[mesa-users] Segfault when redistributing ADIPLS mesh to fewer points

[Bill Paxton]

On Jun 16, 2016, at 9:43 AM, Warrick Ball wrote:

> On the one hand, I guess remeshing onto a sparser mesh is a strange request so could be skipped.
> 
> On the other hand, when I run redistrb.c.in outside of MESA, I don't get an error if I remesh onto fewer points and the "remeshed" model looks OK. (I haven't tested this extensively.)  So I think it'd be more desirable to read the coarser model if that's what the user wants.  But I'm not sure if there's something about how ADIPLS and MESA interact that precludes this.

Hi Warrick,

I think Rob should do the quick solution for now to eliminate the segfault by skipping remesh to smaller grid.
It would be great if you'd like to invest your time in finding a general solution for this. ;)

b











> 
> W
> 
> 
> ------------
> Warrick Ball
> Postdoc, Institut für Astrophysik Göttingen
> wball at astro.physik.uni-goettingen.de
> +49 (0) 551 39 5069
> 
> On Thu, 16 Jun 2016, Bill Paxton wrote:
> 
>> On Jun 16, 2016, at 9:24 AM, Robert Farmer wrote:
>> 
>>      Hi Warwick
>> 
>>      I'll have a look into it.
>> Great.  Since we don't want to go into adipls to make changes if we can avoid it, i'd suggest adding a check in
>> adipls_support to skip the call to srdist in case the new size (nn_new) is not greater than the current (nn).   i
>> think that should work.
>> Thanks,
>> Bill
>> 
>>      Rob
>> 
>>      On Jun 16, 2016 8:43 AM, "Warrick Ball" <wball at astro.physik.uni-goettingen.de> wrote:
>>            Hi all,
>> 
>>            I've just come across some segfaults as a result of accidentally trying to remesh a model
>>            onto a smaller grid for ADIPLS.  I realize this is a slightly odd use case, but I feel
>>            like MESA should rarely segfault.
>> 
>>            To reproduce the error in r7624 with the 20150908 SDK:
>>            - make a copy of the star/astero/work directory,
>>            - set do_redistribute_mesh = .true. in inlist_astero_search_controls,
>>            - change the first number in redistrb.c.pruned.in from 2201 to something smaller than
>>            about 1700 (the first bad model has ~1730 mesh points), and
>>            - run!
>> 
>>            When the code first tries to compute the mode frequencies (once chi2_delta_nu < 20), I get
>>            a segfault:
>> 
>>            Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
>> 
>>            Backtrace for this error:
>>            #0  0x7FF0C5745557
>>            #1  0x7FF0C5745B6E
>>            #2  0x7FF0C482F11F
>>            #3  0x7FF0C5801550
>>            #4  0x7FF0C5801A30
>>            #5  0x7FF0C580F387
>>            #6  0xB25EDC in srdist_
>>            #7  0x432412 in redist_amdl.11484 at adipls_support.f:321
>>            #8  0x432E6D in __adipls_support_MOD_store_model_for_adipls at adipls_support.f:297
>>            #9  0x43154D in __adipls_support_MOD_run_adipls at adipls_support.f:390
>>            #10  0x44BD5B in __astero_support_MOD_get_one_el_info at astero_support.f:144
>>            #11  0x4648A6 in get_radial.11310 at extras_support.f:642
>>            #12  0x45EE77 in __extras_support_MOD_do_astero_extras_check_model at extras_support.f:374
>>            #13  0x4573DA in __extras_support_MOD_astero_extras_check_model at extras_support.f:1432
>>            #14  0x48A189 in __run_star_support_MOD_run1_star
>>            #15  0x472A79 in __run_star_extras_astero_MOD_eval1 at run_star_extras_astero.f:234
>>            #16  0x473652 in __run_star_extras_astero_MOD_do_run_star_astero at
>>            run_star_extras_astero.f:172
>>            #17  0x473967 in MAIN__ at run.f:13
>> 
>>            I'm not sure what the appropriate fix is because I'm not sure exactly how ADIPLS
>>            communicates the results back to MESA.  This doesn't actually affect anything I'm doing. 
>>            Like most reasonable people (?), I usually remesh onto *more* points but I forgot to
>>            change the default in some testing I was doing and stumbled upon this error.
>> 
>>            Cheers,
>>            Warrick
>> 
>>            ------------
>>            Warrick Ball
>>            Postdoc, Institut für Astrophysik Göttingen
>>            wball at astro.physik.uni-goettingen.de
>>            +49 (0) 551 39 5069
>>            ------------------------------------------------------------------------------
>>            What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic
>>            patterns at an interface-level. Reveals which users, apps, and protocols are
>>            consuming the most bandwidth. Provides multi-vendor support for NetFlow,
>>            J-Flow, sFlow and other flows. Make informed decisions using capacity planning
>>            reports. http://pubads.g.doubleclick.net/gampad/clk?id=1444514421&iu=/41014381
>>            _______________________________________________
>>            mesa-users mailing list
>>            mesa-users at lists.sourceforge.net
>>            https://lists.sourceforge.net/lists/listinfo/mesa-users
>> ------------------------------------------------------------------------------
>> What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic
>> patterns at an interface-level. Reveals which users, apps, and protocols are
>> consuming the most bandwidth. Provides multi-vendor support for NetFlow,
>> J-Flow, sFlow and other flows. Make informed decisions using capacity planning
>> reports. http://sdm.link/zohomanageengine_______________________________________________
>> mesa-users mailing list
>> mesa-users at lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/mesa-users