Hi there, I have recently been working with the VectorHelmholtz solver and have generated a test case (https://github.com/CMeinersHH/filled_re ... _waveguide) for which there is also an analytical solution (for comparison). It looks quite good so far. However, during execution I still get several messages of the form:
"CheckKeyword: Found keyword type assuming suffix 1 for: variable 2"
This is not critical, but can anyone give advice on what could be done better?
Christian
Filled rectangular wave guide (VectorHelmholtz)
-
- Posts: 1
- Joined: 25 Apr 2024, 00:24
- Antispam: Yes
-
- Site Admin
- Posts: 4841
- Joined: 22 Aug 2009, 11:57
- Antispam: Yes
- Location: Espoo, Finland
- Contact:
Re: Filled rectangular wave guide (VectorHelmholtz)
Hi
This is not a warning: You just have the code rather verbose and it tells you that "variable 2" was given type based on the keyword "variable 1" that is named in the keyword database. We could make this even lower priority.
-Peter
This is not a warning: You just have the code rather verbose and it tells you that "variable 2" was given type based on the keyword "variable 1" that is named in the keyword database. We could make this even lower priority.
-Peter
-
- Posts: 40
- Joined: 22 Oct 2017, 21:03
- Antispam: Yes
Re: Filled rectangular wave guide (VectorHelmholtz)
Hello Peter,
I tried Max Output Level = 1 (or 0), but I still get those messages, in my case a few hundred lines.Can I use another setting?
Kind regards,
Martin
I tried Max Output Level = 1 (or 0), but I still get those messages, in my case a few hundred lines.Can I use another setting?
Kind regards,
Martin
-
- Posts: 2368
- Joined: 25 Jan 2019, 01:28
- Antispam: Yes
Re: Filled rectangular wave guide (VectorHelmholtz)
Check Keywords "Warn"
usually placed in the beginning of the input file. When this command is given, the solver outputs warning
messages if the input file contains keywords that are not listed in the file of known keywords. If new
keywords are introduced, misleading warning messages can be avoided by adding the new keywords to the
keyword file SOLVER.KEYWORDS, located in the directory of the shared library files of ElmerSolver. The
other options include ignore, abort, silent
from the elmersolver manual
usually placed in the beginning of the input file. When this command is given, the solver outputs warning
messages if the input file contains keywords that are not listed in the file of known keywords. If new
keywords are introduced, misleading warning messages can be avoided by adding the new keywords to the
keyword file SOLVER.KEYWORDS, located in the directory of the shared library files of ElmerSolver. The
other options include ignore, abort, silent
from the elmersolver manual
-
- Posts: 40
- Joined: 22 Oct 2017, 21:03
- Antispam: Yes
Re: Filled rectangular wave guide (VectorHelmholtz)
Hi,
I'm a bit confused now.
On 25 Apr 2024, 16:16 , Peter wrote "This is not a warning". Removing the line 'Check Keywords "Warn"' doesn't help.
So, probably I have to write a script to add the new keywords to SOLVER.KEYWORDS (I have a few hundred new key words).
Kind regards,
Martin
I'm a bit confused now.
On 25 Apr 2024, 16:16 , Peter wrote "This is not a warning". Removing the line 'Check Keywords "Warn"' doesn't help.
So, probably I have to write a script to add the new keywords to SOLVER.KEYWORDS (I have a few hundred new key words).
Kind regards,
Martin
-
- Posts: 2368
- Joined: 25 Jan 2019, 01:28
- Antispam: Yes
Re: Filled rectangular wave guide (VectorHelmholtz)
My mistake, It is not CheckKeyWord Warn, it is a message from the model description saying the variable type was assigned using an assumption.
Therefore the CheckKeyWord flag will not work.
Interesting that in the code that the message level is set to 20, ModelDesription.F90
IF ( HashEqualKeys( Val % TYPE, TYPE ) ) THEN
CALL info('CheckKeyword','Found keyword type assuming suffix 1 for: '//TRIM(Name),Level=20)
RETURN
END IF
So it should not output at message level less than 20, but it does, Messages.F90
IF ( PRESENT( Level ) ) THEN
if (Level > MaxOutputLevel) RETURN ! Level = 20 for this message, but MaxOutputLevel is not working?
IF ( .NOT. OutputLevelMask(Level) ) RETURN
ELSE
! The default level of info
!-------------------------------------------
IF( .NOT. OutputLevelMask(DefLevel) ) RETURN
END IF
Therefore the CheckKeyWord flag will not work.
Interesting that in the code that the message level is set to 20, ModelDesription.F90
IF ( HashEqualKeys( Val % TYPE, TYPE ) ) THEN
CALL info('CheckKeyword','Found keyword type assuming suffix 1 for: '//TRIM(Name),Level=20)
RETURN
END IF
So it should not output at message level less than 20, but it does, Messages.F90
IF ( PRESENT( Level ) ) THEN
if (Level > MaxOutputLevel) RETURN ! Level = 20 for this message, but MaxOutputLevel is not working?
IF ( .NOT. OutputLevelMask(Level) ) RETURN
ELSE
! The default level of info
!-------------------------------------------
IF( .NOT. OutputLevelMask(DefLevel) ) RETURN
END IF
-
- Posts: 2368
- Joined: 25 Jan 2019, 01:28
- Antispam: Yes
Re: Filled rectangular wave guide (VectorHelmholtz)
Put in some messages and it appears that during the model info phase upfront
MaxOutputLevel is set to 31 regardless of user input. This assures that some model info is output upfront.
level: 20 MaxOutputLevel: 31
CheckKeyword: Found keyword type assuming suffix 1 for: variable 2
making this change eliminates the message during model info
IF ( HashEqualKeys( Val % TYPE, TYPE ) ) THEN
CALL info('CheckKeyword','Found keyword type assuming suffix 1 for: '//TRIM(Name),Level=32)
RETURN
END IF
MaxOutputLevel is set to 31 regardless of user input. This assures that some model info is output upfront.
level: 20 MaxOutputLevel: 31
CheckKeyword: Found keyword type assuming suffix 1 for: variable 2
making this change eliminates the message during model info
IF ( HashEqualKeys( Val % TYPE, TYPE ) ) THEN
CALL info('CheckKeyword','Found keyword type assuming suffix 1 for: '//TRIM(Name),Level=32)
RETURN
END IF
-
- Posts: 40
- Joined: 22 Oct 2017, 21:03
- Antispam: Yes
Re: Filled rectangular wave guide (VectorHelmholtz)
Hello Kevin,
Thank you for the explanation, but I think this is not something I can change?
Kind regards,
Martin
Thank you for the explanation, but I think this is not something I can change?
Kind regards,
Martin
-
- Posts: 2368
- Joined: 25 Jan 2019, 01:28
- Antispam: Yes
Re: Filled rectangular wave guide (VectorHelmholtz)
Only if you compile your own version of the solver