0.4/ 
        .
        .                      ! 39996 DATA statements go here
        .
        DATA ARRAY(2)/7.4/ 
        DATA ARRAY(1)/1.5/ 
        END

Additional query words: 1.00 4.00 hang

======================================================================
Keywords          : kbLangFortran 
Technology        : kbAudDeveloper kbFortranSearch kbZNotKeyword2 kbZNotKeyword3 kbFORTRANPower32100NT kbFORTRANPower32400NT kbFORTRANPower100DOS kbFORTRANPower100aDOS
Version           : :1.0,1.0a,4.0

=============================================================================

THE INFORMATION PROVIDED IN THE MICROSOFT KNOWLEDGE BASE IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND. MICROSOFT DISCLAIMS ALL WARRANTIES, EITHER EXPRESS OR IMPLIED, INCLUDING THE WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT SHALL MICROSOFT CORPORATION OR ITS SUPPLIERS BE LIABLE FOR ANY DAMAGES WHATSOEVER INCLUDING DIRECT, INDIRECT, INCIDENTAL, CONSEQUENTIAL, LOSS OF BUSINESS PROFITS OR SPECIAL DAMAGES, EVEN IF MICROSOFT CORPORATION OR ITS SUPPLIERS HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. SOME STATES DO NOT ALLOW THE EXCLUSION OR LIMITATION OF LIABILITY FOR CONSEQUENTIAL OR INCIDENTAL DAMAGES SO THE FOREGOING LIMITATION MAY NOT APPLY.