[mesa-users] Robert Farmer's December 2016 Paper

[Frank Timmes]

> My goal in looking at this is to try to find a way for the Si burning stage to complete with fewer timesteps and so complete quicker.

silicon burning is a slog, although very doable with any net.
there are settings post core o-depletion that can help reduce
the slog and most of these are probably in rob's posted inlists. 
i'm sure several folks would be interested should you find 
meaningful efficiency gains!

fxt





> On Jan 11, 2017, at 7:06 PM, Robert Farmer <rjfarmer at asu.edu> wrote:
> 
> Hi
> 
> >I am also currently running the inlists with your mesa_79.net and I do not get the hydro failures.
> 
> ah, you need to set:
> 
> fe56ec_fake_factor =0
> 
> in your controls inlist (only necessary for approx nets). What this does is turn off a "fake" electron capture rate on fe56, which the approx nets do to turn fe56 to the special isotopes (those isotopes at the end of the approx21_ filename) near core-collapse. This shouldn't be a problem during the pre-MS but the ni56 rate got changed and the 56ec rate is based on fe56ec_fake_factor*ni56_rate so the fe56 starts decaying much sooner than it should.
> 
> If you want to continue to use the approx nets then you'll need to also turn it back on by setting fe56ec_fake_factor=1d-7 once the temperature exceeds ~3d9 (see commit 9155)
> 
> Rob
> 
> ps There is the sdk under cygwin https://sourceforge.net/p/mesa/mailman/message/34636241/ to get crlibm and pgplot, but i think we've had this conversation before :)
> 
> 
> On Wed, Jan 11, 2017 at 6:48 PM, Ian Foley <ifoley2008 at gmail.com> wrote:
> Hi Rob,
> 
> I have attached the inlists.
> 
> I am also currently running the inlists with your mesa_79.net and I do not get the hydro failures.
> 
> I have attached a copy of the terminal output snapshot for mesa_79.net
> 
> Note: I am running mesa on Windows 10 with 4 cores. I therefore cannot use pgstar, but have written a much more primitive graphics program which I can use to give visual snapshots as the simulation proceeds.
> 
> I also cannot use the mesasdk, but do not think this is the source of the problem since the installation tests run pretty well with some differences only in the last couple of decimal places. The test_suite programs run well too. I wish there was a windows version of this.
> 
> Kind regards
> Ian
> 
> On 12 January 2017 at 12:33, Robert Farmer <rjfarmer at asu.edu> wrote:
> Hi Ian
> 
> Can you send me the inlists you've used with your modifications? I've just tried myself with 8845 and the model goes slow but because of the varcontrol rather than the hydro failures with your logs show.
> 
> delta_HR_limit is based on sqrt(delta_logT**2+delta_logL**2) ie the motion on the HR track.
> 
> Rob
> 
> 
> On Wed, Jan 11, 2017 at 5:54 PM, Ian Foley <ifoley2008 at gmail.com> wrote:
> Hi,
> 
> I have tried to use the inlist's that Robert Framer has supplied to run the outputs in his paper. I am using r8845, not r7624 so had to make some minor changes
> 
> Need help because cannot get past the pre_ms stage.
> 
> Around step 70, the luminosity starts to go high, the central temperature drops and the timesteps become small eventually causing program to stop. I have tried many variations to the inlists without being able to overcome this and I am not sure why this is happening.
> 
> In addition, it is puzzling to me how the "delta_HR_limit = 0.0025d0" will work as we are not on the main sequence yet and this inlist entry forces small timesteps in the pre_ms stage.
> 
> I have encountered this problem before and it seemed to be related to a problem with the reaction net.
> 
> I have run the inlists with two approx 21 nets
> 
> new_net_name = 'approx21_plus_co56.net'
> new_net_name = 'approx21_cr60_plus_co56.net' (Used by Robert Farmer)
> 
> They both failed, but the terminal outputs are distinctly different making me think that the reaction net is having quite an impact. Or has this to do with the fact that I am using a later release? Or perhaps something else - like I'm doing something wrong.
> 
> My goal in looking at this is to try to find a way for the Si burning stage to complete with fewer timesteps and so complete quicker. I thought Robert's inlists might give some good clues, but I am stuck at the pre_ms stage.
> 
> I have attached the terminal outputs from the above two nets.
> 15M.txt is for 'approx21_plus_co56.net'
> 15M1.txt is for 'approx21_cr60_plus_co56.net'
> 
> Kind regards
> Ian
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> Developer Access Program for Intel Xeon Phi Processors
> Access to Intel Xeon Phi processor-based developer platforms.
> With one year of Intel Parallel Studio XE.
> Training and support from Colfax.
> Order your platform today. http://sdm.link/xeonphi
> _______________________________________________
> mesa-users mailing list
> mesa-users at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/mesa-users
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> Developer Access Program for Intel Xeon Phi Processors
> Access to Intel Xeon Phi processor-based developer platforms.
> With one year of Intel Parallel Studio XE.
> Training and support from Colfax.
> Order your platform today. http://sdm.link/xeonphi_______________________________________________
> mesa-users mailing list
> mesa-users at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/mesa-users

Frank Timmes, AAS Journals Lead Editor
School of Earth and Space Exploration
Arizona State University
Tempe AZ 85287-1404
Phone: 480-965-4274
Email: frank.timmes at aas.org