|
|
 Up |
|
|
  |
Author: Steve LionelSteve Lionel Date: Jul 22, 2008 10:50
On Tue, 22 Jul 2008 17:28:50 +0200, fred wrote:
> !!! intel compilo
> open(unit=20, file='input.sep', form='binary')
> open(unit=21, file='a.sep', form='binary')
>
> !!! GNU compilo
>!!$ open(unit=20, file='input.sep', form='unformatted', access='stream')
>!!$ open(unit=21, file='a.sep', form='unformatted', access='stream')
Intel Fortran supports access='stream' - unless you are using a version
several years old.
--
Steve Lionel
Developer Products Division
Intel Corporation
Nashua, NH
For email address, replace "invalid" with "com"
|
| Show full article (0.81Kb) |
|
| | 10 Comments |
|
|
Author: Gary ScottGary Scott Date: Jul 23, 2008 04:51
fred wrote:
> fred a écrit :
>
>
>>Steve Lionel a écrit :
>>
>>
>>>Intel Fortran supports access='stream' - unless you are using a version
>>>several years old.
>
> Steve,
>
> I use intel fortran 9.1.045.
>
> marsu[pts/3]:~/tmp/{5}/> ifort essai.f90 -o essai
> fortcom: Error: essai.f90, line 12: Not a valid value for the char-expr in this connect-spec. ['stream']
> open(unit=20, file='input.sep', form='binary', access='stream')
> --------------------------------------------------------^
> fortcom: Error: essai.f90, line 13: Not a valid value for the char-expr in this connect-spec. ['stream']
> open(unit=21, file='a.sep', form='binary', access='stream') ...
|
| Show full article (1.21Kb) |
|
| | no comments |
|
|
Author: Paul van DelstPaul van Delst Date: Jul 23, 2008 06:18
Gary Scott wrote:
> fred wrote:
>> fred a écrit :
>>
>>
>>> Steve Lionel a écrit :
>>>
>>>
>>>> Intel Fortran supports access='stream' - unless you are using a version
>>>> several years old.
>>
>> Steve,
>>
>> I use intel fortran 9.1.045.
>>
>> marsu[pts/3]:~/tmp/{5}/> ifort essai.f90 -o essai
>> fortcom: Error: essai.f90, line 12: Not a valid value for the
>> char-expr in this connect-spec. ['stream']
>> open(unit=20, file='input.sep', form='binary', access='stream')
>> --------------------------------------------------------^ ...
|
| Show full article (1.45Kb) |
| no comments |
|
|
Author: Steve LionelSteve Lionel Date: Jul 23, 2008 06:46
On Wed, 23 Jul 2008 09:18:05 -0400, Paul van Delst noaa.gov>
wrote:
>This is off-topic (and a bit of a peeve/rant), but where I work I see again and again the
>incredulity of people when they realise that a 1-2yo old compiler needs upgrading. In fact
>the very idea of compiler upgrades seems completely foreign to some folks.
For many people, compilers are something that "if ain't broke, don't fix it".
If you're not having problems, then there's rarely a compelling reason to
upgrade a compiler. But if you ARE having problems, or are looking for a new
feature, then you may well find that an upgrade is required.
I might say that this problem is confined to commercial products such as
ifort, but no, we often see in this group people writing in with complaints
about old versions of gfortran and g95 (or even g77!)
All I can ask is that when asking a question, PLEASE identify the exact
compiler (with specific compiler version, OS version and CPU type) to make it
easier to find the answer.
|
| Show full article (1.38Kb) |
| no comments |
|
|
Author: James Van BuskirkJames Van Buskirk Date: Jul 23, 2008 08:58
> fred wrote:
>> fortcom: Error: essai.f90, line 13: Not a valid value for the char-expr
>> in this connect-spec. ['stream']
>> open(unit=21, file='a.sep', form='binary', access='stream')
>> ----------------------------------------------------^
> That means you are using an old version of Intel Fortran
Do even new versions of ifort support the extension form='binary' in
conjunction with the f03 access='stream' ? At best the two are
redundant and at worst conflicting in that access='stream' implies
file positioning primitives and form='binary' does not, at least not
with the same syntax as f03 employs.
--
write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
6.0134700243160014d-154/),(/'x'/)); end
|
| |
| no comments |
|
|
Author: glen herrmannsfeldtglen herrmannsfeldt Date: Jul 23, 2008 11:04
Steve Lionel wrote:
(snip)
> For many people, compilers are something that "if ain't broke, don't fix it".
> If you're not having problems, then there's rarely a compelling reason to
> upgrade a compiler. But if you ARE having problems, or are looking for a new
> feature, then you may well find that an upgrade is required.
> I might say that this problem is confined to commercial products such as
> ifort, but no, we often see in this group people writing in with complaints
> about old versions of gfortran and g95 (or even g77!)
Maybe more often with non-commercial products. With commercial products,
one might feel that, having paid for it, the company owes them a properly
working product. The work of installing the new version is part of
getting what one paid for.
-- glen
|
| |
| no comments |
|
|
Author: GaryScottGaryScott Date: Jul 23, 2008 10:11
On Jul 23, 1:04 pm, glen herrmannsfeldt ugcs.caltech.edu> wrote:> Steve Lionel wrote:
>
> (snip)
>
>> For many people, compilers are something that "if ain't broke, don't fix it".
>> If you're not having problems, then there's rarely a compelling reason to
>> upgrade a compiler. But if you ARE having problems, or are looking for a new
>> feature, then you may well find that an upgrade is required.
>> I might say that this problem is confined to commercial products such as
>> ifort, but no, we often see in this group people writing in with complaints
>> about old versions of gfortran and g95 (or even g77!)
>
> Maybe more often with non-commercial products. With commercial products,
> one might feel that, having paid for it, the company owes them a properly
> working product....
|
| Show full article (0.98Kb) |
| no comments |
|
|
Author: glen herrmannsfeldtglen herrmannsfeldt Date: Jul 23, 2008 11:29
GaryScott wrote:
(snip, I wrote)
>>Maybe more often with non-commercial products. With commercial products,
>>one might feel that, having paid for it, the company owes them a properly
>>working product. The work of installing the new version is part of
>>getting what one paid for.
> Usually fairly painless with most commercial products (install
> wizards, etc.).
Usually, but you never know. If it fails, how easy is it to
get back to the previous version? What else might accidentally
break? (Replacing DLLs that might be used by others?)
Then there are the funny ones. I have seen installers, even fairly
new ones, fail after improperly determining available disk space
or virtual memory. A system with over 2GB virtual memory available,
computed in signed integers, comes out negative. There is a
similar failure in the disk space calculation.
|
| Show full article (1.19Kb) |
| no comments |
|
|
Author: Steve LionelSteve Lionel Date: Jul 23, 2008 12:40
On Wed, 23 Jul 2008 09:58:04 -0600, "James Van Buskirk"
comcast.net> wrote:
>Do even new versions of ifort support the extension form='binary' in
>conjunction with the f03 access='stream' ? At best the two are
>redundant and at worst conflicting in that access='stream' implies
>file positioning primitives and form='binary' does not, at least not
>with the same syntax as f03 employs.
Thanks for picking up on that. No, form='binary' should not be used with
access='stream'. One typically wants form='unformatted'.
--
Steve Lionel
Developer Products Division
Intel Corporation
Nashua, NH
For email address, replace "invalid" with "com"
|
| Show full article (0.89Kb) |
| no comments |
|
|
|
|
|
Author: Colin Paul GlosterColin Paul Gloster Date: Jul 24, 2008 07:54
On Wed, 23 Jul 2008, Glen Herrmannsfeldt wrote:
|-----------------------------------------------------------------------------|
|"GaryScott wrote: |
|[..] ...
|
| Show full article (2.49Kb) |
| no comments |
|
|
|
|
