xolotl-psi-development — Xolotl Dev List

Re: [Xolotl-psi-development] Figuring out parts of IFunction in the ADR prototype

[Jay J. B. <bil...@or...>]

Barry,

Ah, I see. That makes sense. I thought we were just updating the 
concentration there. I didn't notice that F was actually Cdot and that 
we needed to minimize it, (cdot-...=0).

Thanks,
Jay

On 07/01/2013 02:30 PM, Barry Smith wrote:
> On Jul 1, 2013, at 1:25 PM, Jay Jay Billings <bil...@or...> wrote:
>
>> Barry,
>>
>> OK, thanks. But... why is the population of He[He] still decreasing
>> (i.e. - f[xi].He[He] -= ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he])?
>>
>> Jay
>      The - sign actually indicates a increase and that having a + would be a decrease. The equation is written as
> cdot[xi].He[He] - ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he] = 0. If you move the ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he]  to the other side and write it as
> cdot[xi].He[He] = +ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he]
> it is now a + and thus He would be increasing.
>
>      Barry
>
>
>> On 07/01/2013 02:03 PM, Barry Smith wrote:
>>>     Jay,
>>>
>>>      Sorry the initial comment is incorrect; probably I started that way and then changed the code and not the comment. It should
>>> read
>>>
>>>      /*He[He-he] + He[he] -> He[He] */
>>>
>>> On Jul 1, 2013, at 11:39 AM, Jay Jay Billings <jay...@gm...> wrote:
>>>
>>>> Barry,
>>>>
>>>> I am looking at the first part of IFunction in the ADR prototype and noticed something that seemed strange to me. Here's the part in question:
>>>>
>>>>      /* ----------------------------------------------------------------
>>>>       ---- Compute reaction terms that can create a cluster of given size
>>>>      */
>>>>      /*   He[He] + He[he] -> He[He+he]  */
>>>>      for (He=2; He<N+1; He++) {
>>>>        /* compute all pairs of clusters of smaller size that can combine to create a cluster of size He,
>>>>           remove the upper half since they are symmetric to the lower half of the pairs. For example
>>>>                when He = 5 (cluster size 5) the pairs are
>>>>                   1   4
>>>>                   2   2
>>>>                   3   2  these last two are not needed in the sum since they repeat from above
>>>>                   4   1  this is why he < (He/2) + 1            */
>>>>        for (he=1; he<(He/2)+1; he++) {
>>>>          f[xi].He[He] -= ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
>>>>
>>>>          /* remove the two clusters that merged to form the larger cluster */
>>>>          f[xi].He[he]    += ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
>>>>          f[xi].He[He-he] += ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
>>>>        }
>>>>      }
>>>>
>>>> In the inner most for loop, I can see why you decrease the concentration of He by the specified amount, but why are you increasing the concentration of he? Shouldn't it decrease too and the concentration of He+he increase? I also notice that you are increasing the population of He-he, but shouldn't it be He+he?
>>>>
>>>> It occurs to me that I could be counting incorrectly and if so, sorry for the trouble!
>>>>
>>>> Thanks,
>>>> Jay
>>>>
>>>> -- 
>>>> Jay Jay Billings
>>>> Oak Ridge National Laboratory
>>>> Twitter Handle: @jayjaybillings
>>>> ------------------------------------------------------------------------------
>>>> This SF.net email is sponsored by Windows:
>>>>
>>>> Build for Windows Store.
>>>>
>>>> http://p.sf.net/sfu/windows-dev2dev_______________________________________________
>>>> Xolotl-psi-development mailing list
>>>> Xol...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>> ------------------------------------------------------------------------------
>>> This SF.net email is sponsored by Windows:
>>>
>>> Build for Windows Store.
>>>
>>> http://p.sf.net/sfu/windows-dev2dev
>>> _______________________________________________
>>> Xolotl-psi-development mailing list
>>> Xol...@li...
>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>
>> -- 
>> Jay Jay Billings
>> Oak Ridge National Laboratory
>> Telephone: (865) 272-9420
>> Email: bil...@or...
>> Twitter Handle: @jayjaybillings
>>
>>
>> ------------------------------------------------------------------------------
>> This SF.net email is sponsored by Windows:
>>
>> Build for Windows Store.
>>
>> http://p.sf.net/sfu/windows-dev2dev
>> _______________________________________________
>> Xolotl-psi-development mailing list
>> Xol...@li...
>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development


-- 
Jay Jay Billings
Oak Ridge National Laboratory
Telephone: (865) 272-9420
Email: bil...@or...
Twitter Handle: @jayjaybillings

Re: [Xolotl-psi-development] Figuring out parts of IFunction in the ADR prototype

[Barry S. <bs...@mc...>]

On Jul 1, 2013, at 1:25 PM, Jay Jay Billings <bil...@or...> wrote:

> Barry,
> 
> OK, thanks. But... why is the population of He[He] still decreasing 
> (i.e. - f[xi].He[He] -= ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he])?
> 
> Jay

    The - sign actually indicates a increase and that having a + would be a decrease. The equation is written as 
cdot[xi].He[He] - ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he] = 0. If you move the ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he]  to the other side and write it as 
cdot[xi].He[He] = +ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he] 
it is now a + and thus He would be increasing.

    Barry


> 
> On 07/01/2013 02:03 PM, Barry Smith wrote:
>>    Jay,
>> 
>>     Sorry the initial comment is incorrect; probably I started that way and then changed the code and not the comment. It should
>> read
>> 
>>     /*He[He-he] + He[he] -> He[He] */
>> 
>> On Jul 1, 2013, at 11:39 AM, Jay Jay Billings <jay...@gm...> wrote:
>> 
>>> Barry,
>>> 
>>> I am looking at the first part of IFunction in the ADR prototype and noticed something that seemed strange to me. Here's the part in question:
>>> 
>>>     /* ----------------------------------------------------------------
>>>      ---- Compute reaction terms that can create a cluster of given size
>>>     */
>>>     /*   He[He] + He[he] -> He[He+he]  */
>>>     for (He=2; He<N+1; He++) {
>>>       /* compute all pairs of clusters of smaller size that can combine to create a cluster of size He,
>>>          remove the upper half since they are symmetric to the lower half of the pairs. For example
>>>               when He = 5 (cluster size 5) the pairs are
>>>                  1   4
>>>                  2   2
>>>                  3   2  these last two are not needed in the sum since they repeat from above
>>>                  4   1  this is why he < (He/2) + 1            */
>>>       for (he=1; he<(He/2)+1; he++) {
>>>         f[xi].He[He] -= ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
>>> 
>>>         /* remove the two clusters that merged to form the larger cluster */
>>>         f[xi].He[he]    += ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
>>>         f[xi].He[He-he] += ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
>>>       }
>>>     }
>>> 
>>> In the inner most for loop, I can see why you decrease the concentration of He by the specified amount, but why are you increasing the concentration of he? Shouldn't it decrease too and the concentration of He+he increase? I also notice that you are increasing the population of He-he, but shouldn't it be He+he?
>>> 
>>> It occurs to me that I could be counting incorrectly and if so, sorry for the trouble!
>>> 
>>> Thanks,
>>> Jay
>>> 
>>> -- 
>>> Jay Jay Billings
>>> Oak Ridge National Laboratory
>>> Twitter Handle: @jayjaybillings
>>> ------------------------------------------------------------------------------
>>> This SF.net email is sponsored by Windows:
>>> 
>>> Build for Windows Store.
>>> 
>>> http://p.sf.net/sfu/windows-dev2dev_______________________________________________
>>> Xolotl-psi-development mailing list
>>> Xol...@li...
>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>> 
>> ------------------------------------------------------------------------------
>> This SF.net email is sponsored by Windows:
>> 
>> Build for Windows Store.
>> 
>> http://p.sf.net/sfu/windows-dev2dev
>> _______________________________________________
>> Xolotl-psi-development mailing list
>> Xol...@li...
>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
> 
> 
> -- 
> Jay Jay Billings
> Oak Ridge National Laboratory
> Telephone: (865) 272-9420
> Email: bil...@or...
> Twitter Handle: @jayjaybillings
> 
> 
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
> 
> Build for Windows Store.
> 
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development

Re: [Xolotl-psi-development] Figuring out parts of IFunction in the ADR prototype

[Jay J. B. <bil...@or...>]

Barry,

OK, thanks. But... why is the population of He[He] still decreasing 
(i.e. - f[xi].He[He] -= ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he])?

Jay

On 07/01/2013 02:03 PM, Barry Smith wrote:
>     Jay,
>
>      Sorry the initial comment is incorrect; probably I started that way and then changed the code and not the comment. It should
> read
>
>      /*He[He-he] + He[he] -> He[He] */
>
> On Jul 1, 2013, at 11:39 AM, Jay Jay Billings <jay...@gm...> wrote:
>
>> Barry,
>>
>> I am looking at the first part of IFunction in the ADR prototype and noticed something that seemed strange to me. Here's the part in question:
>>
>>      /* ----------------------------------------------------------------
>>       ---- Compute reaction terms that can create a cluster of given size
>>      */
>>      /*   He[He] + He[he] -> He[He+he]  */
>>      for (He=2; He<N+1; He++) {
>>        /* compute all pairs of clusters of smaller size that can combine to create a cluster of size He,
>>           remove the upper half since they are symmetric to the lower half of the pairs. For example
>>                when He = 5 (cluster size 5) the pairs are
>>                   1   4
>>                   2   2
>>                   3   2  these last two are not needed in the sum since they repeat from above
>>                   4   1  this is why he < (He/2) + 1            */
>>        for (he=1; he<(He/2)+1; he++) {
>>          f[xi].He[He] -= ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
>>
>>          /* remove the two clusters that merged to form the larger cluster */
>>          f[xi].He[he]    += ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
>>          f[xi].He[He-he] += ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
>>        }
>>      }
>>
>> In the inner most for loop, I can see why you decrease the concentration of He by the specified amount, but why are you increasing the concentration of he? Shouldn't it decrease too and the concentration of He+he increase? I also notice that you are increasing the population of He-he, but shouldn't it be He+he?
>>
>> It occurs to me that I could be counting incorrectly and if so, sorry for the trouble!
>>
>> Thanks,
>> Jay
>>
>> -- 
>> Jay Jay Billings
>> Oak Ridge National Laboratory
>> Twitter Handle: @jayjaybillings
>> ------------------------------------------------------------------------------
>> This SF.net email is sponsored by Windows:
>>
>> Build for Windows Store.
>>
>> http://p.sf.net/sfu/windows-dev2dev_______________________________________________
>> Xolotl-psi-development mailing list
>> Xol...@li...
>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
>
> Build for Windows Store.
>
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development


-- 
Jay Jay Billings
Oak Ridge National Laboratory
Telephone: (865) 272-9420
Email: bil...@or...
Twitter Handle: @jayjaybillings

Re: [Xolotl-psi-development] Figuring out parts of IFunction in the ADR prototype

[Barry S. <bs...@mc...>]

   Jay,

    Sorry the initial comment is incorrect; probably I started that way and then changed the code and not the comment. It should
read 

    /*He[He-he] + He[he] -> He[He] */

On Jul 1, 2013, at 11:39 AM, Jay Jay Billings <jay...@gm...> wrote:

> Barry,
> 
> I am looking at the first part of IFunction in the ADR prototype and noticed something that seemed strange to me. Here's the part in question:
> 
>     /* ----------------------------------------------------------------
>      ---- Compute reaction terms that can create a cluster of given size
>     */
>     /*   He[He] + He[he] -> He[He+he]  */
>     for (He=2; He<N+1; He++) {
>       /* compute all pairs of clusters of smaller size that can combine to create a cluster of size He,
>          remove the upper half since they are symmetric to the lower half of the pairs. For example
>               when He = 5 (cluster size 5) the pairs are
>                  1   4
>                  2   2
>                  3   2  these last two are not needed in the sum since they repeat from above
>                  4   1  this is why he < (He/2) + 1            */
>       for (he=1; he<(He/2)+1; he++) {
>         f[xi].He[He] -= ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
> 
>         /* remove the two clusters that merged to form the larger cluster */
>         f[xi].He[he]    += ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
>         f[xi].He[He-he] += ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
>       }
>     }
> 
> In the inner most for loop, I can see why you decrease the concentration of He by the specified amount, but why are you increasing the concentration of he? Shouldn't it decrease too and the concentration of He+he increase? I also notice that you are increasing the population of He-he, but shouldn't it be He+he?
> 
> It occurs to me that I could be counting incorrectly and if so, sorry for the trouble!
> 
> Thanks,
> Jay
> 
> -- 
> Jay Jay Billings
> Oak Ridge National Laboratory
> Twitter Handle: @jayjaybillings
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
> 
> Build for Windows Store.
> 
> http://p.sf.net/sfu/windows-dev2dev_______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development

[Xolotl-psi-development] Figuring out parts of IFunction in the ADR prototype

[Jay J. B. <jay...@gm...>]

Barry,

I am looking at the first part of IFunction in the ADR prototype and
noticed something that seemed strange to me. Here's the part in question:

    /* ----------------------------------------------------------------
     ---- Compute reaction terms that can create a cluster of given size
    */
    /*   He[He] + He[he] -> He[He+he]  */
    for (He=2; He<N+1; He++) {
      /* compute all pairs of clusters of smaller size that can combine to
create a cluster of size He,
         remove the upper half since they are symmetric to the lower half
of the pairs. For example
              when He = 5 (cluster size 5) the pairs are
                 1   4
                 2   2
                 3   2  these last two are not needed in the sum since they
repeat from above
                 4   1  this is why he < (He/2) + 1            */
      for (he=1; he<(He/2)+1; he++) {
        f[xi].He[He] -= ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];

        /* remove the two clusters that merged to form the larger cluster */
        f[xi].He[he]    += ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
        f[xi].He[He-he] += ctx->reactionScale*c[xi].He[he]*c[xi].He[He-he];
      }
    }

In the inner most for loop, I can see why you decrease the concentration of
He by the specified amount, but why are you increasing the concentration of
he? Shouldn't it decrease too and the concentration of He+he increase? I
also notice that you are increasing the population of He-he, but shouldn't
it be He+he?

It occurs to me that I could be counting incorrectly and if so, sorry for
the trouble!

Thanks,
Jay

-- 
Jay Jay Billings
Oak Ridge National Laboratory
Twitter Handle: @jayjaybillings

Re: [Xolotl-psi-development] Diagonal block vs. off-diagonal block in PETSc?

[Jay J. B. <jay...@gm...>]

Barry,

Thanks for getting back to me on this! I thought that was what they meant,
but I didn't want to make any assumptions.

I think I have covered the diagonal block sufficiently based on your
prototype and Tibo's notes. It will be up for discussion during tomorrow's
call.

Out of curiosity, why did you pick the names diagonal block and
off-diagonal block? Does it have something to do with the way PETSc uses
them in its solve?

Jay


On Sun, Jun 30, 2013 at 5:30 PM, Barry Smith <bs...@mc...> wrote:

>
>    Jay,
>
>      The "diagonal block" merely represents the couplings between
> different components AT a grid point (or cell). The off diagonal block
> represents coupling between components BETWEEN two neighboring grid cells.
>
>     In this project the diagonal block represents reactions and the
> off-diagonal block represents diffusion. So an entry in the diagonal block
> i,j means that component i reacts with or creates in a reaction component
> j.  An entry i,i in the off diagonal block represents diffusion of
> component i. (Note there is no entry i,j  i != j in the off diagonal block
> (diffusion) since each component diffuses independently.
>
>    I just made up the terms diagonal block and off diagonal block for this
> purpose many years ago, I don't know of standard terminology for it.
>
>    Barry
>
> In this project component means cluster type/size
>
>
> On Jun 30, 2013, at 3:03 PM, Jay Jay Billings <jay...@gm...>
> wrote:
>
> > Can someone explain the difference between the diagonal block and
> off-diagonal block in PETSc or point me to a mathematical description of
> the difference please? I am meaning, specifically, those arrays described at
> >
> >
> http://www.mcs.anl.gov/petsc/petsc-current/docs/manualpages/DM/DMDASetBlockFills.html
> >
> > I don't want to sound like an idiot when I talk about these tomorrow.
> >
> > Thanks,
> > Jay
> >
> > --
> > Jay Jay Billings
> > Oak Ridge National Laboratory
> > Twitter Handle: @jayjaybillings
> >
> ------------------------------------------------------------------------------
> > This SF.net email is sponsored by Windows:
> >
> > Build for Windows Store.
> >
> >
> http://p.sf.net/sfu/windows-dev2dev_______________________________________________
> > Xolotl-psi-development mailing list
> > Xol...@li...
> > https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>
>


-- 
Jay Jay Billings
Oak Ridge National Laboratory
Twitter Handle: @jayjaybillings

Re: [Xolotl-psi-development] Diagonal block vs. off-diagonal block in PETSc?

[Barry S. <bs...@mc...>]

   Jay,

     The "diagonal block" merely represents the couplings between different components AT a grid point (or cell). The off diagonal block represents coupling between components BETWEEN two neighboring grid cells. 

    In this project the diagonal block represents reactions and the off-diagonal block represents diffusion. So an entry in the diagonal block i,j means that component i reacts with or creates in a reaction component j.  An entry i,i in the off diagonal block represents diffusion of component i. (Note there is no entry i,j  i != j in the off diagonal block (diffusion) since each component diffuses independently. 

   I just made up the terms diagonal block and off diagonal block for this purpose many years ago, I don't know of standard terminology for it.

   Barry

In this project component means cluster type/size


On Jun 30, 2013, at 3:03 PM, Jay Jay Billings <jay...@gm...> wrote:

> Can someone explain the difference between the diagonal block and off-diagonal block in PETSc or point me to a mathematical description of the difference please? I am meaning, specifically, those arrays described at 
> 
> http://www.mcs.anl.gov/petsc/petsc-current/docs/manualpages/DM/DMDASetBlockFills.html
> 
> I don't want to sound like an idiot when I talk about these tomorrow. 
> 
> Thanks,
> Jay
> 
> -- 
> Jay Jay Billings
> Oak Ridge National Laboratory
> Twitter Handle: @jayjaybillings
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
> 
> Build for Windows Store.
> 
> http://p.sf.net/sfu/windows-dev2dev_______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development

[Xolotl-psi-development] Diagonal block vs. off-diagonal block in PETSc?

[Jay J. B. <jay...@gm...>]

Can someone explain the difference between the diagonal block and
off-diagonal block in PETSc or point me to a mathematical description of
the difference please? I am meaning, specifically, those arrays described
at

http://www.mcs.anl.gov/petsc/petsc-current/docs/manualpages/DM/DMDASetBlockFills.html

I don't want to sound like an idiot when I talk about these tomorrow.

Thanks,
Jay

-- 
Jay Jay Billings
Oak Ridge National Laboratory
Twitter Handle: @jayjaybillings

Re: [Xolotl-psi-development] Possible to MPI Broadcast with PETSc mixed in?

[Jay J. B. <jay...@gm...>]

Thanks, guys. I will broadcast it.

At least for the foreseeable future, it should be very small and homogenous
across the elements at start up.

Jay

-----
Jay Jay Billings
Computational Scientist
Oak Ridge National Laboratory
Twitter Handle: @jayjaybillings
On Jun 13, 2013 5:26 PM, "Tim Tautges" <ta...@mc...> wrote:

> Me too, though it also depends on what that data looks like.  Will it vary
> by element, vertex, or something else?
> Sometimes there'll be good reasons to associate it with the mesh,
> sometimes not, and as Barry mentions, the volume may
> not be high so bcasting it will work fine.
>
> - tim
>
> On 06/13/2013 04:06 PM, Barry Smith wrote:
> >
> >     Presumably the reaction network is small by parallel computing
> standards so I would have process 0 read it and broadcast it to all.
> >
> >     Barry
> >
> > On Jun 13, 2013, at 3:45 PM, Jay Jay Billings <bil...@or...>
> wrote:
> >
> >> Everyone,
> >>
> >> I'm trying to figure out the best way to broadcast the contents of the
> >> initial reaction network across the nodes in Xolotl.
> >>
> >> In the long run, I imagine putting the contents of the reaction network
> >> on the MOAB mesh when we generate it. However, for now, I just want to
> >> broadcast it to get the FD version working. I imagine simply packing up
> >> the structure, sending it out to all nodes and loading it onto the FD
> >> grid generated by PETSc. I'm thinking that all I need to do is pass it
> >> the correct communicator and then play catch, so to say.
> >>
> >> Barry mentioned using a PETSc Vec file too. However, when we get to the
> >> point of doing that I think it will be best to create a preprocessor
> >> that generates our mesh, puts the data on it and dumps it all to a MOAB
> >> file.
> >>
> >> One thing I could try since we are running small is just having every
> >> node read the file since reads are safe. This won't work for long,
> though.
> >>
> >> Is a plain, old MPI broadcast possible with PETSc in the mix, or should
> >> I consider something else?
> >>
> >> Jay
> >>
> >>
> ------------------------------------------------------------------------------
> >> This SF.net email is sponsored by Windows:
> >>
> >> Build for Windows Store.
> >>
> >> http://p.sf.net/sfu/windows-dev2dev
> >> _______________________________________________
> >> Xolotl-psi-development mailing list
> >> Xol...@li...
> >> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
> >
> >
> >
> ------------------------------------------------------------------------------
> > This SF.net email is sponsored by Windows:
> >
> > Build for Windows Store.
> >
> > http://p.sf.net/sfu/windows-dev2dev
> > _______________________________________________
> > Xolotl-psi-development mailing list
> > Xol...@li...
> > https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
> >
>
> --
> ================================================================
> "You will keep in perfect peace him whose mind is
>    steadfast, because he trusts in you."               Isaiah 26:3
>
>               Tim Tautges            Argonne National Laboratory
>           (ta...@mc...)      (telecommuting from UW-Madison)
>   phone (gvoice): (608) 354-1459      1500 Engineering Dr.
>              fax: (608) 263-4499      Madison, WI 53706
>
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
>
> Build for Windows Store.
>
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>

Re: [Xolotl-psi-development] Possible to MPI Broadcast with PETSc mixed in?

[Tim T. <ta...@mc...>]

Me too, though it also depends on what that data looks like.  Will it vary by element, vertex, or something else? 
Sometimes there'll be good reasons to associate it with the mesh, sometimes not, and as Barry mentions, the volume may 
not be high so bcasting it will work fine.

- tim

On 06/13/2013 04:06 PM, Barry Smith wrote:
>
>     Presumably the reaction network is small by parallel computing standards so I would have process 0 read it and broadcast it to all.
>
>     Barry
>
> On Jun 13, 2013, at 3:45 PM, Jay Jay Billings <bil...@or...> wrote:
>
>> Everyone,
>>
>> I'm trying to figure out the best way to broadcast the contents of the
>> initial reaction network across the nodes in Xolotl.
>>
>> In the long run, I imagine putting the contents of the reaction network
>> on the MOAB mesh when we generate it. However, for now, I just want to
>> broadcast it to get the FD version working. I imagine simply packing up
>> the structure, sending it out to all nodes and loading it onto the FD
>> grid generated by PETSc. I'm thinking that all I need to do is pass it
>> the correct communicator and then play catch, so to say.
>>
>> Barry mentioned using a PETSc Vec file too. However, when we get to the
>> point of doing that I think it will be best to create a preprocessor
>> that generates our mesh, puts the data on it and dumps it all to a MOAB
>> file.
>>
>> One thing I could try since we are running small is just having every
>> node read the file since reads are safe. This won't work for long, though.
>>
>> Is a plain, old MPI broadcast possible with PETSc in the mix, or should
>> I consider something else?
>>
>> Jay
>>
>> ------------------------------------------------------------------------------
>> This SF.net email is sponsored by Windows:
>>
>> Build for Windows Store.
>>
>> http://p.sf.net/sfu/windows-dev2dev
>> _______________________________________________
>> Xolotl-psi-development mailing list
>> Xol...@li...
>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
>
> Build for Windows Store.
>
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>

-- 
================================================================
"You will keep in perfect peace him whose mind is
   steadfast, because he trusts in you."               Isaiah 26:3

              Tim Tautges            Argonne National Laboratory
          (ta...@mc...)      (telecommuting from UW-Madison)
  phone (gvoice): (608) 354-1459      1500 Engineering Dr.
             fax: (608) 263-4499      Madison, WI 53706

Re: [Xolotl-psi-development] Possible to MPI Broadcast with PETSc mixed in?

[Barry S. <bs...@mc...>]

   Presumably the reaction network is small by parallel computing standards so I would have process 0 read it and broadcast it to all.

   Barry

On Jun 13, 2013, at 3:45 PM, Jay Jay Billings <bil...@or...> wrote:

> Everyone,
> 
> I'm trying to figure out the best way to broadcast the contents of the 
> initial reaction network across the nodes in Xolotl.
> 
> In the long run, I imagine putting the contents of the reaction network 
> on the MOAB mesh when we generate it. However, for now, I just want to 
> broadcast it to get the FD version working. I imagine simply packing up 
> the structure, sending it out to all nodes and loading it onto the FD 
> grid generated by PETSc. I'm thinking that all I need to do is pass it 
> the correct communicator and then play catch, so to say.
> 
> Barry mentioned using a PETSc Vec file too. However, when we get to the 
> point of doing that I think it will be best to create a preprocessor 
> that generates our mesh, puts the data on it and dumps it all to a MOAB 
> file.
> 
> One thing I could try since we are running small is just having every 
> node read the file since reads are safe. This won't work for long, though.
> 
> Is a plain, old MPI broadcast possible with PETSc in the mix, or should 
> I consider something else?
> 
> Jay
> 
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
> 
> Build for Windows Store.
> 
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development

[Xolotl-psi-development] Possible to MPI Broadcast with PETSc mixed in?

[Jay J. B. <bil...@or...>]

Everyone,

I'm trying to figure out the best way to broadcast the contents of the 
initial reaction network across the nodes in Xolotl.

In the long run, I imagine putting the contents of the reaction network 
on the MOAB mesh when we generate it. However, for now, I just want to 
broadcast it to get the FD version working. I imagine simply packing up 
the structure, sending it out to all nodes and loading it onto the FD 
grid generated by PETSc. I'm thinking that all I need to do is pass it 
the correct communicator and then play catch, so to say.

Barry mentioned using a PETSc Vec file too. However, when we get to the 
point of doing that I think it will be best to create a preprocessor 
that generates our mesh, puts the data on it and dumps it all to a MOAB 
file.

One thing I could try since we are running small is just having every 
node read the file since reads are safe. This won't work for long, though.

Is a plain, old MPI broadcast possible with PETSc in the mix, or should 
I consider something else?

Jay

Re: [Xolotl-psi-development] Testing const vs. constexpr

[Jay J. B. <bil...@or...>]

Barry,

Done & committed.

Jay

On 06/05/2013 10:52 PM, Barry Smith wrote:
> On Jun 5, 2013, at 4:11 PM, Jay Jay Billings <bil...@or...> wrote:
>
>> Barry,
>>
>> The data is generated automatically from Jed Brown's FindPETSc.cmake
>> file. Look in xolotl/cmake-modules/FindPETSc.cmake.
>>
>> Jay
>>
>     Jay,
>
>      No the problem is here in CMakeLists.txt
>
> # Setup the library list
> set(XOLOTL_LIBS ${PETSC_LIBRARIES} xolotlReactants xolotlSolver)
>
> # Add an executable for the main program
> ADD_EXECUTABLE ( xolotl main.cpp)
>
> # Link the reactants library
> target_link_libraries(xolotl ${XOLOTL_LIBS})
>
> You need to list the xolotlReactants xolotlSolver  BEFORE the ${PETSC_LIBRARIES}
>
> Why the order didn't matter in your versions of linux I cannot explain.
>
>     Barry
>
>
>
>> On 06/05/2013 05:09 PM, Barry Smith wrote:
>>> On Jun 5, 2013, at 3:21 PM, Jay Jay Billings <bil...@or...> wrote:
>>>
>>>> Barry,
>>>>
>>>> That file is not in the repository. It is auto-generated when you run
>>>> CMake in your build directory. I'm not sure how to fix that.
>>>      Where is the data provided to cmake to generate that particular auto-generated file. Possibly the ordering could be changed in that location and then cmake would generate the correct thing?
>>>
>>>     Barry
>>>
>>>> In your build directory, you should find an executable named "xolotl."
>>>> You can run that with...
>>>>
>>>> time ./xolotl ../benchmarks/tungsten.txt -ts_final_time 2000 -da_refine
>>>> 2 -log_summary -ts_adapt_dt_max 10 -ts_monitor
>>>>
>>>> with the location to the benchmark directory specified as fits your
>>>> system. It runs for about 58min on my desktop.
>>>>
>>>> The data in that benchmark file is not yet passed to the PETSc code. I
>>>> am using the values that you generate for now.
>>>>
>>>> Jay
>>>>
>>>> On 06/05/2013 03:49 PM, Barry Smith wrote:
>>>>>     Thanks. I ran and got
>>>>>
>>>>> Linking CXX executable xolotl
>>>>> /usr/bin/cmake -E cmake_link_script CMakeFiles/xolotl.dir/link.txt --verbose=1
>>>>> /usr/bin/c++   -std=c++0x     CMakeFiles/xolotl.dir/main.cpp.o  -o xolotl -rdynamic /sandbox/bsmith/petsc/arch-xolotl/lib/libpetsc.so xolotlCore/reactants/libxolotlReactants.so xolotlSolver/libxolotlSolver.so -Wl,-rpath,/sandbox/bsmith/petsc/arch-xolotl/lib:/sandbox/bsmith/trunk/xolotl/build/xolotlCore/reactants:/sandbox/bsmith/trunk/xolotl/build/xolotlSolver
>>>>> xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
>>>>> …
>>>>>
>>>>> so I edited cmake_link_script CMakeFiles/xolotl.dir/link.txt  to list /sandbox/bsmith/petsc/arch-xolotl/lib/libpetsc.so AFTER the other two .so  then the link went through and all the compiles/links worked (haven't actually run anything).
>>>>>
>>>>>     So it looks like the order of those libraries is important and the PETSc one (since the others depended on it) must be listed last.
>>>>>
>>>>>     Will you fix the repository so I can test again without my editing hack?
>>>>>
>>>>>     Also how do I run something?
>>>>>
>>>>>     Barry
>>>>>
>>>>>
>>>>>
>>>>> On Jun 5, 2013, at 1:31 PM, Jay Jay Billings <bil...@or...> wrote:
>>>>>
>>>>>> Barry,
>>>>>>
>>>>>> Do
>>>>>>
>>>>>> make VERBOSE=1
>>>>>>
>>>>>> Jay
>>>>>>
>>>>>> On 06/05/2013 02:10 PM, Barry Smith wrote:
>>>>>>> On Jun 5, 2013, at 11:35 AM, "Billings, Jay Jay" <bil...@or...> wrote:
>>>>>>>
>>>>>>>> Tim and Barry,
>>>>>>>>
>>>>>>>> I was able to make a fix along the lines that Tim suggested. You should be able to build it now.
>>>>>>>     That fixed that problem but I have others.
>>>>>>>
>>>>>>>     How can I get cmake to actually print what it is doing when I use make?
>>>>>>>
>>>>>>>     I would like to see the commands it is trying to use so I can debug it.  I have set LD_LIBRARY_PATH  but somehow it is not finding or linking in the PETSc library. It says "Linking CXX executable xolotl" but it would useful if it showed the exact link line.
>>>>>>>
>>>>>>>     Barry
>>>>>>>
>>>>>>>
>>>>>>> # fast build rule for target.
>>>>>>> xolotl/fast:
>>>>>>>   $(MAKE) -f CMakeFiles/xolotl.dir/build.make CMakeFiles/xolotl.dir/build
>>>>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ emacs -nw Makefile
>>>>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ make
>>>>>>> [  6%] Built target xolotlSolver
>>>>>>> [ 73%] Built target xolotlReactants
>>>>>>> Linking CXX executable xolotl
>>>>>>> xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
>>>>>>> …..
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>> Please let me know if you encounter more errors.
>>>>>>>>
>>>>>>>> I'll try to fix the compiler identification bug next.
>>>>>>>>
>>>>>>>> Jay
>>>>>>>> ________________________________________
>>>>>>>> From: Billings, Jay Jay [bil...@or...]
>>>>>>>> Sent: Wednesday, June 05, 2013 12:28 PM
>>>>>>>> To: xol...@li...
>>>>>>>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>>>>>>>
>>>>>>>> Tim,
>>>>>>>>
>>>>>>>> That could work. I'm initializing it in a class now, but let me try this out.
>>>>>>>>
>>>>>>>> Jay
>>>>>>>> ________________________________________
>>>>>>>> From: Tim Tautges [ta...@mc...]
>>>>>>>> Sent: Wednesday, June 05, 2013 12:26 PM
>>>>>>>> To: xol...@li...
>>>>>>>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>>>>>>>
>>>>>>>> Why not initialize at the top of the .cpp file, e.g.
>>>>>>>>
>>>>>>>> const double xolotlCore::PSICluster::kBoltzmann = 1.0e4;
>>>>>>>>
>>>>>>>> or something like that?  I haven't looked at the actual code, so maybe you're already doing that...
>>>>>>>>
>>>>>>>> - tim
>>>>>>>>
>>>>>>>> On 06/05/2013 11:10 AM, Jay Jay Billings wrote:
>>>>>>>>> Barry,
>>>>>>>>>
>>>>>>>>> Can you try a fix for me please and let me know if it works? Can you change the word "const" in Constants.h to "constexpr" and let me know if that fixes the bug you reported below?
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> In file included from /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
>>>>>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error: ‘constexpr’ needed for in-class initialization of static data member ‘kBoltzmann’ of non-integral type
>>>>>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp: In member function ‘virtual double xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
>>>>>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15: error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
>>>>>>>>> make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o] Error 1
>>>>>>>>> make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
>>>>>>>>> make: *** [all] Error 2
>>>>>>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> This is a C++ vs. C++11 issue. We may need to declare a minimum compiler version for Xolotl or define the constants with
>>>>>>>>> macros.
>>>>>>>>>
>>>>>>>>> Jay
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>> Jay Jay Billings
>>>>>>>>> Oak Ridge National Laboratory
>>>>>>>>> Twitter Handle: @jayjaybillings
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>>>>> 3. A single system of record for all IT processes
>>>>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> _______________________________________________
>>>>>>>>> Xolotl-psi-development mailing list
>>>>>>>>> Xol...@li...
>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>>>>>>
>>>>>>>> --
>>>>>>>> ================================================================
>>>>>>>> "You will keep in perfect peace him whose mind is
>>>>>>>>    steadfast, because he trusts in you."               Isaiah 26:3
>>>>>>>>
>>>>>>>>               Tim Tautges            Argonne National Laboratory
>>>>>>>>           (ta...@mc...)      (telecommuting from UW-Madison)
>>>>>>>>   phone (gvoice): (608) 354-1459      1500 Engineering Dr.
>>>>>>>>              fax: (608) 263-4499      Madison, WI 53706
>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>>>> 3. A single system of record for all IT processes
>>>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>>>> _______________________________________________
>>>>>>>> Xolotl-psi-development mailing list
>>>>>>>> Xol...@li...
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>>>> 3. A single system of record for all IT processes
>>>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>>>> _______________________________________________
>>>>>>>> Xolotl-psi-development mailing list
>>>>>>>> Xol...@li...
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>>>> 3. A single system of record for all IT processes
>>>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>>>> _______________________________________________
>>>>>>>> Xolotl-psi-development mailing list
>>>>>>>> Xol...@li...
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>>> --
>>>>>> Jay Jay Billings
>>>>>> Oak Ridge National Laboratory
>>>>>> Telephone: (865) 272-9420
>>>>>> Email: bil...@or...
>>>>>> Twitter Handle: @jayjaybillings
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>> 3. A single system of record for all IT processes
>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>> _______________________________________________
>>>>>> Xolotl-psi-development mailing list
>>>>>> Xol...@li...
>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>> --
>>>> Jay Jay Billings
>>>> Oak Ridge National Laboratory
>>>> Telephone: (865) 272-9420
>>>> Email: bil...@or...
>>>> Twitter Handle: @jayjaybillings
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> How ServiceNow helps IT people transform IT departments:
>>>> 1. A cloud service to automate IT design, transition and operations
>>>> 2. Dashboards that offer high-level views of enterprise services
>>>> 3. A single system of record for all IT processes
>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>> _______________________________________________
>>>> Xolotl-psi-development mailing list
>>>> Xol...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>
>> --
>> Jay Jay Billings
>> Oak Ridge National Laboratory
>> Telephone: (865) 272-9420
>> Email: bil...@or...
>> Twitter Handle: @jayjaybillings
>>
>>
>> ------------------------------------------------------------------------------
>> How ServiceNow helps IT people transform IT departments:
>> 1. A cloud service to automate IT design, transition and operations
>> 2. Dashboards that offer high-level views of enterprise services
>> 3. A single system of record for all IT processes
>> http://p.sf.net/sfu/servicenow-d2d-j
>> _______________________________________________
>> Xolotl-psi-development mailing list
>> Xol...@li...
>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development


-- 
Jay Jay Billings
Oak Ridge National Laboratory
Telephone: (865) 272-9420
Email: bil...@or...
Twitter Handle: @jayjaybillings

Re: [Xolotl-psi-development] Testing const vs. constexpr

[Barry S. <bs...@mc...>]

On Jun 5, 2013, at 4:11 PM, Jay Jay Billings <bil...@or...> wrote:

> Barry,
> 
> The data is generated automatically from Jed Brown's FindPETSc.cmake 
> file. Look in xolotl/cmake-modules/FindPETSc.cmake.
> 
> Jay
> 
   Jay,

    No the problem is here in CMakeLists.txt 

# Setup the library list
set(XOLOTL_LIBS ${PETSC_LIBRARIES} xolotlReactants xolotlSolver)

# Add an executable for the main program
ADD_EXECUTABLE ( xolotl main.cpp)

# Link the reactants library
target_link_libraries(xolotl ${XOLOTL_LIBS})

You need to list the xolotlReactants xolotlSolver  BEFORE the ${PETSC_LIBRARIES} 

Why the order didn't matter in your versions of linux I cannot explain.

   Barry



> On 06/05/2013 05:09 PM, Barry Smith wrote:
>> On Jun 5, 2013, at 3:21 PM, Jay Jay Billings <bil...@or...> wrote:
>> 
>>> Barry,
>>> 
>>> That file is not in the repository. It is auto-generated when you run
>>> CMake in your build directory. I'm not sure how to fix that.
>>     Where is the data provided to cmake to generate that particular auto-generated file. Possibly the ordering could be changed in that location and then cmake would generate the correct thing?
>> 
>>    Barry
>> 
>>> In your build directory, you should find an executable named "xolotl."
>>> You can run that with...
>>> 
>>> time ./xolotl ../benchmarks/tungsten.txt -ts_final_time 2000 -da_refine
>>> 2 -log_summary -ts_adapt_dt_max 10 -ts_monitor
>>> 
>>> with the location to the benchmark directory specified as fits your
>>> system. It runs for about 58min on my desktop.
>>> 
>>> The data in that benchmark file is not yet passed to the PETSc code. I
>>> am using the values that you generate for now.
>>> 
>>> Jay
>>> 
>>> On 06/05/2013 03:49 PM, Barry Smith wrote:
>>>>    Thanks. I ran and got
>>>> 
>>>> Linking CXX executable xolotl
>>>> /usr/bin/cmake -E cmake_link_script CMakeFiles/xolotl.dir/link.txt --verbose=1
>>>> /usr/bin/c++   -std=c++0x     CMakeFiles/xolotl.dir/main.cpp.o  -o xolotl -rdynamic /sandbox/bsmith/petsc/arch-xolotl/lib/libpetsc.so xolotlCore/reactants/libxolotlReactants.so xolotlSolver/libxolotlSolver.so -Wl,-rpath,/sandbox/bsmith/petsc/arch-xolotl/lib:/sandbox/bsmith/trunk/xolotl/build/xolotlCore/reactants:/sandbox/bsmith/trunk/xolotl/build/xolotlSolver
>>>> xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
>>>> …
>>>> 
>>>> so I edited cmake_link_script CMakeFiles/xolotl.dir/link.txt  to list /sandbox/bsmith/petsc/arch-xolotl/lib/libpetsc.so AFTER the other two .so  then the link went through and all the compiles/links worked (haven't actually run anything).
>>>> 
>>>>    So it looks like the order of those libraries is important and the PETSc one (since the others depended on it) must be listed last.
>>>> 
>>>>    Will you fix the repository so I can test again without my editing hack?
>>>> 
>>>>    Also how do I run something?
>>>> 
>>>>    Barry
>>>> 
>>>> 
>>>> 
>>>> On Jun 5, 2013, at 1:31 PM, Jay Jay Billings <bil...@or...> wrote:
>>>> 
>>>>> Barry,
>>>>> 
>>>>> Do
>>>>> 
>>>>> make VERBOSE=1
>>>>> 
>>>>> Jay
>>>>> 
>>>>> On 06/05/2013 02:10 PM, Barry Smith wrote:
>>>>>> On Jun 5, 2013, at 11:35 AM, "Billings, Jay Jay" <bil...@or...> wrote:
>>>>>> 
>>>>>>> Tim and Barry,
>>>>>>> 
>>>>>>> I was able to make a fix along the lines that Tim suggested. You should be able to build it now.
>>>>>>    That fixed that problem but I have others.
>>>>>> 
>>>>>>    How can I get cmake to actually print what it is doing when I use make?
>>>>>> 
>>>>>>    I would like to see the commands it is trying to use so I can debug it.  I have set LD_LIBRARY_PATH  but somehow it is not finding or linking in the PETSc library. It says "Linking CXX executable xolotl" but it would useful if it showed the exact link line.
>>>>>> 
>>>>>>    Barry
>>>>>> 
>>>>>> 
>>>>>> # fast build rule for target.
>>>>>> xolotl/fast:
>>>>>> 	$(MAKE) -f CMakeFiles/xolotl.dir/build.make CMakeFiles/xolotl.dir/build
>>>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ emacs -nw Makefile
>>>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ make
>>>>>> [  6%] Built target xolotlSolver
>>>>>> [ 73%] Built target xolotlReactants
>>>>>> Linking CXX executable xolotl
>>>>>> xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
>>>>>> …..
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>>> Please let me know if you encounter more errors.
>>>>>>> 
>>>>>>> I'll try to fix the compiler identification bug next.
>>>>>>> 
>>>>>>> Jay
>>>>>>> ________________________________________
>>>>>>> From: Billings, Jay Jay [bil...@or...]
>>>>>>> Sent: Wednesday, June 05, 2013 12:28 PM
>>>>>>> To: xol...@li...
>>>>>>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>>>>>> 
>>>>>>> Tim,
>>>>>>> 
>>>>>>> That could work. I'm initializing it in a class now, but let me try this out.
>>>>>>> 
>>>>>>> Jay
>>>>>>> ________________________________________
>>>>>>> From: Tim Tautges [ta...@mc...]
>>>>>>> Sent: Wednesday, June 05, 2013 12:26 PM
>>>>>>> To: xol...@li...
>>>>>>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>>>>>> 
>>>>>>> Why not initialize at the top of the .cpp file, e.g.
>>>>>>> 
>>>>>>> const double xolotlCore::PSICluster::kBoltzmann = 1.0e4;
>>>>>>> 
>>>>>>> or something like that?  I haven't looked at the actual code, so maybe you're already doing that...
>>>>>>> 
>>>>>>> - tim
>>>>>>> 
>>>>>>> On 06/05/2013 11:10 AM, Jay Jay Billings wrote:
>>>>>>>> Barry,
>>>>>>>> 
>>>>>>>> Can you try a fix for me please and let me know if it works? Can you change the word "const" in Constants.h to "constexpr" and let me know if that fixes the bug you reported below?
>>>>>>>> 
>>>>>>>> 
>>>>>>>> In file included from /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
>>>>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error: ‘constexpr’ needed for in-class initialization of static data member ‘kBoltzmann’ of non-integral type
>>>>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp: In member function ‘virtual double xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
>>>>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15: error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
>>>>>>>> make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o] Error 1
>>>>>>>> make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
>>>>>>>> make: *** [all] Error 2
>>>>>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version
>>>>>>>> 
>>>>>>>> 
>>>>>>>> This is a C++ vs. C++11 issue. We may need to declare a minimum compiler version for Xolotl or define the constants with
>>>>>>>> macros.
>>>>>>>> 
>>>>>>>> Jay
>>>>>>>> 
>>>>>>>> 
>>>>>>>> --
>>>>>>>> Jay Jay Billings
>>>>>>>> Oak Ridge National Laboratory
>>>>>>>> Twitter Handle: @jayjaybillings
>>>>>>>> 
>>>>>>>> 
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>>>> 3. A single system of record for all IT processes
>>>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>>>> 
>>>>>>>> 
>>>>>>>> 
>>>>>>>> _______________________________________________
>>>>>>>> Xolotl-psi-development mailing list
>>>>>>>> Xol...@li...
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>>>>> 
>>>>>>> --
>>>>>>> ================================================================
>>>>>>> "You will keep in perfect peace him whose mind is
>>>>>>>   steadfast, because he trusts in you."               Isaiah 26:3
>>>>>>> 
>>>>>>>              Tim Tautges            Argonne National Laboratory
>>>>>>>          (ta...@mc...)      (telecommuting from UW-Madison)
>>>>>>>  phone (gvoice): (608) 354-1459      1500 Engineering Dr.
>>>>>>>             fax: (608) 263-4499      Madison, WI 53706
>>>>>>> 
>>>>>>> 
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>>> 3. A single system of record for all IT processes
>>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>>> _______________________________________________
>>>>>>> Xolotl-psi-development mailing list
>>>>>>> Xol...@li...
>>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>>>> 
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>>> 3. A single system of record for all IT processes
>>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>>> _______________________________________________
>>>>>>> Xolotl-psi-development mailing list
>>>>>>> Xol...@li...
>>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>>>> 
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>>> 3. A single system of record for all IT processes
>>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>>> _______________________________________________
>>>>>>> Xolotl-psi-development mailing list
>>>>>>> Xol...@li...
>>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>> -- 
>>>>> Jay Jay Billings
>>>>> Oak Ridge National Laboratory
>>>>> Telephone: (865) 272-9420
>>>>> Email: bil...@or...
>>>>> Twitter Handle: @jayjaybillings
>>>>> 
>>>>> 
>>>>> ------------------------------------------------------------------------------
>>>>> How ServiceNow helps IT people transform IT departments:
>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>> 3. A single system of record for all IT processes
>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>> _______________________________________________
>>>>> Xolotl-psi-development mailing list
>>>>> Xol...@li...
>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>> 
>>> -- 
>>> Jay Jay Billings
>>> Oak Ridge National Laboratory
>>> Telephone: (865) 272-9420
>>> Email: bil...@or...
>>> Twitter Handle: @jayjaybillings
>>> 
>>> 
>>> ------------------------------------------------------------------------------
>>> How ServiceNow helps IT people transform IT departments:
>>> 1. A cloud service to automate IT design, transition and operations
>>> 2. Dashboards that offer high-level views of enterprise services
>>> 3. A single system of record for all IT processes
>>> http://p.sf.net/sfu/servicenow-d2d-j
>>> _______________________________________________
>>> Xolotl-psi-development mailing list
>>> Xol...@li...
>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
> 
> 
> -- 
> Jay Jay Billings
> Oak Ridge National Laboratory
> Telephone: (865) 272-9420
> Email: bil...@or...
> Twitter Handle: @jayjaybillings
> 
> 
> ------------------------------------------------------------------------------
> How ServiceNow helps IT people transform IT departments:
> 1. A cloud service to automate IT design, transition and operations
> 2. Dashboards that offer high-level views of enterprise services
> 3. A single system of record for all IT processes
> http://p.sf.net/sfu/servicenow-d2d-j
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development

Re: [Xolotl-psi-development] Testing const vs. constexpr

[Jay J. B. <bil...@or...>]

Barry,

The data is generated automatically from Jed Brown's FindPETSc.cmake 
file. Look in xolotl/cmake-modules/FindPETSc.cmake.

Jay

On 06/05/2013 05:09 PM, Barry Smith wrote:
> On Jun 5, 2013, at 3:21 PM, Jay Jay Billings <bil...@or...> wrote:
>
>> Barry,
>>
>> That file is not in the repository. It is auto-generated when you run
>> CMake in your build directory. I'm not sure how to fix that.
>      Where is the data provided to cmake to generate that particular auto-generated file. Possibly the ordering could be changed in that location and then cmake would generate the correct thing?
>
>     Barry
>
>> In your build directory, you should find an executable named "xolotl."
>> You can run that with...
>>
>> time ./xolotl ../benchmarks/tungsten.txt -ts_final_time 2000 -da_refine
>> 2 -log_summary -ts_adapt_dt_max 10 -ts_monitor
>>
>> with the location to the benchmark directory specified as fits your
>> system. It runs for about 58min on my desktop.
>>
>> The data in that benchmark file is not yet passed to the PETSc code. I
>> am using the values that you generate for now.
>>
>> Jay
>>
>> On 06/05/2013 03:49 PM, Barry Smith wrote:
>>>     Thanks. I ran and got
>>>
>>> Linking CXX executable xolotl
>>> /usr/bin/cmake -E cmake_link_script CMakeFiles/xolotl.dir/link.txt --verbose=1
>>> /usr/bin/c++   -std=c++0x     CMakeFiles/xolotl.dir/main.cpp.o  -o xolotl -rdynamic /sandbox/bsmith/petsc/arch-xolotl/lib/libpetsc.so xolotlCore/reactants/libxolotlReactants.so xolotlSolver/libxolotlSolver.so -Wl,-rpath,/sandbox/bsmith/petsc/arch-xolotl/lib:/sandbox/bsmith/trunk/xolotl/build/xolotlCore/reactants:/sandbox/bsmith/trunk/xolotl/build/xolotlSolver
>>> xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
>>> …
>>>
>>> so I edited cmake_link_script CMakeFiles/xolotl.dir/link.txt  to list /sandbox/bsmith/petsc/arch-xolotl/lib/libpetsc.so AFTER the other two .so  then the link went through and all the compiles/links worked (haven't actually run anything).
>>>
>>>     So it looks like the order of those libraries is important and the PETSc one (since the others depended on it) must be listed last.
>>>
>>>     Will you fix the repository so I can test again without my editing hack?
>>>
>>>     Also how do I run something?
>>>
>>>     Barry
>>>
>>>
>>>
>>> On Jun 5, 2013, at 1:31 PM, Jay Jay Billings <bil...@or...> wrote:
>>>
>>>> Barry,
>>>>
>>>> Do
>>>>
>>>> make VERBOSE=1
>>>>
>>>> Jay
>>>>
>>>> On 06/05/2013 02:10 PM, Barry Smith wrote:
>>>>> On Jun 5, 2013, at 11:35 AM, "Billings, Jay Jay" <bil...@or...> wrote:
>>>>>
>>>>>> Tim and Barry,
>>>>>>
>>>>>> I was able to make a fix along the lines that Tim suggested. You should be able to build it now.
>>>>>     That fixed that problem but I have others.
>>>>>
>>>>>     How can I get cmake to actually print what it is doing when I use make?
>>>>>
>>>>>     I would like to see the commands it is trying to use so I can debug it.  I have set LD_LIBRARY_PATH  but somehow it is not finding or linking in the PETSc library. It says "Linking CXX executable xolotl" but it would useful if it showed the exact link line.
>>>>>
>>>>>     Barry
>>>>>
>>>>>
>>>>> # fast build rule for target.
>>>>> xolotl/fast:
>>>>> 	$(MAKE) -f CMakeFiles/xolotl.dir/build.make CMakeFiles/xolotl.dir/build
>>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ emacs -nw Makefile
>>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ make
>>>>> [  6%] Built target xolotlSolver
>>>>> [ 73%] Built target xolotlReactants
>>>>> Linking CXX executable xolotl
>>>>> xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
>>>>> …..
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>> Please let me know if you encounter more errors.
>>>>>>
>>>>>> I'll try to fix the compiler identification bug next.
>>>>>>
>>>>>> Jay
>>>>>> ________________________________________
>>>>>> From: Billings, Jay Jay [bil...@or...]
>>>>>> Sent: Wednesday, June 05, 2013 12:28 PM
>>>>>> To: xol...@li...
>>>>>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>>>>>
>>>>>> Tim,
>>>>>>
>>>>>> That could work. I'm initializing it in a class now, but let me try this out.
>>>>>>
>>>>>> Jay
>>>>>> ________________________________________
>>>>>> From: Tim Tautges [ta...@mc...]
>>>>>> Sent: Wednesday, June 05, 2013 12:26 PM
>>>>>> To: xol...@li...
>>>>>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>>>>>
>>>>>> Why not initialize at the top of the .cpp file, e.g.
>>>>>>
>>>>>> const double xolotlCore::PSICluster::kBoltzmann = 1.0e4;
>>>>>>
>>>>>> or something like that?  I haven't looked at the actual code, so maybe you're already doing that...
>>>>>>
>>>>>> - tim
>>>>>>
>>>>>> On 06/05/2013 11:10 AM, Jay Jay Billings wrote:
>>>>>>> Barry,
>>>>>>>
>>>>>>> Can you try a fix for me please and let me know if it works? Can you change the word "const" in Constants.h to "constexpr" and let me know if that fixes the bug you reported below?
>>>>>>>
>>>>>>>
>>>>>>> In file included from /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
>>>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error: ‘constexpr’ needed for in-class initialization of static data member ‘kBoltzmann’ of non-integral type
>>>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp: In member function ‘virtual double xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
>>>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15: error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
>>>>>>> make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o] Error 1
>>>>>>> make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
>>>>>>> make: *** [all] Error 2
>>>>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version
>>>>>>>
>>>>>>>
>>>>>>> This is a C++ vs. C++11 issue. We may need to declare a minimum compiler version for Xolotl or define the constants with
>>>>>>> macros.
>>>>>>>
>>>>>>> Jay
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Jay Jay Billings
>>>>>>> Oak Ridge National Laboratory
>>>>>>> Twitter Handle: @jayjaybillings
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>>> 3. A single system of record for all IT processes
>>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Xolotl-psi-development mailing list
>>>>>>> Xol...@li...
>>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>>>>
>>>>>> --
>>>>>> ================================================================
>>>>>> "You will keep in perfect peace him whose mind is
>>>>>>    steadfast, because he trusts in you."               Isaiah 26:3
>>>>>>
>>>>>>               Tim Tautges            Argonne National Laboratory
>>>>>>           (ta...@mc...)      (telecommuting from UW-Madison)
>>>>>>   phone (gvoice): (608) 354-1459      1500 Engineering Dr.
>>>>>>              fax: (608) 263-4499      Madison, WI 53706
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>> 3. A single system of record for all IT processes
>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>> _______________________________________________
>>>>>> Xolotl-psi-development mailing list
>>>>>> Xol...@li...
>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>> 3. A single system of record for all IT processes
>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>> _______________________________________________
>>>>>> Xolotl-psi-development mailing list
>>>>>> Xol...@li...
>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>> 3. A single system of record for all IT processes
>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>> _______________________________________________
>>>>>> Xolotl-psi-development mailing list
>>>>>> Xol...@li...
>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>> -- 
>>>> Jay Jay Billings
>>>> Oak Ridge National Laboratory
>>>> Telephone: (865) 272-9420
>>>> Email: bil...@or...
>>>> Twitter Handle: @jayjaybillings
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> How ServiceNow helps IT people transform IT departments:
>>>> 1. A cloud service to automate IT design, transition and operations
>>>> 2. Dashboards that offer high-level views of enterprise services
>>>> 3. A single system of record for all IT processes
>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>> _______________________________________________
>>>> Xolotl-psi-development mailing list
>>>> Xol...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>
>> -- 
>> Jay Jay Billings
>> Oak Ridge National Laboratory
>> Telephone: (865) 272-9420
>> Email: bil...@or...
>> Twitter Handle: @jayjaybillings
>>
>>
>> ------------------------------------------------------------------------------
>> How ServiceNow helps IT people transform IT departments:
>> 1. A cloud service to automate IT design, transition and operations
>> 2. Dashboards that offer high-level views of enterprise services
>> 3. A single system of record for all IT processes
>> http://p.sf.net/sfu/servicenow-d2d-j
>> _______________________________________________
>> Xolotl-psi-development mailing list
>> Xol...@li...
>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development


-- 
Jay Jay Billings
Oak Ridge National Laboratory
Telephone: (865) 272-9420
Email: bil...@or...
Twitter Handle: @jayjaybillings

Re: [Xolotl-psi-development] Testing const vs. constexpr

[Barry S. <bs...@mc...>]

On Jun 5, 2013, at 3:21 PM, Jay Jay Billings <bil...@or...> wrote:

> Barry,
> 
> That file is not in the repository. It is auto-generated when you run 
> CMake in your build directory. I'm not sure how to fix that.

    Where is the data provided to cmake to generate that particular auto-generated file. Possibly the ordering could be changed in that location and then cmake would generate the correct thing?

   Barry

> 
> In your build directory, you should find an executable named "xolotl." 
> You can run that with...
> 
> time ./xolotl ../benchmarks/tungsten.txt -ts_final_time 2000 -da_refine 
> 2 -log_summary -ts_adapt_dt_max 10 -ts_monitor
> 
> with the location to the benchmark directory specified as fits your 
> system. It runs for about 58min on my desktop.
> 
> The data in that benchmark file is not yet passed to the PETSc code. I 
> am using the values that you generate for now.
> 
> Jay
> 
> On 06/05/2013 03:49 PM, Barry Smith wrote:
>>    Thanks. I ran and got
>> 
>> Linking CXX executable xolotl
>> /usr/bin/cmake -E cmake_link_script CMakeFiles/xolotl.dir/link.txt --verbose=1
>> /usr/bin/c++   -std=c++0x     CMakeFiles/xolotl.dir/main.cpp.o  -o xolotl -rdynamic /sandbox/bsmith/petsc/arch-xolotl/lib/libpetsc.so xolotlCore/reactants/libxolotlReactants.so xolotlSolver/libxolotlSolver.so -Wl,-rpath,/sandbox/bsmith/petsc/arch-xolotl/lib:/sandbox/bsmith/trunk/xolotl/build/xolotlCore/reactants:/sandbox/bsmith/trunk/xolotl/build/xolotlSolver
>> xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
>> …
>> 
>> so I edited cmake_link_script CMakeFiles/xolotl.dir/link.txt  to list /sandbox/bsmith/petsc/arch-xolotl/lib/libpetsc.so AFTER the other two .so  then the link went through and all the compiles/links worked (haven't actually run anything).
>> 
>>    So it looks like the order of those libraries is important and the PETSc one (since the others depended on it) must be listed last.
>> 
>>    Will you fix the repository so I can test again without my editing hack?
>> 
>>    Also how do I run something?
>> 
>>    Barry
>> 
>> 
>> 
>> On Jun 5, 2013, at 1:31 PM, Jay Jay Billings <bil...@or...> wrote:
>> 
>>> Barry,
>>> 
>>> Do
>>> 
>>> make VERBOSE=1
>>> 
>>> Jay
>>> 
>>> On 06/05/2013 02:10 PM, Barry Smith wrote:
>>>> On Jun 5, 2013, at 11:35 AM, "Billings, Jay Jay" <bil...@or...> wrote:
>>>> 
>>>>> Tim and Barry,
>>>>> 
>>>>> I was able to make a fix along the lines that Tim suggested. You should be able to build it now.
>>>>    That fixed that problem but I have others.
>>>> 
>>>>    How can I get cmake to actually print what it is doing when I use make?
>>>> 
>>>>    I would like to see the commands it is trying to use so I can debug it.  I have set LD_LIBRARY_PATH  but somehow it is not finding or linking in the PETSc library. It says "Linking CXX executable xolotl" but it would useful if it showed the exact link line.
>>>> 
>>>>    Barry
>>>> 
>>>> 
>>>> # fast build rule for target.
>>>> xolotl/fast:
>>>> 	$(MAKE) -f CMakeFiles/xolotl.dir/build.make CMakeFiles/xolotl.dir/build
>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ emacs -nw Makefile
>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ make
>>>> [  6%] Built target xolotlSolver
>>>> [ 73%] Built target xolotlReactants
>>>> Linking CXX executable xolotl
>>>> xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
>>>> …..
>>>> 
>>>> 
>>>> 
>>>> 
>>>>> Please let me know if you encounter more errors.
>>>>> 
>>>>> I'll try to fix the compiler identification bug next.
>>>>> 
>>>>> Jay
>>>>> ________________________________________
>>>>> From: Billings, Jay Jay [bil...@or...]
>>>>> Sent: Wednesday, June 05, 2013 12:28 PM
>>>>> To: xol...@li...
>>>>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>>>> 
>>>>> Tim,
>>>>> 
>>>>> That could work. I'm initializing it in a class now, but let me try this out.
>>>>> 
>>>>> Jay
>>>>> ________________________________________
>>>>> From: Tim Tautges [ta...@mc...]
>>>>> Sent: Wednesday, June 05, 2013 12:26 PM
>>>>> To: xol...@li...
>>>>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>>>> 
>>>>> Why not initialize at the top of the .cpp file, e.g.
>>>>> 
>>>>> const double xolotlCore::PSICluster::kBoltzmann = 1.0e4;
>>>>> 
>>>>> or something like that?  I haven't looked at the actual code, so maybe you're already doing that...
>>>>> 
>>>>> - tim
>>>>> 
>>>>> On 06/05/2013 11:10 AM, Jay Jay Billings wrote:
>>>>>> Barry,
>>>>>> 
>>>>>> Can you try a fix for me please and let me know if it works? Can you change the word "const" in Constants.h to "constexpr" and let me know if that fixes the bug you reported below?
>>>>>> 
>>>>>> 
>>>>>> In file included from /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
>>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error: ‘constexpr’ needed for in-class initialization of static data member ‘kBoltzmann’ of non-integral type
>>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp: In member function ‘virtual double xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
>>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15: error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
>>>>>> make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o] Error 1
>>>>>> make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
>>>>>> make: *** [all] Error 2
>>>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version
>>>>>> 
>>>>>> 
>>>>>> This is a C++ vs. C++11 issue. We may need to declare a minimum compiler version for Xolotl or define the constants with
>>>>>> macros.
>>>>>> 
>>>>>> Jay
>>>>>> 
>>>>>> 
>>>>>> --
>>>>>> Jay Jay Billings
>>>>>> Oak Ridge National Laboratory
>>>>>> Twitter Handle: @jayjaybillings
>>>>>> 
>>>>>> 
>>>>>> ------------------------------------------------------------------------------
>>>>>> How ServiceNow helps IT people transform IT departments:
>>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>>> 3. A single system of record for all IT processes
>>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> _______________________________________________
>>>>>> Xolotl-psi-development mailing list
>>>>>> Xol...@li...
>>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>>> 
>>>>> --
>>>>> ================================================================
>>>>> "You will keep in perfect peace him whose mind is
>>>>>   steadfast, because he trusts in you."               Isaiah 26:3
>>>>> 
>>>>>              Tim Tautges            Argonne National Laboratory
>>>>>          (ta...@mc...)      (telecommuting from UW-Madison)
>>>>>  phone (gvoice): (608) 354-1459      1500 Engineering Dr.
>>>>>             fax: (608) 263-4499      Madison, WI 53706
>>>>> 
>>>>> 
>>>>> ------------------------------------------------------------------------------
>>>>> How ServiceNow helps IT people transform IT departments:
>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>> 3. A single system of record for all IT processes
>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>> _______________________________________________
>>>>> Xolotl-psi-development mailing list
>>>>> Xol...@li...
>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>> 
>>>>> ------------------------------------------------------------------------------
>>>>> How ServiceNow helps IT people transform IT departments:
>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>> 3. A single system of record for all IT processes
>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>> _______________________________________________
>>>>> Xolotl-psi-development mailing list
>>>>> Xol...@li...
>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>> 
>>>>> ------------------------------------------------------------------------------
>>>>> How ServiceNow helps IT people transform IT departments:
>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>> 3. A single system of record for all IT processes
>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>> _______________________________________________
>>>>> Xolotl-psi-development mailing list
>>>>> Xol...@li...
>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>> 
>>> -- 
>>> Jay Jay Billings
>>> Oak Ridge National Laboratory
>>> Telephone: (865) 272-9420
>>> Email: bil...@or...
>>> Twitter Handle: @jayjaybillings
>>> 
>>> 
>>> ------------------------------------------------------------------------------
>>> How ServiceNow helps IT people transform IT departments:
>>> 1. A cloud service to automate IT design, transition and operations
>>> 2. Dashboards that offer high-level views of enterprise services
>>> 3. A single system of record for all IT processes
>>> http://p.sf.net/sfu/servicenow-d2d-j
>>> _______________________________________________
>>> Xolotl-psi-development mailing list
>>> Xol...@li...
>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
> 
> 
> -- 
> Jay Jay Billings
> Oak Ridge National Laboratory
> Telephone: (865) 272-9420
> Email: bil...@or...
> Twitter Handle: @jayjaybillings
> 
> 
> ------------------------------------------------------------------------------
> How ServiceNow helps IT people transform IT departments:
> 1. A cloud service to automate IT design, transition and operations
> 2. Dashboards that offer high-level views of enterprise services
> 3. A single system of record for all IT processes
> http://p.sf.net/sfu/servicenow-d2d-j
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development

Re: [Xolotl-psi-development] Testing const vs. constexpr

[Jay J. B. <bil...@or...>]

Barry,

That file is not in the repository. It is auto-generated when you run 
CMake in your build directory. I'm not sure how to fix that.

In your build directory, you should find an executable named "xolotl." 
You can run that with...

time ./xolotl ../benchmarks/tungsten.txt -ts_final_time 2000 -da_refine 
2 -log_summary -ts_adapt_dt_max 10 -ts_monitor

with the location to the benchmark directory specified as fits your 
system. It runs for about 58min on my desktop.

The data in that benchmark file is not yet passed to the PETSc code. I 
am using the values that you generate for now.

Jay

On 06/05/2013 03:49 PM, Barry Smith wrote:
>     Thanks. I ran and got
>
> Linking CXX executable xolotl
> /usr/bin/cmake -E cmake_link_script CMakeFiles/xolotl.dir/link.txt --verbose=1
> /usr/bin/c++   -std=c++0x     CMakeFiles/xolotl.dir/main.cpp.o  -o xolotl -rdynamic /sandbox/bsmith/petsc/arch-xolotl/lib/libpetsc.so xolotlCore/reactants/libxolotlReactants.so xolotlSolver/libxolotlSolver.so -Wl,-rpath,/sandbox/bsmith/petsc/arch-xolotl/lib:/sandbox/bsmith/trunk/xolotl/build/xolotlCore/reactants:/sandbox/bsmith/trunk/xolotl/build/xolotlSolver
> xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
> …
>
> so I edited cmake_link_script CMakeFiles/xolotl.dir/link.txt  to list /sandbox/bsmith/petsc/arch-xolotl/lib/libpetsc.so AFTER the other two .so  then the link went through and all the compiles/links worked (haven't actually run anything).
>
>     So it looks like the order of those libraries is important and the PETSc one (since the others depended on it) must be listed last.
>
>     Will you fix the repository so I can test again without my editing hack?
>
>     Also how do I run something?
>
>     Barry
>
>
>
> On Jun 5, 2013, at 1:31 PM, Jay Jay Billings <bil...@or...> wrote:
>
>> Barry,
>>
>> Do
>>
>> make VERBOSE=1
>>
>> Jay
>>
>> On 06/05/2013 02:10 PM, Barry Smith wrote:
>>> On Jun 5, 2013, at 11:35 AM, "Billings, Jay Jay" <bil...@or...> wrote:
>>>
>>>> Tim and Barry,
>>>>
>>>> I was able to make a fix along the lines that Tim suggested. You should be able to build it now.
>>>     That fixed that problem but I have others.
>>>
>>>     How can I get cmake to actually print what it is doing when I use make?
>>>
>>>     I would like to see the commands it is trying to use so I can debug it.  I have set LD_LIBRARY_PATH  but somehow it is not finding or linking in the PETSc library. It says "Linking CXX executable xolotl" but it would useful if it showed the exact link line.
>>>
>>>     Barry
>>>
>>>
>>> # fast build rule for target.
>>> xolotl/fast:
>>> 	$(MAKE) -f CMakeFiles/xolotl.dir/build.make CMakeFiles/xolotl.dir/build
>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ emacs -nw Makefile
>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ make
>>> [  6%] Built target xolotlSolver
>>> [ 73%] Built target xolotlReactants
>>> Linking CXX executable xolotl
>>> xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
>>> …..
>>>
>>>
>>>
>>>
>>>> Please let me know if you encounter more errors.
>>>>
>>>> I'll try to fix the compiler identification bug next.
>>>>
>>>> Jay
>>>> ________________________________________
>>>> From: Billings, Jay Jay [bil...@or...]
>>>> Sent: Wednesday, June 05, 2013 12:28 PM
>>>> To: xol...@li...
>>>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>>>
>>>> Tim,
>>>>
>>>> That could work. I'm initializing it in a class now, but let me try this out.
>>>>
>>>> Jay
>>>> ________________________________________
>>>> From: Tim Tautges [ta...@mc...]
>>>> Sent: Wednesday, June 05, 2013 12:26 PM
>>>> To: xol...@li...
>>>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>>>
>>>> Why not initialize at the top of the .cpp file, e.g.
>>>>
>>>> const double xolotlCore::PSICluster::kBoltzmann = 1.0e4;
>>>>
>>>> or something like that?  I haven't looked at the actual code, so maybe you're already doing that...
>>>>
>>>> - tim
>>>>
>>>> On 06/05/2013 11:10 AM, Jay Jay Billings wrote:
>>>>> Barry,
>>>>>
>>>>> Can you try a fix for me please and let me know if it works? Can you change the word "const" in Constants.h to "constexpr" and let me know if that fixes the bug you reported below?
>>>>>
>>>>>
>>>>> In file included from /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error: ‘constexpr’ needed for in-class initialization of static data member ‘kBoltzmann’ of non-integral type
>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp: In member function ‘virtual double xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
>>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15: error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
>>>>> make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o] Error 1
>>>>> make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
>>>>> make: *** [all] Error 2
>>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version
>>>>>
>>>>>
>>>>> This is a C++ vs. C++11 issue. We may need to declare a minimum compiler version for Xolotl or define the constants with
>>>>> macros.
>>>>>
>>>>> Jay
>>>>>
>>>>>
>>>>> --
>>>>> Jay Jay Billings
>>>>> Oak Ridge National Laboratory
>>>>> Twitter Handle: @jayjaybillings
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> How ServiceNow helps IT people transform IT departments:
>>>>> 1. A cloud service to automate IT design, transition and operations
>>>>> 2. Dashboards that offer high-level views of enterprise services
>>>>> 3. A single system of record for all IT processes
>>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Xolotl-psi-development mailing list
>>>>> Xol...@li...
>>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>>
>>>> --
>>>> ================================================================
>>>> "You will keep in perfect peace him whose mind is
>>>>    steadfast, because he trusts in you."               Isaiah 26:3
>>>>
>>>>               Tim Tautges            Argonne National Laboratory
>>>>           (ta...@mc...)      (telecommuting from UW-Madison)
>>>>   phone (gvoice): (608) 354-1459      1500 Engineering Dr.
>>>>              fax: (608) 263-4499      Madison, WI 53706
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> How ServiceNow helps IT people transform IT departments:
>>>> 1. A cloud service to automate IT design, transition and operations
>>>> 2. Dashboards that offer high-level views of enterprise services
>>>> 3. A single system of record for all IT processes
>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>> _______________________________________________
>>>> Xolotl-psi-development mailing list
>>>> Xol...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>
>>>> ------------------------------------------------------------------------------
>>>> How ServiceNow helps IT people transform IT departments:
>>>> 1. A cloud service to automate IT design, transition and operations
>>>> 2. Dashboards that offer high-level views of enterprise services
>>>> 3. A single system of record for all IT processes
>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>> _______________________________________________
>>>> Xolotl-psi-development mailing list
>>>> Xol...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>>
>>>> ------------------------------------------------------------------------------
>>>> How ServiceNow helps IT people transform IT departments:
>>>> 1. A cloud service to automate IT design, transition and operations
>>>> 2. Dashboards that offer high-level views of enterprise services
>>>> 3. A single system of record for all IT processes
>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>> _______________________________________________
>>>> Xolotl-psi-development mailing list
>>>> Xol...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>
>> -- 
>> Jay Jay Billings
>> Oak Ridge National Laboratory
>> Telephone: (865) 272-9420
>> Email: bil...@or...
>> Twitter Handle: @jayjaybillings
>>
>>
>> ------------------------------------------------------------------------------
>> How ServiceNow helps IT people transform IT departments:
>> 1. A cloud service to automate IT design, transition and operations
>> 2. Dashboards that offer high-level views of enterprise services
>> 3. A single system of record for all IT processes
>> http://p.sf.net/sfu/servicenow-d2d-j
>> _______________________________________________
>> Xolotl-psi-development mailing list
>> Xol...@li...
>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development


-- 
Jay Jay Billings
Oak Ridge National Laboratory
Telephone: (865) 272-9420
Email: bil...@or...
Twitter Handle: @jayjaybillings

Re: [Xolotl-psi-development] Testing const vs. constexpr

[Barry S. <bs...@mc...>]

   Thanks. I ran and got

Linking CXX executable xolotl
/usr/bin/cmake -E cmake_link_script CMakeFiles/xolotl.dir/link.txt --verbose=1
/usr/bin/c++   -std=c++0x     CMakeFiles/xolotl.dir/main.cpp.o  -o xolotl -rdynamic /sandbox/bsmith/petsc/arch-xolotl/lib/libpetsc.so xolotlCore/reactants/libxolotlReactants.so xolotlSolver/libxolotlSolver.so -Wl,-rpath,/sandbox/bsmith/petsc/arch-xolotl/lib:/sandbox/bsmith/trunk/xolotl/build/xolotlCore/reactants:/sandbox/bsmith/trunk/xolotl/build/xolotlSolver 
xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
…

so I edited cmake_link_script CMakeFiles/xolotl.dir/link.txt  to list /sandbox/bsmith/petsc/arch-xolotl/lib/libpetsc.so AFTER the other two .so  then the link went through and all the compiles/links worked (haven't actually run anything). 

   So it looks like the order of those libraries is important and the PETSc one (since the others depended on it) must be listed last.

   Will you fix the repository so I can test again without my editing hack?

   Also how do I run something?

   Barry



On Jun 5, 2013, at 1:31 PM, Jay Jay Billings <bil...@or...> wrote:

> Barry,
> 
> Do
> 
> make VERBOSE=1
> 
> Jay
> 
> On 06/05/2013 02:10 PM, Barry Smith wrote:
>> On Jun 5, 2013, at 11:35 AM, "Billings, Jay Jay" <bil...@or...> wrote:
>> 
>>> Tim and Barry,
>>> 
>>> I was able to make a fix along the lines that Tim suggested. You should be able to build it now.
>>    That fixed that problem but I have others.
>> 
>>    How can I get cmake to actually print what it is doing when I use make?
>> 
>>    I would like to see the commands it is trying to use so I can debug it.  I have set LD_LIBRARY_PATH  but somehow it is not finding or linking in the PETSc library. It says "Linking CXX executable xolotl" but it would useful if it showed the exact link line.
>> 
>>    Barry
>> 
>> 
>> # fast build rule for target.
>> xolotl/fast:
>> 	$(MAKE) -f CMakeFiles/xolotl.dir/build.make CMakeFiles/xolotl.dir/build
>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ emacs -nw Makefile
>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ make
>> [  6%] Built target xolotlSolver
>> [ 73%] Built target xolotlReactants
>> Linking CXX executable xolotl
>> xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
>> …..
>> 
>> 
>> 
>> 
>>> Please let me know if you encounter more errors.
>>> 
>>> I'll try to fix the compiler identification bug next.
>>> 
>>> Jay
>>> ________________________________________
>>> From: Billings, Jay Jay [bil...@or...]
>>> Sent: Wednesday, June 05, 2013 12:28 PM
>>> To: xol...@li...
>>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>> 
>>> Tim,
>>> 
>>> That could work. I'm initializing it in a class now, but let me try this out.
>>> 
>>> Jay
>>> ________________________________________
>>> From: Tim Tautges [ta...@mc...]
>>> Sent: Wednesday, June 05, 2013 12:26 PM
>>> To: xol...@li...
>>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>> 
>>> Why not initialize at the top of the .cpp file, e.g.
>>> 
>>> const double xolotlCore::PSICluster::kBoltzmann = 1.0e4;
>>> 
>>> or something like that?  I haven't looked at the actual code, so maybe you're already doing that...
>>> 
>>> - tim
>>> 
>>> On 06/05/2013 11:10 AM, Jay Jay Billings wrote:
>>>> Barry,
>>>> 
>>>> Can you try a fix for me please and let me know if it works? Can you change the word "const" in Constants.h to "constexpr" and let me know if that fixes the bug you reported below?
>>>> 
>>>> 
>>>> In file included from /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error: ‘constexpr’ needed for in-class initialization of static data member ‘kBoltzmann’ of non-integral type
>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp: In member function ‘virtual double xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
>>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15: error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
>>>> make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o] Error 1
>>>> make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
>>>> make: *** [all] Error 2
>>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version
>>>> 
>>>> 
>>>> This is a C++ vs. C++11 issue. We may need to declare a minimum compiler version for Xolotl or define the constants with
>>>> macros.
>>>> 
>>>> Jay
>>>> 
>>>> 
>>>> --
>>>> Jay Jay Billings
>>>> Oak Ridge National Laboratory
>>>> Twitter Handle: @jayjaybillings
>>>> 
>>>> 
>>>> ------------------------------------------------------------------------------
>>>> How ServiceNow helps IT people transform IT departments:
>>>> 1. A cloud service to automate IT design, transition and operations
>>>> 2. Dashboards that offer high-level views of enterprise services
>>>> 3. A single system of record for all IT processes
>>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>> 
>>>> 
>>>> 
>>>> _______________________________________________
>>>> Xolotl-psi-development mailing list
>>>> Xol...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>> 
>>> --
>>> ================================================================
>>> "You will keep in perfect peace him whose mind is
>>>   steadfast, because he trusts in you."               Isaiah 26:3
>>> 
>>>              Tim Tautges            Argonne National Laboratory
>>>          (ta...@mc...)      (telecommuting from UW-Madison)
>>>  phone (gvoice): (608) 354-1459      1500 Engineering Dr.
>>>             fax: (608) 263-4499      Madison, WI 53706
>>> 
>>> 
>>> ------------------------------------------------------------------------------
>>> How ServiceNow helps IT people transform IT departments:
>>> 1. A cloud service to automate IT design, transition and operations
>>> 2. Dashboards that offer high-level views of enterprise services
>>> 3. A single system of record for all IT processes
>>> http://p.sf.net/sfu/servicenow-d2d-j
>>> _______________________________________________
>>> Xolotl-psi-development mailing list
>>> Xol...@li...
>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>> 
>>> ------------------------------------------------------------------------------
>>> How ServiceNow helps IT people transform IT departments:
>>> 1. A cloud service to automate IT design, transition and operations
>>> 2. Dashboards that offer high-level views of enterprise services
>>> 3. A single system of record for all IT processes
>>> http://p.sf.net/sfu/servicenow-d2d-j
>>> _______________________________________________
>>> Xolotl-psi-development mailing list
>>> Xol...@li...
>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>> 
>>> ------------------------------------------------------------------------------
>>> How ServiceNow helps IT people transform IT departments:
>>> 1. A cloud service to automate IT design, transition and operations
>>> 2. Dashboards that offer high-level views of enterprise services
>>> 3. A single system of record for all IT processes
>>> http://p.sf.net/sfu/servicenow-d2d-j
>>> _______________________________________________
>>> Xolotl-psi-development mailing list
>>> Xol...@li...
>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
> 
> 
> -- 
> Jay Jay Billings
> Oak Ridge National Laboratory
> Telephone: (865) 272-9420
> Email: bil...@or...
> Twitter Handle: @jayjaybillings
> 
> 
> ------------------------------------------------------------------------------
> How ServiceNow helps IT people transform IT departments:
> 1. A cloud service to automate IT design, transition and operations
> 2. Dashboards that offer high-level views of enterprise services
> 3. A single system of record for all IT processes
> http://p.sf.net/sfu/servicenow-d2d-j
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development

Re: [Xolotl-psi-development] Testing const vs. constexpr

[Jay J. B. <bil...@or...>]

Barry,

Do

make VERBOSE=1

Jay

On 06/05/2013 02:10 PM, Barry Smith wrote:
> On Jun 5, 2013, at 11:35 AM, "Billings, Jay Jay" <bil...@or...> wrote:
>
>> Tim and Barry,
>>
>> I was able to make a fix along the lines that Tim suggested. You should be able to build it now.
>     That fixed that problem but I have others.
>
>     How can I get cmake to actually print what it is doing when I use make?
>
>     I would like to see the commands it is trying to use so I can debug it.  I have set LD_LIBRARY_PATH  but somehow it is not finding or linking in the PETSc library. It says "Linking CXX executable xolotl" but it would useful if it showed the exact link line.
>
>     Barry
>
>
> # fast build rule for target.
> xolotl/fast:
> 	$(MAKE) -f CMakeFiles/xolotl.dir/build.make CMakeFiles/xolotl.dir/build
> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ emacs -nw Makefile
> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ make
> [  6%] Built target xolotlSolver
> [ 73%] Built target xolotlReactants
> Linking CXX executable xolotl
> xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
> …..
>
>
>
>
>> Please let me know if you encounter more errors.
>>
>> I'll try to fix the compiler identification bug next.
>>
>> Jay
>> ________________________________________
>> From: Billings, Jay Jay [bil...@or...]
>> Sent: Wednesday, June 05, 2013 12:28 PM
>> To: xol...@li...
>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>
>> Tim,
>>
>> That could work. I'm initializing it in a class now, but let me try this out.
>>
>> Jay
>> ________________________________________
>> From: Tim Tautges [ta...@mc...]
>> Sent: Wednesday, June 05, 2013 12:26 PM
>> To: xol...@li...
>> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
>>
>> Why not initialize at the top of the .cpp file, e.g.
>>
>> const double xolotlCore::PSICluster::kBoltzmann = 1.0e4;
>>
>> or something like that?  I haven't looked at the actual code, so maybe you're already doing that...
>>
>> - tim
>>
>> On 06/05/2013 11:10 AM, Jay Jay Billings wrote:
>>> Barry,
>>>
>>> Can you try a fix for me please and let me know if it works? Can you change the word "const" in Constants.h to "constexpr" and let me know if that fixes the bug you reported below?
>>>
>>>
>>> In file included from /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error: ‘constexpr’ needed for in-class initialization of static data member ‘kBoltzmann’ of non-integral type
>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp: In member function ‘virtual double xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
>>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15: error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
>>> make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o] Error 1
>>> make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
>>> make: *** [all] Error 2
>>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version
>>>
>>>
>>> This is a C++ vs. C++11 issue. We may need to declare a minimum compiler version for Xolotl or define the constants with
>>> macros.
>>>
>>> Jay
>>>
>>>
>>> --
>>> Jay Jay Billings
>>> Oak Ridge National Laboratory
>>> Twitter Handle: @jayjaybillings
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> How ServiceNow helps IT people transform IT departments:
>>> 1. A cloud service to automate IT design, transition and operations
>>> 2. Dashboards that offer high-level views of enterprise services
>>> 3. A single system of record for all IT processes
>>> http://p.sf.net/sfu/servicenow-d2d-j
>>>
>>>
>>>
>>> _______________________________________________
>>> Xolotl-psi-development mailing list
>>> Xol...@li...
>>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>>
>> --
>> ================================================================
>> "You will keep in perfect peace him whose mind is
>>    steadfast, because he trusts in you."               Isaiah 26:3
>>
>>               Tim Tautges            Argonne National Laboratory
>>           (ta...@mc...)      (telecommuting from UW-Madison)
>>   phone (gvoice): (608) 354-1459      1500 Engineering Dr.
>>              fax: (608) 263-4499      Madison, WI 53706
>>
>>
>> ------------------------------------------------------------------------------
>> How ServiceNow helps IT people transform IT departments:
>> 1. A cloud service to automate IT design, transition and operations
>> 2. Dashboards that offer high-level views of enterprise services
>> 3. A single system of record for all IT processes
>> http://p.sf.net/sfu/servicenow-d2d-j
>> _______________________________________________
>> Xolotl-psi-development mailing list
>> Xol...@li...
>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>
>> ------------------------------------------------------------------------------
>> How ServiceNow helps IT people transform IT departments:
>> 1. A cloud service to automate IT design, transition and operations
>> 2. Dashboards that offer high-level views of enterprise services
>> 3. A single system of record for all IT processes
>> http://p.sf.net/sfu/servicenow-d2d-j
>> _______________________________________________
>> Xolotl-psi-development mailing list
>> Xol...@li...
>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>>
>> ------------------------------------------------------------------------------
>> How ServiceNow helps IT people transform IT departments:
>> 1. A cloud service to automate IT design, transition and operations
>> 2. Dashboards that offer high-level views of enterprise services
>> 3. A single system of record for all IT processes
>> http://p.sf.net/sfu/servicenow-d2d-j
>> _______________________________________________
>> Xolotl-psi-development mailing list
>> Xol...@li...
>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development


-- 
Jay Jay Billings
Oak Ridge National Laboratory
Telephone: (865) 272-9420
Email: bil...@or...
Twitter Handle: @jayjaybillings

Re: [Xolotl-psi-development] Testing const vs. constexpr

[Barry S. <bs...@mc...>]

On Jun 5, 2013, at 11:35 AM, "Billings, Jay Jay" <bil...@or...> wrote:

> Tim and Barry, 
> 
> I was able to make a fix along the lines that Tim suggested. You should be able to build it now.

   That fixed that problem but I have others. 

   How can I get cmake to actually print what it is doing when I use make?

   I would like to see the commands it is trying to use so I can debug it.  I have set LD_LIBRARY_PATH  but somehow it is not finding or linking in the PETSc library. It says "Linking CXX executable xolotl" but it would useful if it showed the exact link line.

   Barry


# fast build rule for target.
xolotl/fast:
	$(MAKE) -f CMakeFiles/xolotl.dir/build.make CMakeFiles/xolotl.dir/build
bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ emacs -nw Makefile 
bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ make
[  6%] Built target xolotlSolver
[ 73%] Built target xolotlReactants
Linking CXX executable xolotl
xolotlSolver/libxolotlSolver.so: undefined reference to `DMDAGetOwnershipRanges'
…..




> Please let me know if you encounter more errors.
> 
> I'll try to fix the compiler identification bug next.
> 
> Jay
> ________________________________________
> From: Billings, Jay Jay [bil...@or...]
> Sent: Wednesday, June 05, 2013 12:28 PM
> To: xol...@li...
> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
> 
> Tim,
> 
> That could work. I'm initializing it in a class now, but let me try this out.
> 
> Jay
> ________________________________________
> From: Tim Tautges [ta...@mc...]
> Sent: Wednesday, June 05, 2013 12:26 PM
> To: xol...@li...
> Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr
> 
> Why not initialize at the top of the .cpp file, e.g.
> 
> const double xolotlCore::PSICluster::kBoltzmann = 1.0e4;
> 
> or something like that?  I haven't looked at the actual code, so maybe you're already doing that...
> 
> - tim
> 
> On 06/05/2013 11:10 AM, Jay Jay Billings wrote:
>> Barry,
>> 
>> Can you try a fix for me please and let me know if it works? Can you change the word "const" in Constants.h to "constexpr" and let me know if that fixes the bug you reported below?
>> 
>> 
>> In file included from /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
>> /sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error: ‘constexpr’ needed for in-class initialization of static data member ‘kBoltzmann’ of non-integral type
>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp: In member function ‘virtual double xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
>> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15: error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
>> make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o] Error 1
>> make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
>> make: *** [all] Error 2
>> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version
>> 
>> 
>> This is a C++ vs. C++11 issue. We may need to declare a minimum compiler version for Xolotl or define the constants with
>> macros.
>> 
>> Jay
>> 
>> 
>> --
>> Jay Jay Billings
>> Oak Ridge National Laboratory
>> Twitter Handle: @jayjaybillings
>> 
>> 
>> ------------------------------------------------------------------------------
>> How ServiceNow helps IT people transform IT departments:
>> 1. A cloud service to automate IT design, transition and operations
>> 2. Dashboards that offer high-level views of enterprise services
>> 3. A single system of record for all IT processes
>> http://p.sf.net/sfu/servicenow-d2d-j
>> 
>> 
>> 
>> _______________________________________________
>> Xolotl-psi-development mailing list
>> Xol...@li...
>> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>> 
> 
> --
> ================================================================
> "You will keep in perfect peace him whose mind is
>   steadfast, because he trusts in you."               Isaiah 26:3
> 
>              Tim Tautges            Argonne National Laboratory
>          (ta...@mc...)      (telecommuting from UW-Madison)
>  phone (gvoice): (608) 354-1459      1500 Engineering Dr.
>             fax: (608) 263-4499      Madison, WI 53706
> 
> 
> ------------------------------------------------------------------------------
> How ServiceNow helps IT people transform IT departments:
> 1. A cloud service to automate IT design, transition and operations
> 2. Dashboards that offer high-level views of enterprise services
> 3. A single system of record for all IT processes
> http://p.sf.net/sfu/servicenow-d2d-j
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
> 
> ------------------------------------------------------------------------------
> How ServiceNow helps IT people transform IT departments:
> 1. A cloud service to automate IT design, transition and operations
> 2. Dashboards that offer high-level views of enterprise services
> 3. A single system of record for all IT processes
> http://p.sf.net/sfu/servicenow-d2d-j
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
> 
> ------------------------------------------------------------------------------
> How ServiceNow helps IT people transform IT departments:
> 1. A cloud service to automate IT design, transition and operations
> 2. Dashboards that offer high-level views of enterprise services
> 3. A single system of record for all IT processes
> http://p.sf.net/sfu/servicenow-d2d-j
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development

Re: [Xolotl-psi-development] Testing const vs. constexpr

[Billings, J. J. <bil...@or...>]

Tim and Barry, 

I was able to make a fix along the lines that Tim suggested. You should be able to build it now. Please let me know if you encounter more errors.

I'll try to fix the compiler identification bug next.

Jay
________________________________________
From: Billings, Jay Jay [bil...@or...]
Sent: Wednesday, June 05, 2013 12:28 PM
To: xol...@li...
Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr

Tim,

That could work. I'm initializing it in a class now, but let me try this out.

Jay
________________________________________
From: Tim Tautges [ta...@mc...]
Sent: Wednesday, June 05, 2013 12:26 PM
To: xol...@li...
Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr

Why not initialize at the top of the .cpp file, e.g.

const double xolotlCore::PSICluster::kBoltzmann = 1.0e4;

or something like that?  I haven't looked at the actual code, so maybe you're already doing that...

- tim

On 06/05/2013 11:10 AM, Jay Jay Billings wrote:
> Barry,
>
> Can you try a fix for me please and let me know if it works? Can you change the word "const" in Constants.h to "constexpr" and let me know if that fixes the bug you reported below?
>
>
> In file included from /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
> /sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error: ‘constexpr’ needed for in-class initialization of static data member ‘kBoltzmann’ of non-integral type
> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp: In member function ‘virtual double xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15: error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
> make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o] Error 1
> make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
> make: *** [all] Error 2
> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version
>
>
> This is a C++ vs. C++11 issue. We may need to declare a minimum compiler version for Xolotl or define the constants with
> macros.
>
> Jay
>
>
> --
> Jay Jay Billings
> Oak Ridge National Laboratory
> Twitter Handle: @jayjaybillings
>
>
> ------------------------------------------------------------------------------
> How ServiceNow helps IT people transform IT departments:
> 1. A cloud service to automate IT design, transition and operations
> 2. Dashboards that offer high-level views of enterprise services
> 3. A single system of record for all IT processes
> http://p.sf.net/sfu/servicenow-d2d-j
>
>
>
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>

--
================================================================
"You will keep in perfect peace him whose mind is
   steadfast, because he trusts in you."               Isaiah 26:3

              Tim Tautges            Argonne National Laboratory
          (ta...@mc...)      (telecommuting from UW-Madison)
  phone (gvoice): (608) 354-1459      1500 Engineering Dr.
             fax: (608) 263-4499      Madison, WI 53706


------------------------------------------------------------------------------
How ServiceNow helps IT people transform IT departments:
1. A cloud service to automate IT design, transition and operations
2. Dashboards that offer high-level views of enterprise services
3. A single system of record for all IT processes
http://p.sf.net/sfu/servicenow-d2d-j
_______________________________________________
Xolotl-psi-development mailing list
Xol...@li...
https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development

------------------------------------------------------------------------------
How ServiceNow helps IT people transform IT departments:
1. A cloud service to automate IT design, transition and operations
2. Dashboards that offer high-level views of enterprise services
3. A single system of record for all IT processes
http://p.sf.net/sfu/servicenow-d2d-j
_______________________________________________
Xolotl-psi-development mailing list
Xol...@li...
https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development

Re: [Xolotl-psi-development] Testing const vs. constexpr

[Billings, J. J. <bil...@or...>]

Tim,

That could work. I'm initializing it in a class now, but let me try this out. 

Jay
________________________________________
From: Tim Tautges [ta...@mc...]
Sent: Wednesday, June 05, 2013 12:26 PM
To: xol...@li...
Subject: Re: [Xolotl-psi-development] Testing const vs. constexpr

Why not initialize at the top of the .cpp file, e.g.

const double xolotlCore::PSICluster::kBoltzmann = 1.0e4;

or something like that?  I haven't looked at the actual code, so maybe you're already doing that...

- tim

On 06/05/2013 11:10 AM, Jay Jay Billings wrote:
> Barry,
>
> Can you try a fix for me please and let me know if it works? Can you change the word "const" in Constants.h to "constexpr" and let me know if that fixes the bug you reported below?
>
>
> In file included from /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
> /sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error: ‘constexpr’ needed for in-class initialization of static data member ‘kBoltzmann’ of non-integral type
> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp: In member function ‘virtual double xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15: error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
> make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o] Error 1
> make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
> make: *** [all] Error 2
> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version
>
>
> This is a C++ vs. C++11 issue. We may need to declare a minimum compiler version for Xolotl or define the constants with
> macros.
>
> Jay
>
>
> --
> Jay Jay Billings
> Oak Ridge National Laboratory
> Twitter Handle: @jayjaybillings
>
>
> ------------------------------------------------------------------------------
> How ServiceNow helps IT people transform IT departments:
> 1. A cloud service to automate IT design, transition and operations
> 2. Dashboards that offer high-level views of enterprise services
> 3. A single system of record for all IT processes
> http://p.sf.net/sfu/servicenow-d2d-j
>
>
>
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>

--
================================================================
"You will keep in perfect peace him whose mind is
   steadfast, because he trusts in you."               Isaiah 26:3

              Tim Tautges            Argonne National Laboratory
          (ta...@mc...)      (telecommuting from UW-Madison)
  phone (gvoice): (608) 354-1459      1500 Engineering Dr.
             fax: (608) 263-4499      Madison, WI 53706


------------------------------------------------------------------------------
How ServiceNow helps IT people transform IT departments:
1. A cloud service to automate IT design, transition and operations
2. Dashboards that offer high-level views of enterprise services
3. A single system of record for all IT processes
http://p.sf.net/sfu/servicenow-d2d-j
_______________________________________________
Xolotl-psi-development mailing list
Xol...@li...
https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development

Re: [Xolotl-psi-development] Testing const vs. constexpr

[Tim T. <ta...@mc...>]

Why not initialize at the top of the .cpp file, e.g.

const double xolotlCore::PSICluster::kBoltzmann = 1.0e4;

or something like that?  I haven't looked at the actual code, so maybe you're already doing that...

- tim

On 06/05/2013 11:10 AM, Jay Jay Billings wrote:
> Barry,
>
> Can you try a fix for me please and let me know if it works? Can you change the word "const" in Constants.h to "constexpr" and let me know if that fixes the bug you reported below?
>
>
> In file included from /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
> /sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error: ‘constexpr’ needed for in-class initialization of static data member ‘kBoltzmann’ of non-integral type
> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp: In member function ‘virtual double xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
> /sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15: error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
> make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o] Error 1
> make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
> make: *** [all] Error 2
> bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version
>
>
> This is a C++ vs. C++11 issue. We may need to declare a minimum compiler version for Xolotl or define the constants with
> macros.
>
> Jay
>
>
> --
> Jay Jay Billings
> Oak Ridge National Laboratory
> Twitter Handle: @jayjaybillings
>
>
> ------------------------------------------------------------------------------
> How ServiceNow helps IT people transform IT departments:
> 1. A cloud service to automate IT design, transition and operations
> 2. Dashboards that offer high-level views of enterprise services
> 3. A single system of record for all IT processes
> http://p.sf.net/sfu/servicenow-d2d-j
>
>
>
> _______________________________________________
> Xolotl-psi-development mailing list
> Xol...@li...
> https://lists.sourceforge.net/lists/listinfo/xolotl-psi-development
>

-- 
================================================================
"You will keep in perfect peace him whose mind is
   steadfast, because he trusts in you."               Isaiah 26:3

              Tim Tautges            Argonne National Laboratory
          (ta...@mc...)      (telecommuting from UW-Madison)
  phone (gvoice): (608) 354-1459      1500 Engineering Dr.
             fax: (608) 263-4499      Madison, WI 53706

[Xolotl-psi-development] Testing const vs. constexpr

[Jay J. B. <jay...@gm...>]

Barry,

Can you try a fix for me please and let me know if it works? Can you
change the word "const" in Constants.h to "constexpr" and let me know
if that fixes the bug you reported below?


In file included from
/sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
/sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error:
‘constexpr’ needed for in-class initialization of static data member
‘kBoltzmann’ of non-integral type
/sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:
In member function ‘virtual double
xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
/sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15:
error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o]
Error 1
make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
make: *** [all] Error 2
bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version


This is a C++ vs. C++11 issue. We may need to declare a minimum compiler
version for Xolotl or define the constants with macros.

Jay


-- 
Jay Jay Billings
Oak Ridge National Laboratory
Twitter Handle: @jayjaybillings

[Xolotl-psi-development] Barry's build problems on Linux

[Jay J. B. <jay...@gm...>]

Everyone,

The list is setup.

Jay

-----


#5 might take a bit. Sourceforge.net is taking its time creating the list. Grr.

Jay
________________________________________
From: Billings, Jay Jay
Sent: Monday, June 03, 2013 3:47 PM
To: Barry Smith
Cc: Jungho Lee; Tim Tautges
Subject: RE: Speaking at 2 pm

Barry,

1.) Yes, of course.
2.) I have no idea how it found that.
3.) I am using FindETSc (from Jed Brown,
https://github.com/jedbrown/cmake-modules/blob/master/FindPETSc.cmake).
It is in the modules directory of the source. I don't know why it
fails that test, but it still builds.
4.) I'm not sure why that is failing. I am trying to build on another
machine right now to check.
5.) Look for the Welcome email in a few minutes.

Jay
________________________________________
From: Barry Smith [bs...@mc...]
Sent: Monday, June 03, 2013 3:41 PM
To: Billings, Jay Jay
Cc: Jungho Lee; Tim Tautges
Subject: Re: Speaking at 2 pm

  On linux

   1) I would argue the cmake is all wrong to be finding a raw gcc and
g++ it should be finding the mpicc and mlicxx compilers directly and
using them.

    2) Found PETSc:
*/sandbox/bsmith/petsc/include;/sandbox/bsmith/petsc/arch-xolotl/include;/usr/include/mpich2
  it should NOT have found that mpich2 include directory. Just because
there happens to be an MPI include somewhere on the system doesn't
mean it is the one I used.  My MPI is in
/sandbox/bsmith/petsc/arch-xolotl/*

    3) Don't know why the PETSc test failed and where the detailed
information on it is available? It should print the log file location
so I can just look at it and see why it failed.

    4) Seems to be some bug in the C++ code.

     5) Is there, will there be a xolotl-dev mailing list or "issues"
website where we can shared issues with compiling etc rather than
sending emails directly to random peoples email addresses?

   Barry

bsmith@petsc:*/sandbox/bsmith/trunk/xolotl$ cd build/*
bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ cmake ..
-- The C compiler identification is GNU
-- The CXX compiler identification is GNU
-- Check for working C compiler: /usr/bin/gcc
-- Check for working C compiler: /usr/bin/gcc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Boost version: 1.46.1
-- petsc_lib_dir /sandbox/bsmith/petsc/arch-xolotl/lib
-- Recognized PETSc install with single library for all packages
-- Performing Test MULTIPASS_TEST_1_petsc_works_minimal
-- Performing Test MULTIPASS_TEST_1_petsc_works_minimal - Failed
-- Performing Test MULTIPASS_TEST_2_petsc_works_allincludes
-- Performing Test MULTIPASS_TEST_2_petsc_works_allincludes - Success
-- PETSc requires extra include paths, but links correctly with only
interface libraries.  This is an unexpected configuration (but it
seems to work fine).
-- Found PETSc:
/sandbox/bsmith/petsc/include;/sandbox/bsmith/petsc/arch-xolotl/include;/usr/include/mpich2
-- Making test /sandbox/bsmith/trunk/xolotl/tests/io/TokenizedLineReaderTester.cpp
-- Making test /sandbox/bsmith/trunk/xolotl/tests/reactants/PSIClusterNetworkLoaderTester.cpp
-- Making test /sandbox/bsmith/trunk/xolotl/tests/reactants/PSIClusterTester.cpp
-- Configuring done
-- Generating done
-- Build files have been written to: /sandbox/bsmith/trunk/xolotl/build
bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ make
Scanning dependencies of target xolotlSolver
[  6%] Building CXX object
xolotlSolver/CMakeFiles/xolotlSolver.dir/PetscSolver.cpp.o
Linking CXX shared library libxolotlSolver.so
[  6%] Built target xolotlSolver
Scanning dependencies of target xolotlReactants
[ 13%] Building CXX object
xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/Reactant.cpp.o
[ 20%] Building CXX object
xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/MixedSpeciesCluster.cpp.o
[ 26%] Building CXX object
xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSIClusterNetworkLoader.cpp.o
[ 33%] Building CXX object
xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/TCluster.cpp.o
[ 40%] Building CXX object
xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/InterstitialCluster.cpp.o
[ 46%] Building CXX object
xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/HCluster.cpp.o
[ 53%] Building CXX object
xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/DCluster.cpp.o
[ 60%] Building CXX object
xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/HeCluster.cpp.o
[ 66%] Building CXX object
xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o
In file included from
/sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:2:0:
/sandbox/bsmith/trunk/xolotl/xolotlCore/Constants.h:18:35: error:
‘constexpr’ needed for in-class initialization of static data member
‘kBoltzmann’ of non-integral type
/sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:
In member function ‘virtual double
xolotlCore::PSICluster::getDiffusionCoefficient(double)’:
/sandbox/bsmith/trunk/xolotl/xolotlCore/reactants/psiclusters/PSICluster.cpp:61:15:
error: ‘kBoltzmann’ is not a member of ‘xolotlCore::Constants’
make[2]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/psiclusters/PSICluster.cpp.o]
Error 1
make[1]: *** [xolotlCore/reactants/CMakeFiles/xolotlReactants.dir/all] Error 2
make: *** [all] Error 2
bsmith@petsc:/sandbox/bsmith/trunk/xolotl/build$ g++ --version


g++ (Ubuntu/Linaro 4.6.3-1ubuntu5) 4.6.3
Copyright (C) 2011 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Linux petsc 3.2.0-39-generic #62-Ubuntu SMP Thu Feb 28 00:28Linux
petsc 3.2.0-39-generic #62-Ubuntu SMP Thu Feb 28 00:28:53 UTC 2013
x86_64 x86_64 x86_64 GNU/Linux



-- 
Jay Jay Billings
Oak Ridge National Laboratory
Twitter Handle: @jayjaybillings