Huge differences in processing speed with different SWAT executables

[Jochen]

Hi,

I did a test comparing 4 different SWAT revisions
rev 510
rev 579 (the latest one)
rev 481 (the one SWAT CUP comes with)
rev 488

rev 510 and rev 488 were very slow compared to rev 481 and rev 579 (up to 5 times slower).
I tested the different revisions in SWAT CUP, too and rev 579 literally rushes through simulations (which I highly appreciate, of course).

Does anyone know where these differences could come from (same input, machine...)?


Best regards
Jochen

[Mauricio Zambrano-Bigiarini]

For the same source code, you can get differences in performance by
using different fortan compilers, or -for the same compiler- by using
different compiler options at the moment of creating the executable
file.

Kinds,

Mauricio

--
=====================================
Water Resources Unit
Institute for Environment and Sustainability
Joint Research Centre, European Commission
webinfo : http://floods.jrc.ec.europa.eu/
=====================================
DISCLAIMER:
"The views expressed are purely those of the writer
and may not in any circumstances be regarded as
stating an official position of the European Commission"
======================================
Linux user #454569 -- Ubuntu user #17469
======================================
"It is a mistake to look too far ahead. Only one
link in the chain of destiny can be handled at a time"
(Sir Winston Churchill)
2012/11/5 Jochen <Jochen....@gmx.de>:

> --
> You received this message because you are subscribed to the Google Groups
> "SWAT-user" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/swatuser/-/_tBfRmUWyG0J.
> To post to this group, send email to swat...@googlegroups.com.
> To unsubscribe from this group, send email to
> swatuser+u...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/swatuser?hl=en.

[Jochen]

Ok, then their compiler obviously had a bad day when creating rev 510.

Thanks for your answer.
Best regards
Jochen

[Mauricio Zambrano-Bigiarini]

I'm not sure where I read it, but I think some versions are compiled
in a different (slower) way on purpose, for allowing to debug possible
errors that appear while running SWAT....

Cheers,

Mauricio

2012/11/5 Jochen <Jochen....@gmx.de>:

> --
> You received this message because you are subscribed to the Google Groups
> "SWAT-user" group.
> To view this discussion on the web visit

> https://groups.google.com/d/msg/swatuser/-/UakhcAjItSwJ.

>
> To post to this group, send email to swat...@googlegroups.com.
> To unsubscribe from this group, send email to
> swatuser+u...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/swatuser?hl=en.

--
=====================================
Water Resources Unit
Institute for Environment and Sustainability
Joint Research Centre, European Commission
webinfo : http://floods.jrc.ec.europa.eu/
=====================================
DISCLAIMER:
"The views expressed are purely those of the writer
and may not in any circumstances be regarded as
stating an official position of the European Commission"
======================================
Linux user #454569 -- Ubuntu user #17469
======================================
"It is a mistake to look too far ahead. Only one
link in the chain of destiny can be handled at a time"
(Sir Winston Churchill)

======================================
http://c2.com/cgi/wiki?HowToAskQuestionsTheSmartWay

[John Joseph]

I've noticed big differences too.  I'm no expert, depending on settings/error reporting options selected for compiling can make a big difference.  Also, one simple thing apart from compiling: the print options under "file.cio" can make a noticeable difference. --- On Mon, 11/5/12, Jochen <Jochen....@gmx.de> wrote:

[Jochen]

That's good news for me. I had no clue in first place and but I doubted that my mgt files were read correctly (as I didn't create them within the ArcSWAT environment).

Best regards
Jochen

[Mikhail Titov]

On Monday, November 5, 2012 10:09:47 PM UTC-6, JJ wrote:

I've noticed big differences too.  I'm no expert, depending on settings/error reporting options selected for compiling can make a big difference.  Also, one simple thing apart from compiling: the print options under "file.cio" can make a noticeable difference.

I've noticed that IO in general takes lots of time. I find convenient especially for SWAT-CUP to set up RAM drive to boost things. It cut down SWAT-CUP iteration time more than twice for me.

Compiler flags do make difference but I thought that there are several executables shipped for that purpose of debugging. Perhaps some checks were left unintentionally.

--
Mikhail