JiscMail Logo
Email discussion lists for the UK Education and Research communities

Help for COMP-FORTRAN-90 Archives


COMP-FORTRAN-90 Archives

COMP-FORTRAN-90 Archives


COMP-FORTRAN-90@JISCMAIL.AC.UK


View:

Message:

[

First

|

Previous

|

Next

|

Last

]

By Topic:

[

First

|

Previous

|

Next

|

Last

]

By Author:

[

First

|

Previous

|

Next

|

Last

]

Font:

Proportional Font

LISTSERV Archives

LISTSERV Archives

COMP-FORTRAN-90 Home

COMP-FORTRAN-90 Home

COMP-FORTRAN-90  2000

COMP-FORTRAN-90 2000

Options

Subscribe or Unsubscribe

Subscribe or Unsubscribe

Log In

Log In

Get Password

Get Password

Subject:

strange allocation behavior

From:

Peter Shenkin <[log in to unmask]>

Reply-To:

Peter Shenkin <[log in to unmask]>

Date:

Wed, 26 Jul 2000 23:59:41 -0400 (EDT)

Content-Type:

TEXT/PLAIN

Parts/Attachments:

Parts/Attachments

TEXT/PLAIN (181 lines)

Hi,

I'm seeing strange allocation behavior (actually, deallocation
behavior) with Absoft Pro Fortran 6.0 on Linux.  I've just sent
email to their support line, but in the meantime I thought I'd ask
others to try my little test program on compilers that I've
not tried.  
 
I'm particularly interested in other f90 compilers for Linux, 
but would like to compile results for as many compilers/platforms
as possible (within reason :-) ).

Here are the results I obtained for Linux-x86 (Absoft Pro F90 6.0),
SGI, TRU64, HPUX, AIX and SunOS.  "Recent" (but not necessarily the
latest) compilers are used in each case.

The program included below goes through the following procedure:

	ALLOCATE int array of size 10000 ints
	DEALLOCATE the same array

	ALLOCATE the same array to size 9999 ints
	DEALLOCATE the same array

	...
	ALLOCATE the same array to size 1 int
	DEALLOCATE the same array

Thus, the sum of the ALLOCATE sizes is 10000*10001/2 = 50005000 ints.

Since each ALLOCATE beyond the first specifies a smaller size 
than the preceding one, the storage should be re-usable, and
the process should not grow significantly after the first allocation.

If no memory is reutilized at all, we expect the memory to grow
by a factor of 50005000/10000, or approx. 5000, over the course of
the process.  To rounding error (getting the process size from
"top" or "ps"), it indeed seems to be the case that memory is
not being reused when the code is compiled with Absoft Pro F90
6.0 / Linux-x86.

Below, please find results from Absoft on Linux, as well as from 
native compilers on SGI, AIX, TRU64, HP and SunOS.  Absoft is
alone in failing to reuse DEALLOCATEd storage.  A similar
C program under Linux-x86 (compiled with gcc) exhibits the expected
reuse of memory.

Colleagues report that at least one other Linux F90 does appear
to reuse memory, as expected.

Here are my results, followed by the program.  The SZ and RSS
values are typed into the program by hand, then reported at
the end.

---------------------------------------------------
absoft pro fortran 6.0, Linux x86
Starting array size, sum of allocations =      10000            50005000
                          When                  SZ                 RSS
            Before first alloc                 372                 372
             After first alloc                 384                 384
              After last alloc               38632               38632
        Over-allocation factor             3188.33             3188.33
---------------------------------------------------
sgi
Starting array size, sum of allocations =      10000            50005000
                          When                  SZ                 RSS
            Before first alloc                 277                 119
             After first alloc                 279                 123
              After last alloc                 279                 124
        Over-allocation factor                1.00                1.25
---------------------------------------------------
aix
Starting array size, sum of allocations =      10000            50005000
                          When                  SZ                 RSS
            Before first alloc                 106                 172
             After first alloc                 112                 184
              After last alloc                 144                 216
        Over-allocation factor                6.33                3.67
---------------------------------------------------
tru64
Starting array size, sum of allocations =      10000            50005000
                          When                  SZ                 RSS
            Before first alloc                 304                 304
             After first alloc                 320                 320
              After last alloc                 344                 344
        Over-allocation factor                2.50                2.50
---------------------------------------------------
hp
Starting array size, sum of allocations =      10000            50005000
                          When                  SZ                 RSS
            Before first alloc                  19                  19
             After first alloc                  23                  23
              After last alloc                  31                  31
        Over-allocation factor                3.00                3.00
---------------------------------------------------
sunos
Starting array size, sum of allocations =      10000            50005000
                          When                  SZ                 RSS
            Before first alloc               10784                1480
             After first alloc               10824                1496
              After last alloc               10824                1536
        Over-allocation factor                1.00                3.50
---------------------------------------------------

! start program
      PROGRAM TRY
      ! Tests out F90 allocation and deallocation
      IMPLICIT NONE
      INTEGER,PARAMETER :: STARTING_ARRAY_SIZE = 10001
      INTEGER,PARAMETER :: MAXITER = 10000
      INTEGER :: ITER, ARRAY_SIZE
      INTEGER,ALLOCATABLE :: IAR (:)
      INTEGER :: SIZES_BEFORE_ALLOC( 2 )
      INTEGER :: SIZES_AFTER_FIRST_ALLOC( 2 )
      INTEGER :: SIZES_AFTER_LAST_ALLOC( 2 )
      INTEGER :: TOTAL_ALLOC = 0

      WRITE( 6, '(A)' )'Please use integers for all values entered.'
      DO ITER = 1, MAXITER

        !!! Prompt user for sizes from 'top' before and after 1st alloc: */
        IF( ITER .EQ. 1 )THEN
          WRITE( 6, '(A)', ADVANCE='NO' )
     &     'Before 1st alloc; enter SZ, RSS from ''top'' or ''ps'': '       
          READ( 5, * ) SIZES_BEFORE_ALLOC
        ELSE IF( ITER .EQ. 2 )THEN
          WRITE( 6, '(A)', ADVANCE='NO' )
     &     'After 1st alloc; enter SZ, RSS from ''top'' or ''ps'': '       
          READ( 5, * ) SIZES_AFTER_FIRST_ALLOC
        ENDIF

        ! Deallocate if the array is already allocated:
        IF( ITER .GT. 1 ) THEN
          DEALLOCATE( IAR )
        END IF

        ! Allocate the array:
        ARRAY_SIZE = STARTING_ARRAY_SIZE - ITER
        TOTAL_ALLOC = TOTAL_ALLOC + ARRAY_SIZE
        ALLOCATE( IAR(ARRAY_SIZE) ) 
      END DO

      WRITE( 6, '(A)', ADVANCE='NO' )
     & 'After last alloc; enter SZ, RSS from ''top'' or ''ps'': '       
      READ( 5, * ) SIZES_AFTER_LAST_ALLOC

      WRITE( 6, '(A,I10,I20)' )
     & 'Starting array size, sum of allocations = ',
     & STARTING_ARRAY_SIZE-1, TOTAL_ALLOC

      WRITE( 6, '(A30,2A20)' )'When', 'SZ', 'RSS'
      WRITE( 6, '(A30,2I20)' )
     & 'Before first alloc',
     &  SIZES_BEFORE_ALLOC
      WRITE( 6, '(A30,2I20)' )
     & 'After first alloc',
     &  SIZES_AFTER_FIRST_ALLOC
      WRITE( 6, '(A30,2I20)' )
     & 'After last alloc',
     &  SIZES_AFTER_LAST_ALLOC
      WRITE( 6, '(A30,2F20.2)' )
     & 'Over-allocation factor',
     & ( SIZES_AFTER_LAST_ALLOC(1) - SIZES_BEFORE_ALLOC(1) )
     & / DBLE( SIZES_AFTER_FIRST_ALLOC(1) - SIZES_BEFORE_ALLOC(1) ),
     & ( SIZES_AFTER_LAST_ALLOC(2) - SIZES_BEFORE_ALLOC(2) )
     & / DBLE( SIZES_AFTER_FIRST_ALLOC(2) - SIZES_BEFORE_ALLOC(2) )

      END
! end program

-- 
--
** Whether the playing field is level depends on the coordinate system. ***
********* Peter S. Shenkin; Schrodinger, Inc.; (201)433-2014 x111 *********
*********** [log in to unmask]; http://www.schrodinger.com ***********




%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

Top of Message | Previous Page | Permalink

JiscMail Tools


RSS Feeds and Sharing


Advanced Options


Archives

December 2023
February 2023
November 2022
September 2022
February 2022
January 2022
June 2021
November 2020
September 2020
June 2020
May 2020
April 2020
December 2019
October 2019
September 2019
March 2019
February 2019
January 2019
November 2018
October 2018
September 2018
August 2018
July 2018
May 2018
April 2018
March 2018
February 2018
January 2018
December 2017
November 2017
October 2017
September 2017
August 2017
July 2017
June 2017
May 2017
April 2017
March 2017
January 2017
December 2016
November 2016
October 2016
September 2016
August 2016
July 2016
June 2016
May 2016
April 2016
March 2016
February 2016
December 2015
November 2015
October 2015
September 2015
August 2015
June 2015
April 2015
March 2015
January 2015
December 2014
November 2014
October 2014
August 2014
July 2014
May 2014
April 2014
March 2014
February 2014
January 2014
December 2013
November 2013
July 2013
June 2013
May 2013
April 2013
February 2013
January 2013
December 2012
November 2012
October 2012
September 2012
August 2012
July 2012
June 2012
April 2012
March 2012
February 2012
January 2012
December 2011
November 2011
October 2011
September 2011
August 2011
July 2011
June 2011
May 2011
April 2011
March 2011
February 2011
January 2011
December 2010
November 2010
October 2010
August 2010
July 2010
June 2010
March 2010
February 2010
January 2010
December 2009
October 2009
August 2009
July 2009
June 2009
March 2009
February 2009
January 2009
December 2008
November 2008
October 2008
September 2008
August 2008
July 2008
June 2008
May 2008
April 2008
March 2008
February 2008
December 2007
November 2007
October 2007
September 2007
August 2007
July 2007
June 2007
May 2007
April 2007
March 2007
January 2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997


JiscMail is a Jisc service.

View our service policies at https://www.jiscmail.ac.uk/policyandsecurity/ and Jisc's privacy policy at https://www.jisc.ac.uk/website/privacy-notice

For help and support help@jisc.ac.uk

Secured by F-Secure Anti-Virus CataList Email List Search Powered by the LISTSERV Email List Manager