[FLASH-USERS] FLASH scalability on parallel computing test

John Zuhone jzuhone at gmail.com
Wed Jan 8 12:39:43 EST 2025


 Hi Ernesto,

Yes, typically HPC resources will allow you to request a small start-up
allocation in order to do such scaling tests on their own machine before
you ask for a larger allocation to do production runs. I imagine that they
will be very reluctant to accept a scaling test from another architecture.

Best,

John ZuHone

On Jan 8, 2025 at 12:34:20 PM, Ryan Farber <rjfarber at umich.edu> wrote:

> Hi Ernesto,
>
> You should perform your scaling study on the machine itself and *not*
> recycle old scaling plots, as the scaling will depend on the properties of
> the machine (e.g., interconnect hardware, specific MPI library plus wrapper
> (intel vs gcc etc)).
>
> But for inspiration, you can search the FLASH mailing list archive and I
> think circa 2016 Sean Couch still has the most impressive scaling plot I’ve
> seen.
>
> Best wishes,
> Ryan
>
> Sent from my iPhone
>
> On 8. Jan 2025, at 06:05, ERNESTO ZURBRIGGEN <ezurbriggen at unc.edu.ar>
> wrote:
>
> 
> Hello everyone!
> I am writing a proposal for HCP computing in a cluster with FLASH as the
> main code to be used. And I was wondering where I can find an MPI
> scalability test to include in the proposal. This is mandatory in the call.
>
> I thought I had seen a scalability plot in the user manual, but I can't
> find it.
>
> Best regards!
>
>
> --
>
> *PhD Ernesto Zurbriggen,*
> *Assistant researcher (CONICET)*
>
>
>
> *Email: ezurbriggen@ <ernesto.zurbriggen at craam.mackenzie.br>unc.edu.ar
> <http://unc.edu.ar/>IP phone: (+54) 351-535 3776 (int. 75601)*
> *Instituto de Astronomía Teórica y Experimental (IATE-CONICET), *
> *Francisco Laprida sth 922**, X5000BGR, **Córdoba, **Argentina. *
> _______________________________________________
> flash-users mailing list
> flash-users at flash.rochester.edu
>
> For list info, including unsubscribe:
> https://flash.rochester.edu/mailman/listinfo/flash-users
>
> _______________________________________________
> flash-users mailing list
> flash-users at flash.rochester.edu
>
> For list info, including unsubscribe:
> https://flash.rochester.edu/mailman/listinfo/flash-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20250108/950020fa/attachment.htm>


More information about the flash-users mailing list