Fortran - A tutorail web site for Fortran 90/95 ?

Author

A tutorail web site for Fortran 90/95 ?

victor

2004-11-13, 8:55 pm

Hi Experts,

Can you tell me a tutorail web site for Fortran 90/95 ?

Thank you in advance.

Danguard

2004-11-13, 8:55 pm

In article <aec1ceb.0411131326.74d7851d@posting.google.com>, victor@in-
box.net says...

> Hi Experts,

Hi, I'm not an expert! :)
However:

> Can you tell me a tutorail web site for Fortran 90/95 ?

I was going to study FORTRAN77, but recent posts by helpful Fortran
programmers led me to study Fortran 90 (or 95).

I've found some tutorials on the Internet, using Google.

e.g.:

<http://www.cs.mtu.edu/~shene/COURSE...ES/fortran.html>

There are also some articles on Fortran from CERN
(I'm a CERN-fan!! :)

They are titled "Michael Metcalf's Fortran 90 CNL Articles", and I found
them here:

<http://wwwasdoc.web.cern.ch/wwwasdoc/f90.html>

Considering that CERN is a top scientific research institution in the
world, and that Fortran is used for scientific computations, I trust in
docs found on CERN website!

Moreover, I think that Google is a very good tool for searching!

> Thank you in advance.

Not at all

Ciao,
Dan

beliavsky@aol.com

2004-11-13, 8:55 pm

Danguard <danguard_robot@hotmail.com> wrote:
>In article <aec1ceb.0411131326.74d7851d@posting.google.com>, victor@in-
>box.net says...
>
>
>Hi, I'm not an expert! :)
>However:
>
>
>I was going to study FORTRAN77, but recent posts by helpful Fortran
>programmers led me to study Fortran 90 (or 95).
>
>I've found some tutorials on the Internet, using Google.

There are links to dozens of Fortran 90/95 tutorials at http://www.dmoz.org/Computers/Progr...tran_90_and_95/
If you find one not listed there, please submit the link to that site.

----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeed.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= 19 East/West-Coast Specialized Servers - Total Privacy via Encryption =---

Danguard

2004-11-14, 8:55 am

In article <41969153_1@127.0.0.1>, beliavsky@127.0.0.1 says...

> There are links to dozens of Fortran 90/95 tutorials at
> http://www.dmoz.org/Computers/Programming/Languages/
> Fortran/Tutorials/Fortran_90_and_95/

Very good! Thank you!

> If you find one not listed there, please submit the link to that site.

OK!

victor

2004-11-14, 8:56 pm

Thanks a lot for the information from both of you. I will try to find
at least one suitable for me :)

Cheers!

victor

Danguard <danguard_robot@hotmail.com> wrote in message news:<MPG.1c01581383aa86b3989a20@powernews.libero.it>...
> In article <41969153_1@127.0.0.1>, beliavsky@127.0.0.1 says...
>
>
> Very good! Thank you!
>
>
> OK!

Dr Ivan D. Reid

2004-11-15, 3:59 pm

On Sat, 13 Nov 2004 22:23:25 GMT, Danguard <danguard_robot@hotmail.com>
wrote in <MPG.1c00a7ca64aa76e4989a1d@powernews.libero.it>:

> Considering that CERN is a top scientific research institution in the
> world, and that Fortran is used for scientific computations, I trust in
> docs found on CERN website!

Guess what language has been all but phased out (actively, in at least
one case) at CERN, and which language is used to replace it.

--
Ivan Reid, Electronic & Computer Engineering, ___ CMS Collaboration,
Brunel University. Ivan.Reid@brunel.ac.uk Room 40-1-B12, CERN
KotPT -- "for stupidity above and beyond the call of duty".

Danguard

2004-11-15, 3:59 pm

In article <slrncph949.35g.Ivan.Reid@loki.brunel.ac.uk>,
Ivan.Reid@brunel.ac.uk says...

> Guess what language has been all but phased out (actively, in at least
> one case) at CERN, and which language is used to replace it.

Fortran 9x is replacing Fortran 77, right?

Or some very-high level language like Mathematica's is replacing
Fortran?

I'm curious...

Thank you.

Ciao
Dan

Dr Ivan D. Reid

2004-11-15, 8:56 pm

On Mon, 15 Nov 2004 15:26:30 GMT, Danguard <danguard_robot@hotmail.com>
wrote in <MPG.1c02e91022098d58989a24@powernews.libero.it>:
> In article <slrncph949.35g.Ivan.Reid@loki.brunel.ac.uk>,
> Ivan.Reid@brunel.ac.uk says...

[color=darkred]
> Fortran 9x is replacing Fortran 77, right?

Nope.

> Or some very-high level language like Mathematica's is replacing
> Fortran?

Nope.

> I'm curious...

All code (that I know about[1]) for CMS is being developed in C++.
CMSIM and GEANT3 have been replaced by C++ versions.
http://cmsdoc.cern.ch/cmsreco/

> Thank you.

[1] Apart from some tools I write myself, typically to produce PostScript
visualisations.

--
Ivan Reid, Electronic & Computer Engineering, ___ CMS Collaboration,
Brunel University. Ivan.Reid@brunel.ac.uk Room 40-1-B12, CERN
KotPT -- "for stupidity above and beyond the call of duty".

Danguard

2004-11-16, 3:57 pm

In article <slrncpifts.748.Ivan.Reid@loki.brunel.ac.uk>,
Ivan.Reid@brunel.ac.uk says...

> All code (that I know about[1]) for CMS is being developed in C++.
> CMSIM and GEANT3 have been replaced by C++ versions.

I don't know of CMS, but it don't understand *why* they're using C++
(?).

I think C++ is a very good language for some kind of software (e.g.: 3D
engines, GUIs, etc.) but I think Fortran 9x is superior to C++ for
numerical computing, in fact Fortran 9x seems to me be more "friendly"
for the programmer.
(...excluding container classes, that I'm searching for Fortran).

e.g. built-in array support of Fortran for arrays is very good feature,
while in C++ you need a separate custom matrix/array class library, with
overloaded operators, etc. and I don't know if this code is comparable
for efficiency to Fortran compiled code.

(I also read that supercomputers like Cray use Fortran as main
scientific numeric programming language...)

....Maybe this "CMS project" has specifications that C++ meets better
than Fortran does?

Dan

Michael Metcalf

2004-11-16, 3:57 pm

"> I don't know of CMS, but it don't understand *why* they're using C++
> (?).
>

I was CERN's representative on X3J3 from 1984 to 1991, so was part of the
story. Basically, it was assumed that F77 code would be migrated to f90 as
soon as suitable compilers became available, and, indeed, I spent a lot of
time testing early versions. However, there had, in the 80s, been a lot of
pressure from SLAC, orchestrated by Paul Kunz, to move to C and, when that
failed, to Objective C. That failed too as it was a proprietary product.
Meanwhile, C++ loomed on the horizon, and Kunz became a keen and competent
proselyte for that language. At CERN and elsewhere, I gave many tutorials on
f90 and about 400 copies of M&R were sold there, so there was widespread
knowledge of what it contained. Its weakness was its inability to handle
objects, for which it was also realised that persistence would be
indispensable for handling the petabytes of data expected from the then
planned LHC (which will, in fact, come into operation in 2007 and for which
grid computing has been invented). Here, it is important to recognise that
computing in high-energy physics is not so much about solving a few huge
problems, as in say weather forecasting, but of keeping track of the
complicated data structures that describe each of the billions of individual
events that a high-energy detector will record. The array language was
interesting but not essential.

It dawned on the community that C++ was better suited for its purpose than
f90, and a final decision was reached at a conference on Computing in
High-Energy Physics (CHEP) around 1992 or 3. As a prototype, CERN rewrote
the simulation program GEANT in C++, and it worked. End of story.

For the record,

Mike Metcalf

beliavsky@aol.com

2004-11-16, 3:57 pm

"Michael Metcalf" <michael.metcalf@t-online.de> wrote:

<good information about CERN moving from Fortran to C++ snipped>

I realize that CERN's decision has been made, but had it existed at the time,
would the full Fortran 2003 language, with its OOP features, have met their
requirements?

I don't know what "object persistence" means, but Googling produces many
links.

----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeed.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= 19 East/West-Coast Specialized Servers - Total Privacy via Encryption =---

Danguard

2004-11-16, 3:57 pm

In article <419a323f_1@127.0.0.1>, beliavsky@127.0.0.1 says...
>
> "Michael Metcalf" <michael.metcalf@t-online.de> wrote:
>
> <good information about CERN moving from Fortran to C++ snipped>

Very interesting, yes!

> I realize that CERN's decision has been made, but had it existed at the time,
> would the full Fortran 2003 language, with its OOP features, have met their
> requirements?

Moreover, I think there are other programming languages well suited for
dynamic data / objects management, e.g.: Java (with its built-in garbage
collector, and its persistance facilities), and maybe Python, too.

The same question queried for F2003 could be also queried regarding Java
or Python, for the purpose of "handling objects".

> I don't know what "object persistence" means, but Googling produces many
> links.

Maybe they mean that you can save and reload/restore the "status" of
complex hierarchies of objects, also distributed on different
computers...?

Ciao,
Dan

Toon Moene

2004-11-16, 3:57 pm

Michael Metcalf wrote:

> Here, it is important to recognise that
> computing in high-energy physics is not so much about solving a few huge
> problems, as in say weather forecasting, but of keeping track of the
> complicated data structures that describe each of the billions of individual
> events that a high-energy detector will record. The array language was
> interesting but not essential.

But, surely, there's work done at CERN on theory ? It can't just be a
huge centre for experiments *only* ?

How about designing those colliders ? You're not going to tell me that
knowing what design parameters to choose for such a monster is a back of
the envelope calculation ...

In weather forecasting, before we sent off an O(1bnEuro) satellite into
space, lots and lots of computation has been spent on simulating the
instrument and simulating the impact of its measurements on weather
forecasts. All in Fortran, BTW.

--
Toon Moene - e-mail: toon@moene.indiv.nluug.nl - phone: +31 346 214290
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
A maintainer of GNU Fortran 95: http://gcc.gnu.org/fortran/

Michael Metcalf

2004-11-16, 3:57 pm

"Toon Moene" <toon@moene.indiv.nluug.nl> wrote in message
news:419a4cfe$0$16491$4d4ebb8e@news.nl.uu.net...
>
> But, surely, there's work done at CERN on theory ? It can't just be a
> huge centre for experiments *only* ?
>
Yes, but it's a minor activity in terms of numbers (rather than quality).
Some code, in fact, was written in f90 but it never really took off.

> How about designing those colliders ? You're not going to tell me that
> knowing what design parameters to choose for such a monster is a back of
> the envelope calculation ...
>

Well there, indeed, Fortran 77 and 90 were and are used, sometimes with a C
interface. In fact, an attempt to rewrite a large simulation code for
accelerator design in C++ was a flop. Here we're back to number crunching
too. An article about this by Forrest appeared in a Fortran Forum about two
years ago (I'm away from my copies).

Both these activities are important but do not engage the thousands of
experimentalists around the globe who carry out the experiments and analyze
the data and where, by common agreement, only one language can be used.

Regards,

Mike Metcalf

Michael Metcalf

2004-11-16, 3:57 pm

"Danguard" <danguard_robot@hotmail.com> wrote in message
news:MPG.1c0467caf8c25d1a989a30@powernews.libero.it...
>
time,[color=darkred]
their[color=darkred]
>
> Moreover, I think there are other programming languages well suited for
> dynamic data / objects management, e.g.: Java (with its built-in garbage
> collector, and its persistance facilities), and maybe Python, too.
>
> The same question queried for F2003 could be also queried regarding Java
> or Python, for the purpose of "handling objects".
>

Yes, but other considerations were the use of the language with, for
instance, COBRA, and the number of physicists learning the various
languages.

>
> Maybe they mean that you can save and reload/restore the "status" of
> complex hierarchies of objects, also distributed on different
> computers...?
>
Yes, and self describing.

Regards,

Mike Metcalf

Sponsored Links