mesa-users — open list for mesa comments, questions, & answers

Re: mesa-users] Using Mesa

[Bill P. <pa...@ki...>]

On Nov 10, 2008, at 10:24 PM, Roni Waldman wrote:

> What's wrong?

Let's go back to check how mesa was installed.
Did you do an svn checkout?  That's currently the best way to go.
The current release version is given in the "getting started" webpage  
as 743, so you'd do

svn co -r 743 https://mesa.svn.sourceforge.net/svnroot/mesa/trunk mesa

After the svn checkout finishes downloading lots of files, open

mesa/utils/makefile_header

and check that it's setup to use the fortran compiler you use.

Once you've specified the compiler, the build is done by:

cd mesa
./install

This script goes through building and testing all of the required  
packages, finishing with mesa/star.  You shouldn't have to do ./mk  
unless you've edited some source file.

Let me know how it goes!

Cheers,
Bill

Re: mesa-users] Using Mesa

[Bill P. <pa...@ki...>]

Hi,

BTW: just a side-note: EZ was fast at what it could do, but it was  
pretty much impossible for me to make it do anything beyond what  
Eggleton's original code could do.  With mesa I've tried to make the  
code much more flexible and robust, but you'll find places where this  
means much slower evolution as it cautiously tip-toes along.  The  
climb up the rgb for low mass stars is a particularly striking case.   
You may want to let some of those run overnight!

Cheers,
Bill

Re: mesa-users] Using Mesa

[Bill P. <pa...@ki...>]

On Nov 10, 2008, at 5:09 AM, Roni Waldman wrote:

> I am interested in trying out mesa,

> Where is the actual evolution code, and how do I run it?


Hi Roni,

I'm happy that you're interested in trying out mesa.   The top-level  
stellar evolution package is "mesa/star".

Like the other packages, mesa/star is organized with sources in  
public and private directories.  In mesa/star/public you'll find  
procedures (in star_lib.f), stellar data (in star_data.dek), and  
control parameters (in star_controls.dek).  Code that actually runs  
the mesa/star evolution package can be found in mesa/star/test/src.   
However, you shouldn't have to edit those files (at least not for  
now).  Instead, the parameters of a run are read from a text file  
called 'inlist' that lives in the star/test directory.  You can edit  
that file to pick things such as initial_mass and initial_Z as well  
as more exotic things like which nuclear net to use.

After you've set up inlist, you can launch the code by using one of  
several little script files that are found in star/test:

./rn
  -- run an evolution using the current contents of inlist.

Information about the state of the star is periodically written to  
the terminal (at a frequency determined by the 'log_cnt' control).

The code also periodically saves "snapshots" of the internal state  
for possible restarts (the 'photostep' control determines how often  
this happens).  For model numbers that are a multiple of 1000, the  
name of the snapshot is simply the model number.  Otherwise, the name  
is 'x' followed by the last 3 digits of the model number. For  
example, a snapshot of model 1325 will be saved as x325.  The  
snapshot files are saved in the 'photos' directory.  (This naming  
scheme was created after I filled up my disc with snapshots during an  
overnight run -- now with photostep set to 10 or 20, I can let it run  
overnight without worrying about creating too many snapshots.)

If you want to change some controls and restart from the snapshot in  
x325, do

./re x325
   -- restart using snapshot in file photos/x325

Logs and profiles are written to the 'LOGS' directory in test  
(profiles are saved at a frequency determined by the  
'profile_interval' control; a star log entry is created for each model).

The files 'log1.data', 'log2.data', etc. have information about a  
particular model.
The file 'profiles.index' has information about what's in those log  
files.
Finally, 'star.log' has the history for the entire run of a variety  
of values.

The LOGS information is automatically reset when you start a new  
evolution using ./rn, but it is not reset when you do a restart  
using ./re -- however it is okay to delete the entire contents of  
LOGS at any time during a run; it will automatically remake the files  
and keep going from there.

You'll find that the profiles and logs are organized as columns of  
data with enough text info included to make it reasonably clear  
what's what.

I use Tioga for plotting (it was created for this sort of thing), and  
the plot routines that I use for mesa/star can be found in  
star_history and star_profile.

At this time, the documentation for mesa/star is limited to emails  
from me!    So don't hesitate to ask.  Just understand that things  
are still in a very preliminary state; I'm sure there are lots of  
bugs out there waiting to be found.  ; - )

Cheers,
Bill

Re: mesa-users] dlnP/dlnMu ?

[Frank T. <fx...@ma...>]

hi bill,

sure.

fxt



> Hi,
>
> Since mesa/eos returns both mu and free_e = 1/mu_e, and their  
> partials wrt T and rho, is it possible to remove the full ionization  
> assumption?
>
> --Bill

Re: mesa-users] dlnP/dlnMu ?

[Bill P. <pa...@ki...>]

On May 4, 2008, at 11:20 AM, Ken Shen wrote:

> Great!  Just to clarify, for example, is h_dpd = dlnP/dlnrho, and is
> h_dpdt = dln(dlnP/dlnrho)/dlnT?  And are these log derivatives (as  
> opposed
> to h_dpd = dP/drho)?

The helm derivatives are not logs, so h_dpd = dP/drho = dP/dd ('d'  
for 'density')

-Bill

Re: mesa-users] dlnP/dlnMu ?

[Bill P. <pa...@ki...>]

The routine eos_get_everything returns all of the HELM results in  
addition to the usual mesa/eos results.
In eos_def you'll find a long list of integer parameter definitions  
for the results from helm.  The ones starting 'h_dp' are the  
derivatives of pressure, and 'h_dpda' is the partial wrt abar,  
'h_dpdz' is the partical wrt zbar, etc.

--Bill



On May 4, 2008, at 11:06 AM, Ken Shen wrote:

> That would definitely work, except I don't know how to access dP/d 
> (abar)
> etc.  Which function would I use?  There is a dmu/dlnT in the more
> complicated eos_get, but I'm not quite sure what that is...
>
>
> Thanks,
> Ken
>
>
>
> On Sun, 4 May 2008, Frank Timmes wrote:
>
>> hi ken,
>>
>> ok. assuming full ionization,
>>
>> mu_i = 1 /(sum Y_i) = Abar
>>
>> mu_e = 1 /(sum Y_i Z_i) = 1/Y_e
>>
>> mu = ( 1/mu_i + 1/mu_e )**(-1) = 1/(sum Y_i (Z_i + 1)) = Abar/ 
>> (Zbar + 1)
>>
>> d(mu) = dAbar/(Zbar + 1) + Abar/(Zbar + 1)**2 d(Zbar)
>>
>> the routines return derivatives with respect
>> to composition like dP/d(Abar) and dP/d(Zbar).
>>
>> given the relation between mu, Abar and Zbar
>> can you fill in the rest?
>>
>> fxt
>>
>>
>>
>>> I mean the mean mass per particle, like, 1/mu = 1/mu_i + 1/mu_e.
>>
>>
>>
>

Re: mesa-users] dlnP/dlnMu ?

[Bill P. <pa...@ki...>]

Hi,

Since mesa/eos returns both mu and free_e = 1/mu_e, and their  
partials wrt T and rho, is it possible to remove the full ionization  
assumption?

--Bill


BTW: I've copied this to mesa-users just to get in the habit of  
leaving a trace of such conservations that will undoubtedly come up  
again in the future.  Please humor me on this and include mesa-users  
in your messages too.


On May 4, 2008, at 11:02 AM, Frank Timmes wrote:

> hi ken,
>
> ok. assuming full ionization,
>
> mu_i = 1 /(sum Y_i) = Abar
>
> mu_e = 1 /(sum Y_i Z_i) = 1/Y_e
>
> mu = ( 1/mu_i + 1/mu_e )**(-1) = 1/(sum Y_i (Z_i + 1)) = Abar/(Zbar  
> + 1)
>
> d(mu) = dAbar/(Zbar + 1) + Abar/(Zbar + 1)**2 d(Zbar)
>
> the routines return derivatives with respect
> to composition like dP/d(Abar) and dP/d(Zbar).
>
> given the relation between mu, Abar and Zbar
> can you fill in the rest?
>
> fxt
>
>
>
>> I mean the mean mass per particle, like, 1/mu = 1/mu_i + 1/mu_e.
>
>
>

Re: mesa-users] mesa eos needs to be extended into the very low T and low density region

[Bill P. <pa...@ki...>]

Hi Falk (and Didier),

> On May 4, 2008, at 7:19 AM, Falk Herwig wrote:
>> Bill, I am forwarding your request to Didier Saumon at Los Alamos.

Thanks Falk -- that's an excellent idea!

Didier, as you may recall, the mesa/eos is a blended combination of  
Frank Timmes' HELM for high T with OPAL and SCVH for lower T.  The  
missing piece is the region that is below OPAL in temperature and  
below SCVH in density.  I'm running models out to the photosphere  
(gray atmosphere, optical depth of 2/3), and with extremely low  
masses (0.02Msun in the example in my original email), and the  
surface of the model is extending into the missing region.  Ideally  
we'd have a continuation of SCVH to much lower densities and/or a  
continuation of OPAL to much lower temperatures.  I think the main  
deviation from ideal gas that comes into play is the H to H2  
transition.  I'm sure there are lots of other difficult things to  
deal with in a full solution, but neutral gas of atoms with H and H2  
in transition would be a wonderful first step.

Cheers,
Bill

mesa-users] Fwd: dlnP/dlnMu ?

[Bill P. <pa...@ki...>]

Hi Frank,

Perhaps you can enlighten us on this one.

Thanks,
Bill

Begin forwarded message:

> From: Ken Shen <ke...@ph...>
> Date: May 3, 2008 9:49:04 PM PDT
> To: Bill Paxton <pa...@ki...>
> Subject: dlnP/dlnMu ?
>
> Hi Bill.  In the MESA EoS stuff, I know I can get dlnP/dlnRho at  
> constant
> T, for example, by calling eos_get_basic_results.  But is there an  
> easy
> way to get dlnP/dlnMu at constant rho and T (i.e., chi_mu)?  I'm  
> trying to
> find it so that I can calculate the Brunt frequency for convection...
>
>
> Thanks,
> Ken
>

Re: mesa-users] Problems with convection package

[Bill P. <pa...@ki...>]

On May 14, 2007, at 3:46 AM, astrophysics wrote:

> Hello.
> I have a problem with extracting con.f file from =20
> mesa_con-0.1.tar.gz package.
> That file is probably corrupted. Could you verify it?
> Thanks a lot.
>

Hello,

That file seems to download okay for me -- strange....   Have you =20
been able to download other packages?

In the meantime, I've attached the tar file and the source files for =20
mesa_con-0.1.

=EF=BF=BC
=EF=BF=BC=EF=BF=BC=EF=BF=BC
Let me know how it goes....


BTW, your email doesn't include your name -- I'd enjoy knowing a bit =20
more such as who you are and what you'd like to do with mesa.  I hope =20=

you understand that mesa is still in early stages of development and =20
as a result is rather unstable.  You are welcome to pick up packages =20
and use them however you like, but use them carefully and with a =20
degree of skepticism appropriate for "raw" code!

Cheers,
Bill Paxton

mesa-users] Problems with convection package

[astrophysics <ast...@o2...>]

Hello.
I have a problem with extracting con.f file from mesa_con-0.1.tar.gz =
package.
That file is probably corrupted. Could you verify it?
Thanks a lot.