[mesa-users] "failed to find all required l=0 modes"
Warrick Ball
wball at astro.physik.uni-goettingen.de
Wed Nov 12 01:16:49 EST 2014
Rich,
I tested 6208, and it too has no instances of the warning and a better fit
than 7184.
chi^2 combined 4.4698878897
chi2_seismo_fraction 0.6670000000
chi^2 seismo 5.9755510897
chi^2 spectro 1.4540399785
Curiously, this is also (slightly) better than 6794... Anyway, I'll also
start hunting for a difference.
W
On Tue, 11 Nov 2014, Richard Townsend wrote:
>> On Nov 11, 2014, at 10:47 AM, Bill Paxton <paxton at kitp.ucsb.edu> wrote:
>>
>> Have you guys considered the option that the message isn't bogus?
>>
>>>>> "failed to find all required l=0 modes"
>>
>> You've set a limited range of frequencies to consider (e.g. freq_min and freq_max in gyre.in).
>>
>> It is quite possible that the required l=0 modes don't fall within that limited range.
>>
>> in that case, you'll fail to find all required l=0 modes. right?
>>
>> perhaps that's what the message is trying to tell you.
>>
>> or perhaps there's a bug.
>>
>> or both.
>>
>
> OK, I’ve compared the output from star/astero/work in the latest release of MESA (7184), against the output from the previous release (6794).
>
> 7184 has many instances of “failed to find all required l=0 modes”, and doesn’t achieve a particularly good minimized chi2:
>
> chi^2 combined 11.6342076263
> chi2_seismo_fraction 0.6670000000
> chi^2 seismo 16.7079283432
> chi^2 spectro 1.4715297940
>
> Whereas 6794 has *no* instances of this warning, and an improved minimized chi2:
>
> chi^2 combined 5.5046603882
> chi2_seismo_fraction 0.6670000000
> chi^2 seismo 7.5258273288
> chi^2 spectro 1.4562569365
>
> The inlists for the two revisions don’t appear to differ substantively. So, I’m not sure where this crept in, but something has gone awry.
>
> The hunt begins. Cry 'Havoc!', and let slip the dogs of diff.
>
> cheers,
>
> Rich
>
>> -b
>>
>>
>>
>>
>>
>>
>>
>> On Nov 11, 2014, at 7:43 AM, Richard Townsend wrote:
>>
>>> Hi Warrick —
>>>
>>> This is interesting; when running the example in star/astero/work, I find lots of similar occurrences of “failed to find all required l=0 modes”. I’m not sure whether these existed in older MESA versions; I’ll check. If not, then it looks like we have a regression which needs tracking down.
>>>
>>> cheers,
>>>
>>> Rich
>>>
>>>> On Nov 11, 2014, at 1:28 AM, Warrick Ball <wball at astro.physik.uni-goettingen.de> wrote:
>>>>
>>>> Hi,
>>>>
>>>> In short, I was wrong about how to fix this issue. If get_radial is set to .true., then the code thinks that chi^2 is small enough that it's worth evaluating the non-radial modes, which isn't the case.
>>>>
>>>> So, for now, I'm just sidestepping the issue by cranking up the limit on the number of "bad" models. When I have some time, I'll have another go at trying to understand how things should be restructured to make more sense.
>>>>
>>>> But my previous "fix" is no fix!
>>>>
>>>> Cheers,
>>>> Warrick
>>>>
>>>> On Fri, 7 Nov 2014, Warrick Ball wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> I was looking at the results of a recent asteroseismic fitting run using ersion 7184, and found errors of the form
>>>>>
>>>>> "failed to find all required l=0 modes"
>>>>>
>>>>> after nearly every chi^2 evaluation in the range where only the radial modes are considered. This came as a surprise, since the code was returning a useful value of chi^2.
>>>>>
>>>>> I grepped the error message and identified it's source at line 375 of star/astero/src/extras_support.f. The message is displayed when the function get_radial returns a false. I then had a look at get_radial (line 631).
>>>>>
>>>>> if (chi2_radial > chi2_radial_limit .and. nl1 + nl2 + nl3 > 0) then
>>>>> write(*,'(a50,i6,99f16.2)') &
>>>>> 'chi2_radial > chi2_radial_limit', &
>>>>> s% model_number, chi2_radial, chi2_radial_limit
>>>>> return
>>>>> end if
>>>>> get_radial = .true.
>>>>>
>>>>> I might be misinterpreting the code, but it looks to me like this will leave get_radial as false if chi2_radial > chi2_radial_limit, and thus the error will be printed. But the error message doesn't correspond to what's actually happened. That is, the modes have all been found, but the current value of chi^2 is too large for it to be worth continuing to evaluate the higher-degree modes.
>>>>>
>>>>> If I'm right, then the fix is simply to put get_radial = .true. before the last if block. But, as always, I might be wrong...
>>>>>
>>>>> Cheers,
>>>>> Warrick
>>>>>
>>>>>
>>>>>
>>>>> ------------
>>>>> Warrick Ball
>>>>> Postdoc, Institut für Astrophysik Göttingen
>>>>> wball at astro.physik.uni-goettingen.de
>>>>> +49 (0) 551 39 5069
>>>>
>>>>
>>>> ------------
>>>> Warrick Ball
>>>> Postdoc, Institut für Astrophysik Göttingen
>>>> wball at astro.physik.uni-goettingen.de
>>>> +49 (0) 551 39 5069------------------------------------------------------------------------------
>>>> Comprehensive Server Monitoring with Site24x7.
>>>> Monitor 10 servers for $9/Month.
>>>> Get alerted through email, SMS, voice calls or mobile push notifications.
>>>> Take corrective actions from your mobile device.
>>>> http://pubads.g.doubleclick.net/gampad/clk?id=154624111&iu=/4140/ostg.clktrk_______________________________________________
>>>> mesa-users mailing list
>>>> mesa-users at lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/mesa-users
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Comprehensive Server Monitoring with Site24x7.
>>> Monitor 10 servers for $9/Month.
>>> Get alerted through email, SMS, voice calls or mobile push notifications.
>>> Take corrective actions from your mobile device.
>>> http://pubads.g.doubleclick.net/gampad/clk?id=154624111&iu=/4140/ostg.clktrk
>>> _______________________________________________
>>> mesa-users mailing list
>>> mesa-users at lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/mesa-users
>
>
------------
Warrick Ball
Postdoc, Institut für Astrophysik Göttingen
wball at astro.physik.uni-goettingen.de
+49 (0) 551 39 5069
More information about the Mesa-users
mailing list