[Mesa-users] MESA multithreaded performance

[Wolf, William M]

If you’re writing profiles often, storage can start to matter. If you are on a budget, I’d get a smallish SSD that you can put the OS on and where you can run the simulations, but then move them to a cheaper and bigger HDD for storage and analysis.

That being said, the processor is probably far more important in most situations.

Bill



________________________________
From: Mesa-users <mesa-users-bounces at lists.mesastar.org> on behalf of Nicholas Herrington via Mesa-users <mesa-users at lists.mesastar.org>
Sent: Friday, July 3, 2020 2:24:30 PM
To: Meisel, Zach <meisel at ohio.edu>; Rob Farmer <r.j.farmer at uva.nl>; sakstein at hawaii.edu <sakstein at hawaii.edu>
Cc: mesa-users at lists.mesastar.org <mesa-users at lists.mesastar.org>
Subject: Re: [Mesa-users] MESA multithreaded performance

Hi Zach, Rob and Jeremy

Thank you for your responses, I will probably go with more cores.

Also has anyone found any difference between running mesa from an SSD or HDD?

Cheers
Nick


On Thu, 2 Jul 2020 at 15:31, Meisel, Zach via Mesa-users <mesa-users at lists.mesastar.org<mailto:mesa-users at lists.mesastar.org>> wrote:

For what it’s worth, for X-ray bursts I find diminishing returns beyond 4 cores.



-Zach



----
Zach Meisel
Associate Professor of Physics & Astronomy

Director, Edwards Accelerator Laboratory
Ohio University
Email: meisel at ohio.edu<mailto:meisel at ohio.edu>
Web: http://inpp.ohio.edu/~meisel<https://nam03.safelinks.protection.outlook.com/?url=http:%2F%2Finpp.ohio.edu%2F~meisel&data=02%7C01%7Cwolfwm%40uwec.edu%7C70e27854333c46e824be08d81f87c2c5%7Cdd068b97759349388b3214faef2af1d8%7C0%7C0%7C637294015230626595&sdata=BPR3dXLfwgq7mQOQ8aszSlqWcDLxfazR581laooS1m8%3D&reserved=0>



From: Rob Farmer via Mesa-users<mailto:mesa-users at lists.mesastar.org>
Sent: Thursday, July 2, 2020 10:23 AM
To: Jeremy Sakstein<mailto:sakstein at hawaii.edu>
Cc: Nicholas.Herrington at myport.ac.uk<mailto:Nicholas.Herrington at myport.ac.uk>; mesa-users at lists.mesastar.org<mailto:mesa-users at lists.mesastar.org>
Subject: Re: [Mesa-users] MESA multithreaded performance



Hi,



There is a discussion of this in section 6 of https://ui.adsabs.harvard.edu/abs/2019ApJS..243...10P<https://nam03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fui.adsabs.harvard.edu%2Fabs%2F2019ApJS..243...10P&data=02%7C01%7Cwolfwm%40uwec.edu%7C70e27854333c46e824be08d81f87c2c5%7Cdd068b97759349388b3214faef2af1d8%7C0%7C0%7C637294015230626595&sdata=A%2FUILCZtV3oR8vIg6rMBnyBhewk4H%2BckEbAJx%2B2y7CA%3D&reserved=0>



The short answer depends on the problem, but you're probably maxing the benefit at ~12 threads for most problems.  I would usually say if you're running one mesa problem at a time get a faster cpu, if you're running multiple mesas at once then get more cores.



Rob



On Thu, 2 Jul 2020 at 16:01, Jeremy Sakstein <sakstein at hawaii.edu<mailto:sakstein at hawaii.edu>> wrote:

Hi Nick,



In my experience this depends on what type of star you are running. Using both r12266 and r12778 I found:



solar mass main-sequence/RGB before helium burning/low-Z blue loops worked best with about 4 cores.



RSP test_suite cases didn't seem to use more than two cores even when I requested more.



Big jobs like the PPISN test_suite worked best with 16 cores.



I hope this helps, I'm sure people who have tested this more systematically than me have a better idea.



Cheers,



Jeremy







On Thu, Jul 2, 2020 at 9:54 AM Nicholas Herrington via Mesa-users <mesa-users at lists.mesastar.org<mailto:mesa-users at lists.mesastar.org>> wrote:

Hi all



I just had a quick question. I am upgrading my Linux machine and I’m just wondering from people’s experience does MESA favour more cpu cores or faster single core, if that makes sense?



Trying to decide between two CPUs of same generation, one has less cores but higher clock speeds and the other has more cores but lower clock speeds.



Thanks

Nick

_______________________________________________
mesa-users at lists.mesastar.org<mailto:mesa-users at lists.mesastar.org>
https://lists.mesastar.org/mailman/listinfo/mesa-users<https://nam03.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.mesastar.org%2Fmailman%2Flistinfo%2Fmesa-users&data=02%7C01%7Cwolfwm%40uwec.edu%7C70e27854333c46e824be08d81f87c2c5%7Cdd068b97759349388b3214faef2af1d8%7C0%7C0%7C637294015230626595&sdata=IG8w5eDAw4%2BLuEAJ4Gi68t6cfiT8ABtoUb1rxoKWoBw%3D&reserved=0>



_______________________________________________
mesa-users at lists.mesastar.org<mailto:mesa-users at lists.mesastar.org>
https://lists.mesastar.org/mailman/listinfo/mesa-users<https://nam03.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.mesastar.org%2Fmailman%2Flistinfo%2Fmesa-users&data=02%7C01%7Cwolfwm%40uwec.edu%7C70e27854333c46e824be08d81f87c2c5%7Cdd068b97759349388b3214faef2af1d8%7C0%7C0%7C637294015230636581&sdata=fxmbazuzrVB8GMwTOlk7E9097pwdtFXKDJu7iM9Cu7g%3D&reserved=0>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20200703/61d23e3c/attachment.htm>