[mesa-users] problem with different results on different platforms -- mac vs. linux

[Bill Paxton]

Hi,

Okay, that's some progress.   I've changed the Subject line -- does it now match your current problem?

If so, then I'm going to assume your previous report of problems with history intervals was a user-error false alarm, right?

We need to keep the false alarm rate down on mesa-users, so in the future please double and triple check that
you can reproduce the problem before reporting it as a problem in mesa.  And when you do report it,
include all necessary supporting files so that other users can try to reproduce it on their machines.

I'm still not clear on all the relevant details of your current reported problem.  As Rich and Max have mentioned,
which SDK are you using on the 2 platforms?   same on each or different?   or if you aren't using the SDK on the linux
supercomputer, what compiler are you using there?   if you are using different compilers, then can you rerun
using the same version of gfortran on each?   

-Bill




On Jan 12, 2013, at 10:21 AM, Jakub Ostrowski wrote:

> Hi Bill,
> 
> OK, let's make it clear:
> 
> 1. I can't recreate my initial problem with different tracks with different history_interval. Today, surprisingly, I get the same results for both settings. Yesterday I was able to recreate it a few times. But other problem exists.
> 
> 2. Tracks obtained on my mac are different than yours and different than my tracks calculated on supercomputer powered by Scientific Linux. And this is real problem now. I posted previously HR diagram where it is clearly visible. The discrepancy is not only visible in evolution tracks but also in profiles. So, basically, main goal of this discussion should now be solving this problem. If the discrepancy exists somewhere must be some kind of bug or other problem.
> 
> 3. I can't tell which results are real and why the other one are wrong. It stops my research and this is the most painful thing for me here.
> 
> 4. Yes, it looks like I'm the only person who experiences this issue (or at least is aware of that). But I can see the differences in results, so the problem is real :)
> 
> Did I clarify the situation? I'll be happy to provide any other information that could be helpful to resolve it. Thank you for any help.
> 
> Cheers,
> Jakub
> 
> PS. I was suggested to download a clean copy of MESA. I can't do it now due to a very slow and unstable internet connection in the place I actually am :(
> 
> 
> On 12 sty 2013, at 18:55, Bill Paxton <paxton at kitp.ucsb.edu> wrote:
> 
>> 
>> Jakub,
>> 
>> Would you please state the problem in clear precise terms.
>> The Subject of this series of messages is "critical problem with history interval".
>> Is that still the problem?   If so, it doesn't seem to exist for anyone but you, right?
>> Or is it now something else?  If so, what precisely is it?
>> 
>> The mesa-users community is a valuable resource.  We need to use it well.
>> 
>> -Bill
>> 
>> 
>> 
>> On Jan 12, 2013, at 2:49 AM, Jakub Ostrowski wrote:
>> 
>>> Grant,
>>> 
>>> Thank you for your help. If I hadn't seen big differences in profiles I wouldn't bother you with this. Differences are huge. Models from my mac (these with the weird small loop after helium depletion) burn helium much slower than models from cluster. So, unfortunately this is a real problem.
>>> 
>>> Take care,
>>> Jakub
>>> ------------------------------------------------------------------------------
>>> Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
>>> MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
>>> with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
>>> MVPs and experts. SALE $99.99 this month only -- learn more at:
>>> http://p.sf.net/sfu/learnmore_122912
>>> _______________________________________________
>>> mesa-users mailing list
>>> mesa-users at lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/mesa-users
>> 
>