IFC & g77 same code, .ne. result :)

I'm sorry for my bad english. th code: ================================ program dadi implicit none integer i,j,a dimension a(2:12) do i=1,6 do j=1,6 a(i+j) = a(i+j) + 1 enddo enddo do i=2,12 write(*,*) i,':',('#',j=1,a(i)) enddo stop end ============================== On intel compiler i have: ~/Documenti/Fortran$ ifc dadi.f -o dadi program DADI 18 Lines Compiled ~/Documenti/Fortran$ ./dadi 2 :# 3 :## 4 :### 5 :#### 6 :##### 7 :###### 8 :##### 9 :#### 10 :### 11 :## 12 :# it's OK :) but with g77 (g77 version 2.95.4 20011002 (Debian prerelease) (from FSF-g77 version 0.5.25 20010319 (prerelease)) ) I have: ~/Documenti/Fortran$ g77 dadi.f -o dadi ~/Documenti/Fortran$ ./dadi ############################################################################# ############################################################################## ############################################################################## ############################################################################## ############################################################################## ############################################################################## ############################################################################## ############################################################################## ############################################################################## ############################################################################## ############################################################################## ############################################################################## [KILL] infinire loop! ???? why?

On Wed, 25 Jun 2003 13:13:30, -=M=M=-@cippa.lippa.net <-=M=M=-@cippa.lippa.net> wrote in <_dhKa.118738$Ny5.3397650@twister2.libero.it>: > program dadi > implicit none > integer i,j,a > dimension a(2:12) > do i=1,6 > do j=1,6 > a(i+j) = a(i+j) + 1 ^^^^^^^^^^^^^^^^^^^ You never initialise the values in a()! There is no guarantee what values they might hold; some compilers/OSs _might_ initialise them to zero, but you can't depend on it. So you must implicitly do this yourself. On all compilers you can use a DO-loop; on modern compilers you can use the array syntax a=0 to initialise the whole array. > enddo > enddo > do i=2,12 > write(*,*) i,':',('#',j=1,a(i)) > enddo > stop > end >============================== > On intel compiler i have: > ~/Documenti/Fortran$ ifc dadi.f -o dadi > program DADI > 18 Lines Compiled > ~/Documenti/Fortran$ ./dadi > 2 :# > 3 :## > 4 :### > 5 :#### > 6 :##### > 7 :###### > 8 :##### > 9 :#### > 10 :### > 11 :## > 12 :# > it's OK :) > but with g77 (g77 version 2.95.4 20011002 (Debian prerelease) (from > FSF-g77 version 0.5.25 20010319 (prerelease)) ) > I have: > ~/Documenti/Fortran$ g77 dadi.f -o dadi > ~/Documenti/Fortran$ ./dadi > ############################################################################# > ############################################################################## > ############################################################################## > ############################################################################## > ############################################################################## > ############################################################################## > ############################################################################## > ############################################################################## > ############################################################################## > ############################################################################## > ############################################################################## > ############################################################################## > [KILL] > infinire loop! -- Ivan Reid, Electronic & Computer Engineering, ___ CMS Collaboration, Brunel University. Ivan.Reid@brunel.ac.uk Room 40-1-B12, CERN GSX600F, RG250WD. "You Porsche. Me pass!" DoD #484 JKLO# 003, 005 WP7# 3000 LC Unit #2368 (tinlc) UKMC#00009 BOTAFOT#16 UKRMMA#7 (Hon) KotPT -- "for stupidity above and beyond the call of duty".

-=M=M=-@cippa.lippa.net wrote: > > but with g77 (g77 version 2.95.4 20011002 (Debian prerelease) (from > FSF-g77 version 0.5.25 20010319 (prerelease)) ) > > I have: > ....snip... > [KILL] > infinire loop! > why? Other than it takes you 36 tries to fill 11 array elements :) I note that the g77 version info says something about "prerelease" -- I guess it's got a bug...

-=M=M=-@cippa.lippa.net wrote: > th code: > ================================ > program dadi > implicit none > integer i,j,a > dimension a(2:12) > > do i=1,6 > do j=1,6 > a(i+j) = a(i+j) + 1 > enddo > enddo > do i=2,12 > write(*,*) i,':',('#',j=1,a(i)) > enddo > stop > end > ============================== > On intel compiler i have: > ~/Documenti/Fortran$ ifc dadi.f -o dadi > ... > it's OK :) > but with g77 (g77 version 2.95.4 20011002 (Debian prerelease) (from > FSF-g77 version 0.5.25 20010319 (prerelease)) ) > ... > infinire loop! > why? You didn't initialize array a to 0. Unless you do so, the array starts out undefined. Usually this means that the contents are whatever left-over stuff is sitting in the memory. IFC is apparently presetting the array contents to 0, but that behavior is not required by the Fortran standard and is compiler-specific. On the other hand, g77 does no presetting, and the output reflects the random contents of a. Both compilers are acting compatibly with the standard. -- John Wingate Learning facts takes valuable time that johnww@worldpath.net could be better spent developing biases. --Richard Maine

Jugoslav Dujic wrote: > > Duane Bozarth wrote: > | -=M=M=-@cippa.lippa.net wrote: > || > || but with g77 (g77 version 2.95.4 20011002 (Debian prerelease) (from > || FSF-g77 version 0.5.25 20010319 (prerelease)) ) > || > || I have: > || > | ...snip... > || [KILL] > || infinire loop! > || why? > | > | Other than it takes you 36 tries to fill 11 array elements :) I note > | that the g77 version info says something about "prerelease" -- I guess > | it's got a bug... > > Do we have the "Blind spot of the year" award ;-))) ? Probably--I shortly after posting realized the problem and tried to withdraw the post, but it apparently had already gotten out before I was able to cancel it--(he says with face of red) :(

-=M=M=-@cippa.lippa.net wrote: > but with g77 (g77 version 2.95.4 20011002 (Debian prerelease) (from > FSF-g77 version 0.5.25 20010319 (prerelease)) ) > ############################################################################## > [KILL] > infinire loop! > > ???? > > why? Well, I hate to say this about my favourite GNU/Linux distribution, but producing Debian GCC 2.95.4 from a prerelease was a mistake. There are several bug reports in our bug database (http://gcc.gnu.org) against this compiler. -- Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290 Saturnushof 14, 3738 XG Maartensdijk, The Netherlands Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html GNU Fortran 95: http://gcc-g95.sourceforge.net/ (under construction)

Toon Moene wrote: > -=M=M=-@cippa.lippa.net wrote: >> [KILL] >> infinire loop! >> >> ???? >> why? > > > Well, I hate to say this about my favourite GNU/Linux distribution, but > producing Debian GCC 2.95.4 from a prerelease was a mistake. There are > several bug reports in our bug database (http://gcc.gnu.org) against > this compiler. Notwithstanding the truthfullness of this statement, the other explanation for the behaviour you see that's circulating on this newsgroup is better. [ open mouth, insert foot :-( ] -- Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290 Saturnushof 14, 3738 XG Maartensdijk, The Netherlands Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html GNU Fortran 95: http://gcc-g95.sourceforge.net/ (under construction)

On Wed, 25 Jun 2003 21:53:41 +0200, Toon Moene <toon@moene.indiv.nluug.nl> wrote in <3EF9FDC5.4060708@moene.indiv.nluug.nl>: > Toon Moene wrote: >> Well, I hate to say this about my favourite GNU/Linux distribution, but >> producing Debian GCC 2.95.4 from a prerelease was a mistake. There are >> several bug reports in our bug database (http://gcc.gnu.org) against >> this compiler. > Notwithstanding the truthfullness of this statement, the other > explanation for the behaviour you see that's circulating on this > newsgroup is better. > [ open mouth, insert foot :-( ] "Success is 99% failure." -- Soichiro Honda :-) -- Ivan Reid, Electronic & Computer Engineering, ___ CMS Collaboration, Brunel University. Ivan.Reid@brunel.ac.uk Room 40-1-B12, CERN

Similiar Articles:

problem in interface - comp.lang.fortran
this is because, this lm_a is from an old code, used ... Libraries compiled with G77 probably will not mix ... month of the same > is likely to give the same results. ....

Where did Fortran go? - comp.lang.fortran
The benchmark results are at https://userpage ... area of computational software, looking at ne ... can widely mix Fortran 77 and Fortran 90 code - even in the same ...

g77 - LinuxCommand.org: Learn the Linux command line. Write shell ...
... IF (J .NE. I) PRINT *, ’Prefix ... some other language for that same name. That is, getting code produced by g77 to ... compiler crashes, that might result from g77 ...

Using and Porting GNU Fortran - Known Causes of Trouble with GNU ...
... it is important that compilers treat code so modified in the same way so that the results of such programs are the same. g77 ... Use of .EQ. and .NE. on LOGICAL operands is ...

7/27/2012 7:19:05 PM