Strange wall time measurements for code profiling SU2

imandralis · May 23, 2021, 08:31

Dear all,

I am trying to test the performance of the SU2 code. To do this I need to measure the wall time taken per iteration. I am using a dual time-stepping approach.

I noticed that for each new outer iteration, the wall time elapsed seems to jump and then decrease for each inner iteration. This doesn't happen for the first and last outer iteration of the simulation.

Is it possible that there is a problem with the wall time screen output, or that I am not choosing the correct options in the configuration file ?

Thanks in advance.

pcg · May 23, 2021, 11:25

The wall time output is an average per iteration.
The total time of one outer iteration is the last wall time value times N+1 (N the inner iteration).

pcg · May 23, 2021, 11:29

By the way, if you are measuring the performance of SU2 please make sure you use the best configuration possible for that
https://su2code.github.io/docs_v7/Bu...-optimizations

imandralis · May 23, 2021, 11:54

Dear Pedro,

Thank you very much for your response - that clarifies how to test the time taken per iteration. I will make sure I am using all the correct optimizations.

May 23, 2021, 08:31	Strange wall time measurements for code profiling SU2	#1
imandralis New Member Ioannis Mandralis Join Date: May 2021 Posts: 7 Rep Power: 5	Dear all, I am trying to test the performance of the SU2 code. To do this I need to measure the wall time taken per iteration. I am using a dual time-stepping approach. I noticed that for each new outer iteration, the wall time elapsed seems to jump and then decrease for each inner iteration. This doesn't happen for the first and last outer iteration of the simulation. Is it possible that there is a problem with the wall time screen output, or that I am not choosing the correct options in the configuration file ? Thanks in advance.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
bash script for pseudo-parallel usage of reconstructPar	kwardle	OpenFOAM Post-Processing	42	May 8, 2024 00:17
SU2 7.0.7 Built on CentOS 7, parallel computation pyscript mpi exit error?	EternalSeekerX	SU2	3	October 9, 2020 19:28
pressure in incompressible solvers e.g. simpleFoam	chrizzl	OpenFOAM Running, Solving & CFD	13	March 28, 2017 06:49
Micro Scale Pore, icoFoam	gooya_kabir	OpenFOAM Running, Solving & CFD	2	November 2, 2013 14:58
Upgraded from Karmic Koala 9.10 to Lucid Lynx10.04.3	bookie56	OpenFOAM Installation	8	August 13, 2011 05:03

May 23, 2021, 11:25		#2
pcg Senior Member Pedro Gomes Join Date: Dec 2017 Posts: 466 Rep Power: 14	The wall time output is an average per iteration. The total time of one outer iteration is the last wall time value times N+1 (N the inner iteration).

May 23, 2021, 11:29		#3
pcg Senior Member Pedro Gomes Join Date: Dec 2017 Posts: 466 Rep Power: 14	By the way, if you are measuring the performance of SU2 please make sure you use the best configuration possible for that https://su2code.github.io/docs_v7/Bu...-optimizations

May 23, 2021, 11:54		#4
imandralis New Member Ioannis Mandralis Join Date: May 2021 Posts: 7 Rep Power: 5	Dear Pedro, Thank you very much for your response - that clarifies how to test the time taken per iteration. I will make sure I am using all the correct optimizations.