f



Intel fortran and fortran 77 standard!

Hi!

I am trying to recompile lapack 3.0 with Intel
fortran, but I am getting compiler errors in ilaenv.f
I tried to skip that file to see what happens, but
there are more errors. Can Lapack be compiled
with Intel compiler at all?

I was using CVF 6.1 by now and any hints
on how to avoid trouble are welcome.

Thanks!
Atmapuri.


0
4/26/2004 11:56:25 AM
comp.lang.fortran 11941 articles. 1 followers. Post Follow

9 Replies
1310 Views

Similar Articles

[PageSpeed] 25

"Atmapuri" <janez.makovsek@usa.net> wrote in message
news:QN6jc.1589$37.232466@news.siol.net...
> Hi!
>
> I am trying to recompile lapack 3.0 with Intel
> fortran, but I am getting compiler errors in ilaenv.f
> I tried to skip that file to see what happens, but
> there are more errors. Can Lapack be compiled
> with Intel compiler at all?
>
> I was using CVF 6.1 by now and any hints
> on how to avoid trouble are welcome.
>
I'm sorry, but I have not been able to guess what problems you have found.
Intel Fortran 7.1 or 8.0 will give you a number of warnings about the
obsolescent syntax.  The version I have in front of me is not suppressing
those warnings when I ask for that.  I'm sure you are aware there are
multiple versions of Intel Fortran and lapack, and many postings about them
accessible by Google, but you ask us to guess which you have, and which you
have chosen to ignore.  Did you not notice that pre-compiled versions of
lapack built with Intel compilers are available?


0
tprince8714 (295)
4/26/2004 12:34:08 PM
Hi!

> I'm sorry, but I have not been able to guess what problems you have found.
> Intel Fortran 7.1 or 8.0 will give you a number of warnings about the
> obsolescent syntax.  The version I have in front of me is not suppressing
> those warnings when I ask for that.  I'm sure you are aware there are
> multiple versions of Intel Fortran and lapack, and many postings about
them
> accessible by Google, but you ask us to guess which you have, and which
you
> have chosen to ignore.  Did you not notice that pre-compiled versions of
> lapack built with Intel compilers are available?

Lapack 3.0, Intel Fortran 7.1 running from VS6.0, Opened working CVF 6.1 dll
project
tried to build all:

 ..\..\..\Flapack\src\ilaenv.f: Error 531 : The variable NAME cannot be a
value dummy argument use CHARACTER(*) instead.
...\..\..\Flapack\src\ilaenv.f: Error 531 : The variable OPTS cannot be a
value dummy argument use CHARACTER(*) instead.

I figured, if I could get this fixed without modifying code, I could make my
other code work work also, within spending years. I have to start
somewhere.

Thanks!
Atmapuri


0
4/26/2004 1:09:09 PM
Atmapuri wrote:
| Hi!
| 
|| I'm sorry, but I have not been able to guess what problems you have found.
|| Intel Fortran 7.1 or 8.0 will give you a number of warnings about the
|| obsolescent syntax.  The version I have in front of me is not suppressing
|| those warnings when I ask for that.  I'm sure you are aware there are
|| multiple versions of Intel Fortran and lapack, and many postings about them
|| accessible by Google, but you ask us to guess which you have, and which you
|| have chosen to ignore.  Did you not notice that pre-compiled versions of
|| lapack built with Intel compilers are available?
| 
| Lapack 3.0, Intel Fortran 7.1 running from VS6.0, Opened working CVF 6.1 dll
| project
| tried to build all:
| 
|  ..\..\..\Flapack\src\ilaenv.f: Error 531 : The variable NAME cannot be a
| value dummy argument use CHARACTER(*) instead.
| ..\..\..\Flapack\src\ilaenv.f: Error 531 : The variable OPTS cannot be a
| value dummy argument use CHARACTER(*) instead.
| 
| I figured, if I could get this fixed without modifying code, I could make my
| other code work work also, within spending years. I have to start
| somewhere.

Well, have you tried replacing CHARACTER*(*) with CHARACTER(*)?

-- 
 Jugoslav
___________
www.geocities.com/jdujic

Please reply to the newsgroup.
You can find my real e-mail on my home page above.

0
jdujic (694)
4/26/2004 1:16:19 PM
Hi!

> |  ..\..\..\Flapack\src\ilaenv.f: Error 531 : The variable NAME cannot be
a
> | value dummy argument use CHARACTER(*) instead.
> | ..\..\..\Flapack\src\ilaenv.f: Error 531 : The variable OPTS cannot be a
> | value dummy argument use CHARACTER(*) instead.
> |
> | I figured, if I could get this fixed without modifying code, I could
make my
> | other code work work also, within spending years. I have to start
> | somewhere.
>
> Well, have you tried replacing CHARACTER*(*) with CHARACTER(*)?

Of course. Did not work, same error. I double checked I fixed the
the right file. I also made another error to see, if it will be picked up
by the compiler. (Besides.
The last I thing I want to do is spend time fixing lapack source. )

I hoped there is a general setting somewhere that I am missing.
For most old fortran code that I am compiling, I dont even know
how to fix it without breaking it.

Intel compiler settings (custom build):
ifl /c /O3 /W1 /GB /QaxW /4I4 /4Na /Qipo /4L72 /Quppercase /Gz /Qpc64 /4Nb
/MT /Qdps /Fo$(OutDir)\ $(InputPath)

Output:
$(OutDir)\$(InputName).obj

Thanks!
Atmapuri.


0
4/26/2004 1:38:17 PM
Hi!

> |  ..\..\..\Flapack\src\ilaenv.f: Error 531 : The variable NAME cannot be
a
> | value dummy argument use CHARACTER(*) instead.

One more thing. There is no error line number next to the error?
Is that usual with IF 7.1?

Thanks!
Atmapuri


0
4/26/2004 1:47:34 PM
Atmapuri wrote:
> 
> I am trying to recompile lapack 3.0 with Intel
> fortran, but I am getting compiler errors in ilaenv.f
> I tried to skip that file to see what happens, but
> there are more errors. Can Lapack be compiled
> with Intel compiler at all?
> 
> I was using CVF 6.1 by now and any hints
> on how to avoid trouble are welcome.

Upgrade to the latest CVF and then stick with it for another couple
of years, or until Intravenously Fed Contraption is disconnected from
the development tubes.

   ftp://ftp.compaq.com/pub/products/fortran/vf/CVF-66-66C.exe

A Fortran compiler that fails on a piece of code as stable as *lapack*
should not even be offered to the public, period!

0
kia (104)
4/28/2004 8:20:09 PM
On Wed, 28 Apr 2004 20:20:09 GMT, kia <kia@sparrow.com> wrote:
> Atmapuri wrote:
>> 
>> I am trying to recompile lapack 3.0 with Intel
>> fortran, but I am getting compiler errors in ilaenv.f
>> I tried to skip that file to see what happens, but
>> there are more errors. Can Lapack be compiled
>> with Intel compiler at all?
>> 
>> I was using CVF 6.1 by now and any hints
>> on how to avoid trouble are welcome.
> 
> Upgrade to the latest CVF and then stick with it for another couple
> of years, or until Intravenously Fed Contraption is disconnected from
> the development tubes.
> 
>    ftp://ftp.compaq.com/pub/products/fortran/vf/CVF-66-66C.exe
> 
> A Fortran compiler that fails on a piece of code as stable as *lapack*
> should not even be offered to the public, period!

Is the compiler actually failing, or LAPACK assuming something
not really right about floating point?  


0
4/29/2004 12:15:10 AM
On Thu, 29 Apr 2004 00:15:10 +0000 (UTC), Dr Chaos
<mbkennelSPAMBEGONE@NOSPAMyahoo.com> wrote:

>> A Fortran compiler that fails on a piece of code as stable as *lapack*
>> should not even be offered to the public, period!
>
>Is the compiler actually failing, or LAPACK assuming something
>not really right about floating point?  

Or is there a corrupted source being compiled?  I know that LAPACK is part of
the regular testing that Intel Fortran undergoes. There isn't clear evidence
here of a compiler problem.

Not that this should affect compiling LAPACK, but it is an older version of
Intel Fortran being discussed here.

Steve Lionel
Software Products Division
Intel Corporation
Nashua, NH

User communities for Intel Software Development Products
  http://softwareforums.intel.com/
Intel Fortran Support
  http://developer.intel.com/software/products/support/
0
Steve.Lionel (766)
4/29/2004 1:14:15 PM
Steve Lionel <Steve.Lionel@intel.com> writes:

> On Thu, 29 Apr 2004 00:15:10 +0000 (UTC), Dr Chaos
> <mbkennelSPAMBEGONE@NOSPAMyahoo.com> wrote:
> 
> >> A Fortran compiler that fails on a piece of code as stable as *lapack*
> >> should not even be offered to the public, period!
> >
> >Is the compiler actually failing, or LAPACK assuming something
> >not really right about floating point?  
> 
> Or is there a corrupted source being compiled?  I know that LAPACK is part of
> the regular testing that Intel Fortran undergoes. There isn't clear evidence
> here of a compiler problem.

I don't recall what evidence was presented, but another posssibility is
that it is a version of LAPACK for some different hardware.  It's been
a while, so my memory may be confusing it with something else, or things
might have changed, but I seem to recall that LAPACK "out of the box"
had one routine that you really *HAD* to customize because it came
defaulted for some hardware that doesn't even exist any more.  The
customization was easy (and could even be done portably with f90
intrinsics, but LAPACK long predates f90).  No compiler is going to
"fix" compiling with the wrong code.

-- 
Richard Maine                       |  Good judgment comes from experience;
email: my first.last at org.domain  |  experience comes from bad judgment.
org: nasa, domain: gov              |        -- Mark Twain
0
nospam47 (9748)
4/29/2004 3:53:34 PM
Reply: