[Bug fortran/45175] New: In gfortran, the calculation of small differences seems to depend on optimization flags

[Bug fortran/45175] New: In gfortran, the calculation of small differences seems to depend on optimization flags

[CycleTimeChart at yahoo dot com]

Just a little background:

I wrote a program that does high accuracy factorials.  It saves a given
factorial for later use as a double precision real between 1 and 2 and an
integer power of 2.  It does this by calculating the factorial to a practically
unlimited number of bits, rounding it to 53 bits only as the final step,
thereby avoiding the cumulative round off errors that occur.

I wanted to compare the accuracy of this method to a straight-forward way of
calculating factorials.  So I also scaled factorials calculated in a
straight-forward way the same way as above, that is as a double precision real
between 1 and 2 and an integer power of 2.

In the comparison, the integer power of 2 can be ignored, as it is practically
always the same.  As to the real portion, the difference between the two reals
should always be an integer multiple of 2**(-52).  The crux of the bug is that
when I compiled using the flag -O, this was true.  Using -O2, it was not. 
Stated another way, the calculation of small differences seems to depend on
optimization flags.


-- 
           Summary: In gfortran, the calculation of small differences seems
                    to depend on optimization flags
           Product: gcc
           Version: 4.5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: CycleTimeChart at yahoo dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45175

[Bug fortran/45175] In gfortran, the calculation of small differences seems to depend on optimization flags

[CycleTimeChart at yahoo dot com]

------- Comment #1 from CycleTimeChart at yahoo dot com  2010-08-04 02:05 -------
Created an attachment (id=21388)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=21388&action=view)
Main program


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45175

[Bug fortran/45175] In gfortran, the calculation of small differences seems to depend on optimization flags

[CycleTimeChart at yahoo dot com]

------- Comment #3 from CycleTimeChart at yahoo dot com  2010-08-04 02:06 -------
Created an attachment (id=21390)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=21390&action=view)
Implicit typecasting include file


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45175

[Bug fortran/45175] In gfortran, the calculation of small differences seems to depend on optimization flags

[CycleTimeChart at yahoo dot com]

------- Comment #2 from CycleTimeChart at yahoo dot com  2010-08-04 02:05 -------
Created an attachment (id=21389)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=21389&action=view)
Support routines


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45175

[Bug fortran/45175] In gfortran, the calculation of small differences seems to depend on optimization flags

[CycleTimeChart at yahoo dot com]

------- Comment #4 from CycleTimeChart at yahoo dot com  2010-08-04 02:22 -------
Created an attachment (id=21391)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=21391&action=view)
Compilation batch file


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45175

[Bug fortran/45175] In gfortran, the calculation of small differences seems to depend on optimization flags

[dominiq at lps dot ens dot fr]

------- Comment #5 from dominiq at lps dot ens dot fr  2010-08-04 02:38 -------
You do not say on which platform you did the tests (gfortran -v). From

C:\FI\Implicit4.inc

I assume that you are using some *86* processor under Windows. 

This is not a gfortran bug, but a duplicate of an infamous pr I cannot find the
number. You should try -ffloat-store or -mfpmath=sse and have a look at the gcc
manual

http://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html#Optimize-Options

under -ffloat-store and -fexcess-precision=style.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45175

[Bug fortran/45175] In gfortran, the calculation of small differences seems to depend on optimization flags

[jvdelisle at gcc dot gnu dot org]

------- Comment #6 from jvdelisle at gcc dot gnu dot org  2010-08-04 02:42 -------
This is a duplicate of PR 323

See  http://gcc.gnu.org/bugzilla/show_bug.cgi?id=323

Just use -ffloat-store.



*** This bug has been marked as a duplicate of 323 ***


-- 

jvdelisle at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |DUPLICATE


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45175