From yfeng1 at tigers.lsu.edu  Sat Jul  4 00:39:04 2009
From: yfeng1 at tigers.lsu.edu (Yin Feng)
Date: Sat, 4 Jul 2009 00:39:04 -0500
Subject: [Users] A beginner's question
Message-ID: <1e8c69dc0907032239m5e405c84i74a612d560d190fd@mail.gmail.com>

I just want to know whether there is any CCTK parameter or function
can return how many times my code in EVOL have run.
The parameter or function may like a counter.

Thanks in advance!

Sincerely,

From schnetter at cct.lsu.edu  Sat Jul  4 05:13:46 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Sat, 4 Jul 2009 12:13:46 +0200
Subject: [Users] A beginner's question
In-Reply-To: <1e8c69dc0907032239m5e405c84i74a612d560d190fd@mail.gmail.com>
References: <1e8c69dc0907032239m5e405c84i74a612d560d190fd@mail.gmail.com>
Message-ID: <B310D009-376C-4C13-85A9-14F6F14370C9@cct.lsu.edu>

On Jul 4, 2009, at 07:39:04, Yin Feng wrote:

> I just want to know whether there is any CCTK parameter or function
> can return how many times my code in EVOL have run.
> The parameter or function may like a counter.

There is a variable cctk_iteration.  It counts the number of  
iterations, where each iteration (typically) executes the code in  
CCTK_EVOL once.  If you activate info output via IOBasic, then the  
left-most column shows the value of cctk_iteration.  Does this help?   
Do you need this in your code, or do you need this for timing  
statistics?

This works nicely for unigrid; with mesh refinement, things are more  
complex, since fine grids are evolved more often than coarse grids.   
No such counter exists for mesh refinement.

-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 203 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/users/attachments/20090704/66a09930/attachment.bin 

From yfeng1 at tigers.lsu.edu  Sat Jul  4 22:25:08 2009
From: yfeng1 at tigers.lsu.edu (Yin Feng)
Date: Sat, 4 Jul 2009 22:25:08 -0500
Subject: [Users] A beginner's question
In-Reply-To: <B310D009-376C-4C13-85A9-14F6F14370C9@cct.lsu.edu>
References: <1e8c69dc0907032239m5e405c84i74a612d560d190fd@mail.gmail.com>
	<B310D009-376C-4C13-85A9-14F6F14370C9@cct.lsu.edu>
Message-ID: <1e8c69dc0907042025u5634c75cuc3a43ecff11dcf24@mail.gmail.com>

Yes, that's helpful.
I just want use it like
if (cctk_iteration==1000) then {//do somthing}

By the way, is there any thing can help me obtain how long it takes to
run my code?
My code is on parallel computation using OpenMPI.

Thank you very much!

Sincerely,
Yin

On Sat, Jul 4, 2009 at 5:13 AM, Erik Schnetter<schnetter at cct.lsu.edu> wrote:
> On Jul 4, 2009, at 07:39:04, Yin Feng wrote:
>
>> I just want to know whether there is any CCTK parameter or function
>> can return how many times my code in EVOL have run.
>> The parameter or function may like a counter.
>
> There is a variable cctk_iteration. ?It counts the number of iterations,
> where each iteration (typically) executes the code in CCTK_EVOL once. ?If
> you activate info output via IOBasic, then the left-most column shows the
> value of cctk_iteration. ?Does this help? ?Do you need this in your code, or
> do you need this for timing statistics?
>
> This works nicely for unigrid; with mesh refinement, things are more
> complex, since fine grids are evolved more often than coarse grids. ?No such
> counter exists for mesh refinement.
>
> -erik
>
> --
> Erik Schnetter <schnetter at cct.lsu.edu> ? http://www.cct.lsu.edu/~eschnett/
>
> My email is as private as my paper mail. ?I therefore support encrypting
> and signing email messages. ?Get my PGP key from <http://pgp.mit.edu>.
>
>
>
>
> _______________________________________________
> Users mailing list
> Users at cactuscode.org
> http://www.cactuscode.org/mailman/listinfo/users
>
>


From yfeng1 at tigers.lsu.edu  Sun Jul  5 00:22:44 2009
From: yfeng1 at tigers.lsu.edu (Yin Feng)
Date: Sun, 5 Jul 2009 00:22:44 -0500
Subject: [Users] A beginner's question
In-Reply-To: <1e8c69dc0907042025u5634c75cuc3a43ecff11dcf24@mail.gmail.com>
References: <1e8c69dc0907032239m5e405c84i74a612d560d190fd@mail.gmail.com>
	<B310D009-376C-4C13-85A9-14F6F14370C9@cct.lsu.edu>
	<1e8c69dc0907042025u5634c75cuc3a43ecff11dcf24@mail.gmail.com>
Message-ID: <1e8c69dc0907042222r675009adl21cdd644f3d704aa@mail.gmail.com>

when I use cctk_iteration in my code, the system reports error saying
"undeclared parameter".
So, do I need to include other thorns before using it?

On Sat, Jul 4, 2009 at 10:25 PM, Yin Feng<yfeng1 at tigers.lsu.edu> wrote:
> Yes, that's helpful.
> I just want use it like
> if (cctk_iteration==1000) then {//do somthing}
>
> By the way, is there any thing can help me obtain how long it takes to
> run my code?
> My code is on parallel computation using OpenMPI.
>
> Thank you very much!
>
> Sincerely,
> Yin
>
> On Sat, Jul 4, 2009 at 5:13 AM, Erik Schnetter<schnetter at cct.lsu.edu> wrote:
>> On Jul 4, 2009, at 07:39:04, Yin Feng wrote:
>>
>>> I just want to know whether there is any CCTK parameter or function
>>> can return how many times my code in EVOL have run.
>>> The parameter or function may like a counter.
>>
>> There is a variable cctk_iteration. ?It counts the number of iterations,
>> where each iteration (typically) executes the code in CCTK_EVOL once. ?If
>> you activate info output via IOBasic, then the left-most column shows the
>> value of cctk_iteration. ?Does this help? ?Do you need this in your code, or
>> do you need this for timing statistics?
>>
>> This works nicely for unigrid; with mesh refinement, things are more
>> complex, since fine grids are evolved more often than coarse grids. ?No such
>> counter exists for mesh refinement.
>>
>> -erik
>>
>> --
>> Erik Schnetter <schnetter at cct.lsu.edu> ? http://www.cct.lsu.edu/~eschnett/
>>
>> My email is as private as my paper mail. ?I therefore support encrypting
>> and signing email messages. ?Get my PGP key from <http://pgp.mit.edu>.
>>
>>
>>
>>
>> _______________________________________________
>> Users mailing list
>> Users at cactuscode.org
>> http://www.cactuscode.org/mailman/listinfo/users
>>
>>
>


From schnetter at cct.lsu.edu  Sun Jul  5 07:01:38 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Sun, 5 Jul 2009 14:01:38 +0200
Subject: [Users] A beginner's question
In-Reply-To: <1e8c69dc0907042222r675009adl21cdd644f3d704aa@mail.gmail.com>
References: <1e8c69dc0907032239m5e405c84i74a612d560d190fd@mail.gmail.com>
	<B310D009-376C-4C13-85A9-14F6F14370C9@cct.lsu.edu>
	<1e8c69dc0907042025u5634c75cuc3a43ecff11dcf24@mail.gmail.com>
	<1e8c69dc0907042222r675009adl21cdd644f3d704aa@mail.gmail.com>
Message-ID: <F58A12C8-37B9-4DE2-885F-1F325F74B406@cct.lsu.edu>

Here is a short example code:

#include <cctk.h>
#include <cctk_Arguments.h>

subroutine Hello1000 (CCTK_ARGUMENTS)
    implicit none
    DECLARE_CCTK_ARGUMENTS
    if (cctk_iteration == 1000) then
       print '("Hello")'
    end if
end subroutine Hello1000

I'm writing this from memory and haven't actually tested it, but it  
shows how to use CCTK_ARGUMENTS etc.

-erik

On Jul 5, 2009, at 07:22:44, Yin Feng wrote:

> when I use cctk_iteration in my code, the system reports error saying
> "undeclared parameter".
> So, do I need to include other thorns before using it?
>
> On Sat, Jul 4, 2009 at 10:25 PM, Yin Feng<yfeng1 at tigers.lsu.edu>  
> wrote:
>> Yes, that's helpful.
>> I just want use it like
>> if (cctk_iteration==1000) then {//do somthing}
>>
>> By the way, is there any thing can help me obtain how long it takes  
>> to
>> run my code?
>> My code is on parallel computation using OpenMPI.
>>
>> Thank you very much!
>>
>> Sincerely,
>> Yin
>>
>> On Sat, Jul 4, 2009 at 5:13 AM, Erik  
>> Schnetter<schnetter at cct.lsu.edu> wrote:
>>> On Jul 4, 2009, at 07:39:04, Yin Feng wrote:
>>>
>>>> I just want to know whether there is any CCTK parameter or function
>>>> can return how many times my code in EVOL have run.
>>>> The parameter or function may like a counter.
>>>
>>> There is a variable cctk_iteration.  It counts the number of  
>>> iterations,
>>> where each iteration (typically) executes the code in CCTK_EVOL  
>>> once.  If
>>> you activate info output via IOBasic, then the left-most column  
>>> shows the
>>> value of cctk_iteration.  Does this help?  Do you need this in  
>>> your code, or
>>> do you need this for timing statistics?
>>>
>>> This works nicely for unigrid; with mesh refinement, things are more
>>> complex, since fine grids are evolved more often than coarse  
>>> grids.  No such
>>> counter exists for mesh refinement.
>>>
>>> -erik
>>>
>>> --
>>> Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/
>>>
>>> My email is as private as my paper mail.  I therefore support  
>>> encrypting
>>> and signing email messages.  Get my PGP key from <http:// 
>>> pgp.mit.edu>.
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Users mailing list
>>> Users at cactuscode.org
>>> http://www.cactuscode.org/mailman/listinfo/users
>>>
>>>
>>
>
> _______________________________________________
> Users mailing list
> Users at cactuscode.org
> http://www.cactuscode.org/mailman/listinfo/users


-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 203 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/users/attachments/20090705/4a3e3b01/attachment.bin 

From roland.haas at physics.gatech.edu  Mon Jul  6 11:44:19 2009
From: roland.haas at physics.gatech.edu (Roland Haas)
Date: Mon, 06 Jul 2009 12:44:19 -0400
Subject: [Users] A beginner's question
In-Reply-To: <F58A12C8-37B9-4DE2-885F-1F325F74B406@cct.lsu.edu>
References: <1e8c69dc0907032239m5e405c84i74a612d560d190fd@mail.gmail.com>	<B310D009-376C-4C13-85A9-14F6F14370C9@cct.lsu.edu>	<1e8c69dc0907042025u5634c75cuc3a43ecff11dcf24@mail.gmail.com>	<1e8c69dc0907042222r675009adl21cdd644f3d704aa@mail.gmail.com>
	<F58A12C8-37B9-4DE2-885F-1F325F74B406@cct.lsu.edu>
Message-ID: <4A5229E3.2020205@mail.gatech.edu>

Hello all,

>>> I just want use it like
>>> if (cctk_iteration==1000) then {//do somthing}
I have got a related question to Yin's but more Carpet centred. When 
using mesh refinement, cctk_iteration is increased in Carpet's 
AdvanceTime routine only. This seems to cause cctk_iteration to have 
different values for routines scheduled in CCTK_EVOL or CCTK_ANALYSIS 
even when dealing with the same (coarse level, the very finest one is 
ok) data. Is there an official way to find out what time (or iteration) 
the current slice lives on in CCTK_EVOL? Right now I query Carpet about 
the current time refinement level and then round cctk_iteration up to 
the next correct refinement level. The idea is to perform a certain 
calculation only on every nth timestep.

Yours,
Roland

From schnetter at cct.lsu.edu  Mon Jul  6 12:15:30 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Mon, 6 Jul 2009 19:15:30 +0200
Subject: [Users] A beginner's question
In-Reply-To: <4A5229E3.2020205@mail.gatech.edu>
References: <1e8c69dc0907032239m5e405c84i74a612d560d190fd@mail.gmail.com>	<B310D009-376C-4C13-85A9-14F6F14370C9@cct.lsu.edu>	<1e8c69dc0907042025u5634c75cuc3a43ecff11dcf24@mail.gmail.com>	<1e8c69dc0907042222r675009adl21cdd644f3d704aa@mail.gmail.com>
	<F58A12C8-37B9-4DE2-885F-1F325F74B406@cct.lsu.edu>
	<4A5229E3.2020205@mail.gatech.edu>
Message-ID: <51F66DAA-A4CC-4535-984A-C0D55ADF5987@cct.lsu.edu>

On Jul 6, 2009, at 18:44:19, Roland Haas wrote:

> Hello all,
>
>>>> I just want use it like
>>>> if (cctk_iteration==1000) then {//do somthing}
> I have got a related question to Yin's but more Carpet centred. When
> using mesh refinement, cctk_iteration is increased in Carpet's
> AdvanceTime routine only. This seems to cause cctk_iteration to have
> different values for routines scheduled in CCTK_EVOL or CCTK_ANALYSIS
> even when dealing with the same (coarse level, the very finest one is
> ok) data. Is there an official way to find out what time (or  
> iteration)
> the current slice lives on in CCTK_EVOL? Right now I query Carpet  
> about
> the current time refinement level and then round cctk_iteration up to
> the next correct refinement level. The idea is to perform a certain
> calculation only on every nth timestep.


Yes, with mesh refinement, the value of cctk_iteration is different  
(on coarse grids) before and after restricting.

Consider the time evolution of two levels.  The coarse level takes one  
step, the fine level takes two steps.  Both levels begin at the same  
time (let's say at t=0), and therefore CCTK_EVOL is called with the  
same value of cctk_iteration as well (say, cctk_iteration=1).

The next iteration on the fine grid is cctk_iteration=2.  There is no  
corresponding coarse grid time step, since the coarse grid time step  
is twice as large.

After this, both grids are aligned in time, and the fine grid is  
restricted to the coarse grid.  Then analysis is run, still at the  
same time, and therefore cctk_iteration=2 for both grids.

It would be possible to have cctk_iteration=2 for CCTK_EVOL on the  
coarse grid as well.  In this case, cctk_iteration would move non- 
monotonically, and this may be confusing to some code, so I decided  
against it.

-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 203 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/users/attachments/20090706/00d4c756/attachment-0001.bin 

From yfeng1 at tigers.lsu.edu  Thu Jul  9 12:48:25 2009
From: yfeng1 at tigers.lsu.edu (Yin Feng)
Date: Thu, 9 Jul 2009 12:48:25 -0500
Subject: [Users] About ghost_size
Message-ID: <1e8c69dc0907091048v4060735ek2e9553e465065862@mail.gmail.com>

If I tried to solve a matrix eqation Ax=b based on parallel
computation using PETSc and Openmpi
under Cactus frame work. So, in the parameter file, what kind of
ghost_size is meaningful? 1?

By then way, if I run my code on queenbee (4 nodes), and I use qshow
to check each node.
I found the load on each node like:
Node         Load
qb0             32
qb1             0.02
qb2               0
qb3               0
So, is there any possible resean to make this problem?

thank you in advance!

Sincerely,
YIN

From schnetter at cct.lsu.edu  Mon Jul 13 10:17:46 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Mon, 13 Jul 2009 10:17:46 -0500
Subject: [Users] About ghost_size
In-Reply-To: <1e8c69dc0907091048v4060735ek2e9553e465065862@mail.gmail.com>
References: <1e8c69dc0907091048v4060735ek2e9553e465065862@mail.gmail.com>
Message-ID: <3F646F32-4DC2-4E1A-BE50-364CB8FCB369@cct.lsu.edu>

On Jul 9, 2009, at 12:48:25, Yin Feng wrote:

> If I tried to solve a matrix eqation Ax=b based on parallel
> computation using PETSc and Openmpi
> under Cactus frame work. So, in the parameter file, what kind of
> ghost_size is meaningful? 1?
>
> By then way, if I run my code on queenbee (4 nodes), and I use qshow
> to check each node.
> I found the load on each node like:
> Node         Load
> qb0             32
> qb1             0.02
> qb2               0
> qb3               0
> So, is there any possible resean to make this problem?

Yin

Depending on what you do (if PETSc calculates all derivatives), you do  
not need any ghost zones.

Are you at LSU?  If so, it is probably easiest for you to come to my  
office at some time.  You can chat with me on AIM at eschnett247 to  
see when I'm available ? tomorrow would be good for me.

-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 203 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/users/attachments/20090713/6fa9f201/attachment.bin 

From yfeng1 at tigers.lsu.edu  Mon Jul 13 18:15:06 2009
From: yfeng1 at tigers.lsu.edu (Yin Feng)
Date: Mon, 13 Jul 2009 18:15:06 -0500
Subject: [Users] About ghost_size
In-Reply-To: <3F646F32-4DC2-4E1A-BE50-364CB8FCB369@cct.lsu.edu>
References: <1e8c69dc0907091048v4060735ek2e9553e465065862@mail.gmail.com>
	<3F646F32-4DC2-4E1A-BE50-364CB8FCB369@cct.lsu.edu>
Message-ID: <1e8c69dc0907131615m1dd5cbf3q7ffdf8c5f68f878d@mail.gmail.com>

Hi, Erik

Can you leave me your office number?
I am from petroleum engineering department and will try to visit you
tommorow afternoon.

Sincerely,
YIN

On Mon, Jul 13, 2009 at 10:17 AM, Erik Schnetter<schnetter at cct.lsu.edu> wrote:
> On Jul 9, 2009, at 12:48:25, Yin Feng wrote:
>
>> If I tried to solve a matrix eqation Ax=b based on parallel
>> computation using PETSc and Openmpi
>> under Cactus frame work. So, in the parameter file, what kind of
>> ghost_size is meaningful? 1?
>>
>> By then way, if I run my code on queenbee (4 nodes), and I use qshow
>> to check each node.
>> I found the load on each node like:
>> Node ? ? ? ? Load
>> qb0 ? ? ? ? ? ? 32
>> qb1 ? ? ? ? ? ? 0.02
>> qb2 ? ? ? ? ? ? ? 0
>> qb3 ? ? ? ? ? ? ? 0
>> So, is there any possible resean to make this problem?
>
> Yin
>
> Depending on what you do (if PETSc calculates all derivatives), you do not
> need any ghost zones.
>
> Are you at LSU? ?If so, it is probably easiest for you to come to my office
> at some time. ?You can chat with me on AIM at eschnett247 to see when I'm
> available ? tomorrow would be good for me.
>
> -erik
>
> --
> Erik Schnetter <schnetter at cct.lsu.edu> ? http://www.cct.lsu.edu/~eschnett/
>
> My email is as private as my paper mail. ?I therefore support encrypting
> and signing email messages. ?Get my PGP key from <http://pgp.mit.edu>.
>
>
>
>
> _______________________________________________
> Users mailing list
> Users at cactuscode.org
> http://www.cactuscode.org/mailman/listinfo/users
>
>


From schnetter at cct.lsu.edu  Tue Jul 14 07:27:48 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Tue, 14 Jul 2009 07:27:48 -0500
Subject: [Users] About ghost_size
In-Reply-To: <1e8c69dc0907131615m1dd5cbf3q7ffdf8c5f68f878d@mail.gmail.com>
References: <1e8c69dc0907091048v4060735ek2e9553e465065862@mail.gmail.com>
	<3F646F32-4DC2-4E1A-BE50-364CB8FCB369@cct.lsu.edu>
	<1e8c69dc0907131615m1dd5cbf3q7ffdf8c5f68f878d@mail.gmail.com>
Message-ID: <6BCB9A3A-7778-4FC2-A789-CFED5025E823@cct.lsu.edu>

I'm in 327 Johnston Hall.

-erik

On Jul 13, 2009, at 18:15:06, Yin Feng wrote:

> Hi, Erik
>
> Can you leave me your office number?
> I am from petroleum engineering department and will try to visit you
> tommorow afternoon.
>
> Sincerely,
> YIN
>
> On Mon, Jul 13, 2009 at 10:17 AM, Erik  
> Schnetter<schnetter at cct.lsu.edu> wrote:
>> On Jul 9, 2009, at 12:48:25, Yin Feng wrote:
>>
>>> If I tried to solve a matrix eqation Ax=b based on parallel
>>> computation using PETSc and Openmpi
>>> under Cactus frame work. So, in the parameter file, what kind of
>>> ghost_size is meaningful? 1?
>>>
>>> By then way, if I run my code on queenbee (4 nodes), and I use qshow
>>> to check each node.
>>> I found the load on each node like:
>>> Node         Load
>>> qb0             32
>>> qb1             0.02
>>> qb2               0
>>> qb3               0
>>> So, is there any possible resean to make this problem?
>>
>> Yin
>>
>> Depending on what you do (if PETSc calculates all derivatives), you  
>> do not
>> need any ghost zones.
>>
>> Are you at LSU?  If so, it is probably easiest for you to come to  
>> my office
>> at some time.  You can chat with me on AIM at eschnett247 to see  
>> when I'm
>> available ? tomorrow would be good for me.
>>
>> -erik
>>
>> --
>> Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/
>>
>> My email is as private as my paper mail.  I therefore support  
>> encrypting
>> and signing email messages.  Get my PGP key from <http:// 
>> pgp.mit.edu>.
>>
>>
>>
>>
>> _______________________________________________
>> Users mailing list
>> Users at cactuscode.org
>> http://www.cactuscode.org/mailman/listinfo/users
>>
>>
>
> _______________________________________________
> Users mailing list
> Users at cactuscode.org
> http://www.cactuscode.org/mailman/listinfo/users


-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 203 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/users/attachments/20090714/acad6615/attachment.bin 

From ian.hinder at aei.mpg.de  Thu Jul 23 07:41:17 2009
From: ian.hinder at aei.mpg.de (Ian Hinder)
Date: Thu, 23 Jul 2009 14:41:17 +0200
Subject: [Users] schedule.ccl error checking
Message-ID: <2C670599-9DB0-47D3-85BF-B27D2DF93B94@aei.mpg.de>

Hi,

If I schedule a function but forget to include a group or timebin,  
there is no CST error and the function is left silently unscheduled.   
I think that the CST should abort with a fatal error in this case.

   schedule function_name
   {
       LANG:C
   } "description"

-- 
Ian Hinder
ian.hinder at aei.mpg.de


From schnetter at cct.lsu.edu  Fri Jul 24 08:05:01 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Fri, 24 Jul 2009 08:05:01 -0500
Subject: [Users] [CactusMaint] Windows compilation
In-Reply-To: <1248356951.6076.25.camel@epona>
References: <1248356951.6076.25.camel@epona>
Message-ID: <D7A718C5-A8E6-49DB-92B2-01A1DA4C3011@cct.lsu.edu>

On Jul 23, 2009, at 08:49:11, Borja Mi?ano wrote:

> Hi,
>
> I have a problem making a compilation for a thorn under windows.
> I configure it with gcc, c++ and gfortran compilers.
>
> The problem occurs when cactus is checking the status of the thorn.
> It returns a multiple target patterns error.
>
> I check the generated files (xxx.d) and i see that they have the paths
> in DOS-style, c:/Cactus/src/include/cctk.h \.
>
> If i change manually the generated files to have POSIX-style paths and
> it continues the compilation.
>
> I saw that the instruction called is:
>
> /usr/bin/cpp -M -traditional -I"c:\cygwin\include" -I"c:\cygwin\usr
> \local\mpich2-install\include" -I"c:\Cactus\arrangements\WaveToy 
> \WaveToy
> \src" -I"c:\Cactus\arrangements\WaveToy\WaveToy\src\include" -I"c:
> \Cactus\configs\WaveToy\config-data" -I"c:\Cactus\configs\WaveToy
> \bindings\include" -I"c:\Cactus\src\include" -I"c:\Cactus 
> \arrangements"
> -I"c:\Cactus\configs\WaveToy\bindings\Configuration\Thorns" -I"c: 
> \Cactus
> \arrangements\WaveToy\WaveToy\src" -I"c:\Cactus\arrangements\WaveToy
> \WaveToy\src\include" -DTHORN_IS_WaveToy -DFCODE
> -DF90CODE /cygdrive/c/Cactus/arrangements/WaveToy/WaveToy/src/ 
> Startup.F
>> Startup.F.d
>
> It uses -I parameters with POSIX-style.
>
> Is there any way to force in make config flags to use POSIX-Style?


Borja

Cactus uses a variable DIRSEP to decide whether to use / or \ as  
directory separator.  When Cactus detects a Windows system, this  
variable is set to \, which seems to cause problems for you.  As a  
work-around, you can edit the file Cactus/lib/make/known-architectures/ 
cygwin.  In line 156 of that file, you would need to comment out that  
assignment, so that DIRSEP is not set, and the default value of / is  
used.

I am not using Windows, so I cannot guarantee that this works.

Can you send us more details about your operating system?  What  
version of Windows and Cygwin, and what compilers are you using?

Good luck,
-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 203 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/users/attachments/20090724/26791a0f/attachment.bin 

From bminyano at iac3.eu  Fri Jul 24 08:40:32 2009
From: bminyano at iac3.eu (Borja =?ISO-8859-1?Q?Mi=F1ano?=)
Date: Fri, 24 Jul 2009 15:40:32 +0200
Subject: [Users] [CactusMaint] Windows compilation
In-Reply-To: <D7A718C5-A8E6-49DB-92B2-01A1DA4C3011@cct.lsu.edu>
References: <1248356951.6076.25.camel@epona>
	<D7A718C5-A8E6-49DB-92B2-01A1DA4C3011@cct.lsu.edu>
Message-ID: <1248442832.6041.32.camel@epona>

Hi Erik,

This solves the problem of the cygdrive and c:\

But, unfortunately, now there is a new problem.
The compiled file cannot be found by the fortran compiler:
gfortran.exe: /cygdrive/c/Cactus/configs/WaveToy/build/WaveToy/Startup.f: No such file or directory
gfortran.exe: no input files

But the file exists (and was created by Cactus).
I'm making the configuration from the dos command line, not inside the
cygwin command line.

I'm using windows 7 (yes, a evaluation copy).
The cygwin version is 6.1

El vie, 24-07-2009 a las 08:05 -0500, Erik Schnetter escribi?:
> On Jul 23, 2009, at 08:49:11, Borja Mi?ano wrote:
> 
> > Hi,
> >
> > I have a problem making a compilation for a thorn under windows.
> > I configure it with gcc, c++ and gfortran compilers.
> >
> > The problem occurs when cactus is checking the status of the thorn.
> > It returns a multiple target patterns error.
> >
> > I check the generated files (xxx.d) and i see that they have the paths
> > in DOS-style, c:/Cactus/src/include/cctk.h \.
> >
> > If i change manually the generated files to have POSIX-style paths and
> > it continues the compilation.
> >
> > I saw that the instruction called is:
> >
> > /usr/bin/cpp -M -traditional -I"c:\cygwin\include" -I"c:\cygwin\usr
> > \local\mpich2-install\include" -I"c:\Cactus\arrangements\WaveToy 
> > \WaveToy
> > \src" -I"c:\Cactus\arrangements\WaveToy\WaveToy\src\include" -I"c:
> > \Cactus\configs\WaveToy\config-data" -I"c:\Cactus\configs\WaveToy
> > \bindings\include" -I"c:\Cactus\src\include" -I"c:\Cactus 
> > \arrangements"
> > -I"c:\Cactus\configs\WaveToy\bindings\Configuration\Thorns" -I"c: 
> > \Cactus
> > \arrangements\WaveToy\WaveToy\src" -I"c:\Cactus\arrangements\WaveToy
> > \WaveToy\src\include" -DTHORN_IS_WaveToy -DFCODE
> > -DF90CODE /cygdrive/c/Cactus/arrangements/WaveToy/WaveToy/src/ 
> > Startup.F
> >> Startup.F.d
> >
> > It uses -I parameters with POSIX-style.
> >
> > Is there any way to force in make config flags to use POSIX-Style?
> 
> 
> Borja
> 
> Cactus uses a variable DIRSEP to decide whether to use / or \ as  
> directory separator.  When Cactus detects a Windows system, this  
> variable is set to \, which seems to cause problems for you.  As a  
> work-around, you can edit the file Cactus/lib/make/known-architectures/ 
> cygwin.  In line 156 of that file, you would need to comment out that  
> assignment, so that DIRSEP is not set, and the default value of / is  
> used.
> 
> I am not using Windows, so I cannot guarantee that this works.
> 
> Can you send us more details about your operating system?  What  
> version of Windows and Cygwin, and what compilers are you using?
> 
> Good luck,
> -erik
> 


From schnetter at cct.lsu.edu  Fri Jul 24 09:48:44 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Fri, 24 Jul 2009 09:48:44 -0500
Subject: [Users] [CactusMaint] Windows compilation
In-Reply-To: <1248442832.6041.32.camel@epona>
References: <1248356951.6076.25.camel@epona>
	<D7A718C5-A8E6-49DB-92B2-01A1DA4C3011@cct.lsu.edu>
	<1248442832.6041.32.camel@epona>
Message-ID: <363908B0-D5B8-42FC-B0E8-B90B40C379C8@cct.lsu.edu>

On Jul 24, 2009, at 08:40:32, Borja Mi?ano wrote:

> Hi Erik,
>
> This solves the problem of the cygdrive and c:\
>
> But, unfortunately, now there is a new problem.
> The compiled file cannot be found by the fortran compiler:
> gfortran.exe: /cygdrive/c/Cactus/configs/WaveToy/build/WaveToy/ 
> Startup.f: No such file or directory
> gfortran.exe: no input files
>
> But the file exists (and was created by Cactus).
> I'm making the configuration from the dos command line, not inside the
> cygwin command line.

I'm afraid I can only offer generic advice here.  Maybe someone else  
is using Cactus on Windows and could offer some hints?

- Does the file exist?  Can you "ls" or "more" it?
- If the file exists, does the name have the same case, or do some  
characters have a different case?  In particular, is there a confusion  
regarding the suffix .f and .F?
- I don't know the difference between a DOS and a Cygwin command  
line.  Since you are using DOS, can you try Cygwin instead?
- Have you used Cactus on a Unix machine before?  That could tell you  
the difference between a system problem, and an error on your side  
that Cactus doesn't detect.

There is a file README.Windows in the Cactus doc directory.  It speaks  
of Cygwin 1.3, which seems quite a bit older than the 6.1 you  
mention.  I'm afraid that Cactus on Windows may not have been tested  
for some time.  Without direct access, these kinds of problems are  
usually difficult to debug.

> I'm using windows 7 (yes, a evaluation copy).
> The cygwin version is 6.1


Thanks.

-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 203 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/users/attachments/20090724/63e7390e/attachment.bin 

From bminyano at iac3.eu  Mon Jul 27 04:18:05 2009
From: bminyano at iac3.eu (Borja =?ISO-8859-1?Q?Mi=F1ano?=)
Date: Mon, 27 Jul 2009 11:18:05 +0200
Subject: [Users] [CactusMaint] Windows compilation
In-Reply-To: <363908B0-D5B8-42FC-B0E8-B90B40C379C8@cct.lsu.edu>
References: <1248356951.6076.25.camel@epona>
	<D7A718C5-A8E6-49DB-92B2-01A1DA4C3011@cct.lsu.edu>
	<1248442832.6041.32.camel@epona>
	<363908B0-D5B8-42FC-B0E8-B90B40C379C8@cct.lsu.edu>
Message-ID: <1248686285.6083.9.camel@epona>

Hi Erik,

Yes, the file exists, so i suppose the problem is related to the paths
(c:\ or /cygdrive/c).
The name is exacly the same case.
I am running the same code in a Linux, and it works perfectly.

I am looking inside the scripts but i still have no idea to solve this.

El vie, 24-07-2009 a las 09:48 -0500, Erik Schnetter escribi?:
> On Jul 24, 2009, at 08:40:32, Borja Mi?ano wrote:
> 
> > Hi Erik,
> >
> > This solves the problem of the cygdrive and c:\
> >
> > But, unfortunately, now there is a new problem.
> > The compiled file cannot be found by the fortran compiler:
> > gfortran.exe: /cygdrive/c/Cactus/configs/WaveToy/build/WaveToy/ 
> > Startup.f: No such file or directory
> > gfortran.exe: no input files
> >
> > But the file exists (and was created by Cactus).
> > I'm making the configuration from the dos command line, not inside the
> > cygwin command line.
> 
> I'm afraid I can only offer generic advice here.  Maybe someone else  
> is using Cactus on Windows and could offer some hints?
> 
> - Does the file exist?  Can you "ls" or "more" it?
> - If the file exists, does the name have the same case, or do some  
> characters have a different case?  In particular, is there a confusion  
> regarding the suffix .f and .F?
> - I don't know the difference between a DOS and a Cygwin command  
> line.  Since you are using DOS, can you try Cygwin instead?
> - Have you used Cactus on a Unix machine before?  That could tell you  
> the difference between a system problem, and an error on your side  
> that Cactus doesn't detect.
> 
> There is a file README.Windows in the Cactus doc directory.  It speaks  
> of Cygwin 1.3, which seems quite a bit older than the 6.1 you  
> mention.  I'm afraid that Cactus on Windows may not have been tested  
> for some time.  Without direct access, these kinds of problems are  
> usually difficult to debug.
> 
> > I'm using windows 7 (yes, a evaluation copy).
> > The cygwin version is 6.1
> 
> 
> Thanks.
> 
> -erik
> 


From bminyano at iac3.eu  Mon Jul 27 05:42:59 2009
From: bminyano at iac3.eu (Borja =?ISO-8859-1?Q?Mi=F1ano?=)
Date: Mon, 27 Jul 2009 12:42:59 +0200
Subject: [Users] [CactusMaint] Windows compilation
In-Reply-To: <363908B0-D5B8-42FC-B0E8-B90B40C379C8@cct.lsu.edu>
References: <1248356951.6076.25.camel@epona>
	<D7A718C5-A8E6-49DB-92B2-01A1DA4C3011@cct.lsu.edu>
	<1248442832.6041.32.camel@epona>
	<363908B0-D5B8-42FC-B0E8-B90B40C379C8@cct.lsu.edu>
Message-ID: <1248691379.6083.15.camel@epona>

Hi again Erik,

I have seen an error.

When i commented the line i got an end-of-line style error that i
ignored. So the cygwin configuration was not taken.
I have corrected the problem, so the configuration is taken, but i have
the previous error c:\ instead of /cygdrive/c.

So commenting the line 156 didn't make the solution.

El vie, 24-07-2009 a las 09:48 -0500, Erik Schnetter escribi?:
> On Jul 24, 2009, at 08:40:32, Borja Mi?ano wrote:
> 
> > Hi Erik,
> >
> > This solves the problem of the cygdrive and c:\
> >
> > But, unfortunately, now there is a new problem.
> > The compiled file cannot be found by the fortran compiler:
> > gfortran.exe: /cygdrive/c/Cactus/configs/WaveToy/build/WaveToy/ 
> > Startup.f: No such file or directory
> > gfortran.exe: no input files
> >
> > But the file exists (and was created by Cactus).
> > I'm making the configuration from the dos command line, not inside the
> > cygwin command line.
> 
> I'm afraid I can only offer generic advice here.  Maybe someone else  
> is using Cactus on Windows and could offer some hints?
> 
> - Does the file exist?  Can you "ls" or "more" it?
> - If the file exists, does the name have the same case, or do some  
> characters have a different case?  In particular, is there a confusion  
> regarding the suffix .f and .F?
> - I don't know the difference between a DOS and a Cygwin command  
> line.  Since you are using DOS, can you try Cygwin instead?
> - Have you used Cactus on a Unix machine before?  That could tell you  
> the difference between a system problem, and an error on your side  
> that Cactus doesn't detect.
> 
> There is a file README.Windows in the Cactus doc directory.  It speaks  
> of Cygwin 1.3, which seems quite a bit older than the 6.1 you  
> mention.  I'm afraid that Cactus on Windows may not have been tested  
> for some time.  Without direct access, these kinds of problems are  
> usually difficult to debug.
> 
> > I'm using windows 7 (yes, a evaluation copy).
> > The cygwin version is 6.1
> 
> 
> Thanks.
> 
> -erik
> 


From bminyano at iac3.eu  Mon Jul 27 10:25:14 2009
From: bminyano at iac3.eu (Borja =?ISO-8859-1?Q?Mi=F1ano?=)
Date: Mon, 27 Jul 2009 17:25:14 +0200
Subject: [Users] [CactusMaint] Windows compilation
In-Reply-To: <363908B0-D5B8-42FC-B0E8-B90B40C379C8@cct.lsu.edu>
References: <1248356951.6076.25.camel@epona>
	<D7A718C5-A8E6-49DB-92B2-01A1DA4C3011@cct.lsu.edu>
	<1248442832.6041.32.camel@epona>
	<363908B0-D5B8-42FC-B0E8-B90B40C379C8@cct.lsu.edu>
Message-ID: <1248708314.6083.23.camel@epona>

Hi again,

I've found where is the problem.

The following lines in cygwin know architecture file are the problematic
ones:

    DIRSEP="\\\\"

# Convert //a to a: 

  GET_WD="cygpath -wa ."

  TRANSFORM_DIRS='$(shell cygpath -wa $(1))'

These lines makes that the paths were converted to DOS paths is several
instructions.
I eliminated them so the file Startup.f.d is generated with UNIX paths,
but another make calling uses these variables.
The fortran compilation uses it and if the lines are deleted then the
not found exception occurs for UNIX paths.

So if the lines exists the generation of the Start.f.d fails, and if
they don't exist then the fortran compilation fails.

Any idea or suggestion about how to solve this?

El vie, 24-07-2009 a las 09:48 -0500, Erik Schnetter escribi?:

> On Jul 24, 2009, at 08:40:32, Borja Mi?ano wrote:
> 
> > Hi Erik,
> >
> > This solves the problem of the cygdrive and c:\
> >
> > But, unfortunately, now there is a new problem.
> > The compiled file cannot be found by the fortran compiler:
> > gfortran.exe: /cygdrive/c/Cactus/configs/WaveToy/build/WaveToy/ 
> > Startup.f: No such file or directory
> > gfortran.exe: no input files
> >
> > But the file exists (and was created by Cactus).
> > I'm making the configuration from the dos command line, not inside the
> > cygwin command line.
> 
> I'm afraid I can only offer generic advice here.  Maybe someone else  
> is using Cactus on Windows and could offer some hints?
> 
> - Does the file exist?  Can you "ls" or "more" it?
> - If the file exists, does the name have the same case, or do some  
> characters have a different case?  In particular, is there a confusion  
> regarding the suffix .f and .F?
> - I don't know the difference between a DOS and a Cygwin command  
> line.  Since you are using DOS, can you try Cygwin instead?
> - Have you used Cactus on a Unix machine before?  That could tell you  
> the difference between a system problem, and an error on your side  
> that Cactus doesn't detect.
> 
> There is a file README.Windows in the Cactus doc directory.  It speaks  
> of Cygwin 1.3, which seems quite a bit older than the 6.1 you  
> mention.  I'm afraid that Cactus on Windows may not have been tested  
> for some time.  Without direct access, these kinds of problems are  
> usually difficult to debug.
> 
> > I'm using windows 7 (yes, a evaluation copy).
> > The cygwin version is 6.1
> 
> 
> Thanks.
> 
> -erik
> 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cactuscode.org/pipermail/users/attachments/20090727/26ad8c35/attachment.html 

From schnetter at cct.lsu.edu  Mon Jul 27 11:29:58 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Mon, 27 Jul 2009 11:29:58 -0500
Subject: [Users] [CactusMaint] Windows compilation
In-Reply-To: <1248708314.6083.23.camel@epona>
References: <1248356951.6076.25.camel@epona>
	<D7A718C5-A8E6-49DB-92B2-01A1DA4C3011@cct.lsu.edu>
	<1248442832.6041.32.camel@epona>
	<363908B0-D5B8-42FC-B0E8-B90B40C379C8@cct.lsu.edu>
	<1248708314.6083.23.camel@epona>
Message-ID: <43DA2E02-7A31-425F-81E5-1378993DA023@cct.lsu.edu>

After removing these lines in the Windows "known architecture" file,  
did you create a configuration from scratch?

-erik

On Jul 27, 2009, at 10:25:14, Borja Mi?ano wrote:

> Hi again,
>
> I've found where is the problem.
>
> The following lines in cygwin know architecture file are the  
> problematic ones:
>
>     DIRSEP="\\\\"
>
> # Convert //a to a:
>
>   GET_WD="cygpath -wa ."
>
>   TRANSFORM_DIRS='$(shell cygpath -wa $(1))'
>
> These lines makes that the paths were converted to DOS paths is  
> several instructions.
> I eliminated them so the file Startup.f.d is generated with UNIX  
> paths, but another make calling uses these variables.
> The fortran compilation uses it and if the lines are deleted then  
> the not found exception occurs for UNIX paths.
>
> So if the lines exists the generation of the Start.f.d fails, and if  
> they don't exist then the fortran compilation fails.
>
> Any idea or suggestion about how to solve this?
>
> El vie, 24-07-2009 a las 09:48 -0500, Erik Schnetter escribi?:
>>
>> On Jul 24, 2009, at 08:40:32, Borja Mi?ano wrote:
>>
>> > Hi Erik,
>> >
>> > This solves the problem of the cygdrive and c:\
>> >
>> > But, unfortunately, now there is a new problem.
>> > The compiled file cannot be found by the fortran compiler:
>> > gfortran.exe: /cygdrive/c/Cactus/configs/WaveToy/build/WaveToy/
>> > Startup.f: No such file or directory
>> > gfortran.exe: no input files
>> >
>> > But the file exists (and was created by Cactus).
>> > I'm making the configuration from the dos command line, not  
>> inside the
>> > cygwin command line.
>>
>> I'm afraid I can only offer generic advice here.  Maybe someone else
>> is using Cactus on Windows and could offer some hints?
>>
>> - Does the file exist?  Can you "ls" or "more" it?
>> - If the file exists, does the name have the same case, or do some
>> characters have a different case?  In particular, is there a  
>> confusion
>> regarding the suffix .f and .F?
>> - I don't know the difference between a DOS and a Cygwin command
>> line.  Since you are using DOS, can you try Cygwin instead?
>> - Have you used Cactus on a Unix machine before?  That could tell you
>> the difference between a system problem, and an error on your side
>> that Cactus doesn't detect.
>>
>> There is a file README.Windows in the Cactus doc directory.  It  
>> speaks
>> of Cygwin 1.3, which seems quite a bit older than the 6.1 you
>> mention.  I'm afraid that Cactus on Windows may not have been tested
>> for some time.  Without direct access, these kinds of problems are
>> usually difficult to debug.
>>
>> > I'm using windows 7 (yes, a evaluation copy).
>> > The cygwin version is 6.1
>>
>>
>> Thanks.
>>
>> -erik
>>


-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 203 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/users/attachments/20090727/e806ddd9/attachment.bin 

From bminyano at iac3.eu  Tue Jul 28 07:58:40 2009
From: bminyano at iac3.eu (Borja =?ISO-8859-1?Q?Mi=F1ano?=)
Date: Tue, 28 Jul 2009 14:58:40 +0200
Subject: [Users] [CactusMaint] Windows compilation
In-Reply-To: <43DA2E02-7A31-425F-81E5-1378993DA023@cct.lsu.edu>
References: <1248356951.6076.25.camel@epona>
	<D7A718C5-A8E6-49DB-92B2-01A1DA4C3011@cct.lsu.edu>
	<1248442832.6041.32.camel@epona>
	<363908B0-D5B8-42FC-B0E8-B90B40C379C8@cct.lsu.edu>
	<1248708314.6083.23.camel@epona>
	<43DA2E02-7A31-425F-81E5-1378993DA023@cct.lsu.edu>
Message-ID: <1248785920.5864.38.camel@epona>

Yes,
Creating it from scratch without the lines the error is:

gfortran.exe: /cygdrive/c/Cactus/configs/WaveToy/build/WaveToy/Startup.f: No such file or directory

El lun, 27-07-2009 a las 11:29 -0500, Erik Schnetter escribi?:
> After removing these lines in the Windows "known architecture" file,  
> did you create a configuration from scratch?
> 
> -erik
> 
> On Jul 27, 2009, at 10:25:14, Borja Mi?ano wrote:
> 
> > Hi again,
> >
> > I've found where is the problem.
> >
> > The following lines in cygwin know architecture file are the  
> > problematic ones:
> >
> >     DIRSEP="\\\\"
> >
> > # Convert //a to a:
> >
> >   GET_WD="cygpath -wa ."
> >
> >   TRANSFORM_DIRS='$(shell cygpath -wa $(1))'
> >
> > These lines makes that the paths were converted to DOS paths is  
> > several instructions.
> > I eliminated them so the file Startup.f.d is generated with UNIX  
> > paths, but another make calling uses these variables.
> > The fortran compilation uses it and if the lines are deleted then  
> > the not found exception occurs for UNIX paths.
> >
> > So if the lines exists the generation of the Start.f.d fails, and if  
> > they don't exist then the fortran compilation fails.
> >
> > Any idea or suggestion about how to solve this?
> >
> > El vie, 24-07-2009 a las 09:48 -0500, Erik Schnetter escribi?:
> >>
> >> On Jul 24, 2009, at 08:40:32, Borja Mi?ano wrote:
> >>
> >> > Hi Erik,
> >> >
> >> > This solves the problem of the cygdrive and c:\
> >> >
> >> > But, unfortunately, now there is a new problem.
> >> > The compiled file cannot be found by the fortran compiler:
> >> > gfortran.exe: /cygdrive/c/Cactus/configs/WaveToy/build/WaveToy/
> >> > Startup.f: No such file or directory
> >> > gfortran.exe: no input files
> >> >
> >> > But the file exists (and was created by Cactus).
> >> > I'm making the configuration from the dos command line, not  
> >> inside the
> >> > cygwin command line.
> >>
> >> I'm afraid I can only offer generic advice here.  Maybe someone else
> >> is using Cactus on Windows and could offer some hints?
> >>
> >> - Does the file exist?  Can you "ls" or "more" it?
> >> - If the file exists, does the name have the same case, or do some
> >> characters have a different case?  In particular, is there a  
> >> confusion
> >> regarding the suffix .f and .F?
> >> - I don't know the difference between a DOS and a Cygwin command
> >> line.  Since you are using DOS, can you try Cygwin instead?
> >> - Have you used Cactus on a Unix machine before?  That could tell you
> >> the difference between a system problem, and an error on your side
> >> that Cactus doesn't detect.
> >>
> >> There is a file README.Windows in the Cactus doc directory.  It  
> >> speaks
> >> of Cygwin 1.3, which seems quite a bit older than the 6.1 you
> >> mention.  I'm afraid that Cactus on Windows may not have been tested
> >> for some time.  Without direct access, these kinds of problems are
> >> usually difficult to debug.
> >>
> >> > I'm using windows 7 (yes, a evaluation copy).
> >> > The cygwin version is 6.1
> >>
> >>
> >> Thanks.
> >>
> >> -erik
> >>
> 
> 


From schnetter at cct.lsu.edu  Wed Jul 29 11:31:28 2009
From: schnetter at cct.lsu.edu (Erik Schnetter)
Date: Wed, 29 Jul 2009 11:31:28 -0500
Subject: [Users] schedule.ccl error checking
In-Reply-To: <2C670599-9DB0-47D3-85BF-B27D2DF93B94@aei.mpg.de>
References: <2C670599-9DB0-47D3-85BF-B27D2DF93B94@aei.mpg.de>
Message-ID: <B8AC4774-5497-4861-B891-02762B7261AD@cct.lsu.edu>

On Jul 23, 2009, at 07:41:17, Ian Hinder wrote:

> Hi,
>
> If I schedule a function but forget to include a group or timebin,
> there is no CST error and the function is left silently unscheduled.
> I think that the CST should abort with a fatal error in this case.
>
>   schedule function_name
>   {
>       LANG:C
>   } "description"


You are right, such a schedule item is always a no-op and most  
certainly caused by an oversight.  This should be an error, or at  
least a warning.

This error should probably be caught in the routine ParseScheduleBlock  
in file lib/sbin/ScheduleParser.pl.  Do you want to give it a try and  
suggest a patch?  For example, this happens for a LANG specifier for a  
schedule group:

             &CST_error(1, "Scheduling group \"$name\" with LANG  
specifier in thorn \"$thorn\"","Groups should not have a LANG  
specificier",__LINE__,__FILE__);

-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from <http://pgp.mit.edu>.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 203 bytes
Desc: This is a digitally signed message part
Url : http://www.cactuscode.org/pipermail/users/attachments/20090729/5ddeb635/attachment.bin