<div dir="ltr"><div class="gmail_default" style="font-family:arial,helvetica,sans-serif">Thanks, Jed.</div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif">That is helpful. </div></div><div class="gmail_extra"><br clear="all"><div><div class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr"><div><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div><div><font size="2" face="georgia, serif"><b>Ajit Desai</b></font></div><div><span style="font-family:georgia,serif;font-size:small"> PhD Scholar</span><font size="2" face="georgia, serif">, Carleton University, </font>Canada</div></div><div><font face="georgia, serif"><b></b></font></div><span style="font-family:verdana,sans-serif"><span style="font-family:times new roman,serif"></span></span></div></div></div></div></div></div></div></div></div></div></div>
<br><div class="gmail_quote">On Thu, Mar 16, 2017 at 8:00 AM, Jed Brown <span dir="ltr"><<a href="mailto:jed@jedbrown.org" target="_blank">jed@jedbrown.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">Ajit Desai <<a href="mailto:aero.aju@gmail.com">aero.aju@gmail.com</a>> writes:<br>
<br>
> Hello Everyone,<br>
</span>> A couple of questions on the *-log_summary* provided by PETSc.<br>
<br>
-log_view is the preferred name.<br>
<br>
> 1. *Avg-Flops & Avg-Flops/sec* are averaged among the participating cores<br>
<span class="">> or averaged over the simulation time or both?<br>
<br>
</span>Flops on each process is just the count (as logged by PetscLogFlops,<br>
which PETSc numerical functions call and you can too) over the time<br>
between PetscInitialize until the log is printed (usually<br>
PetscFinalize). The time on each process is the wall clock time from<br>
PetscInitialize until the log is printed. The Average, Max, Min, and<br>
total are over processes.<br>
<br>
> 2. *Max/Min Flops & Flops/sec* is an indication of load balancing?<br>
<span class="">> In simulation-2 these ratios are high compared to simulation-1. Does that<br>
> mean simulation-2 is not well balanced?<br>
<br>
</span>Yes, not balanced in terms of flops (correlated with time, but not a<br>
measure of time). Look at the events to help see where.<br>
<span class=""><br>
> Please follow the outputs from two different simulations<br>
> (Note: the problem size and the number of processors used are different).<br>
><br>
</span>> *Simulation-1*<br>
<span class="">> Max Max/Min Avg<br>
> Total<br>
> Time (sec): 4.208e+02 1.00005 4.208e+02<br>
> Objects: 7.100e+01 1.00000 7.100e+01<br>
> Flops: 3.326e+11 1.31175 3.017e+11 4.826e+13<br>
> Flops/sec: 7.904e+08 1.31175 7.169e+08 1.147e+11<br>
> MPI Messages: 0.000e+00 0.00000 0.000e+00 0.000e+00<br>
><br>
><br>
</span>> *SImulation-2:*<br>
<span class="">> Max Max/Min Avg<br>
> Total<br>
> Time (sec): 8.434e+02 1.00000 8.434e+02<br>
> Objects: 7.300e+01 1.02817 7.102e+01<br>
> Flops: 6.555e+11 1.85115 5.798e+11 3.711e+14<br>
> Flops/sec: 7.772e+08 1.85115 6.874e+08 4.400e+11<br>
> MPI Messages: 0.000e+00 0.00000 0.000e+00 0.000e+00<br>
><br>
> We're trying to understand the performance of our solver using these<br>
> outputs.<br>
> Any comments in relation to that will be helpful.<br>
><br>
> Thanks & Regards,<br>
</span>> *Ajit Desai*<br>
<div class="HOEnZb"><div class="h5">> PhD Scholar, Carleton University, Canada<br>
</div></div></blockquote></div><br></div>