<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body>
<div dir="ltr">
<div dir="ltr">
<div dir="ltr">
<div dir="ltr">
<div dir="ltr">Myriam:<br>
</div>
<div>Checking MatPtAP() in petsc-3.6.4, I realized that it uses different algorithm than petsc-10 and later versions. petsc-3.6 uses out-product for C=P^T * AP, while petsc-3.10 uses local transpose of P. petsc-3.10 accelerates data accessing, but doubles the
memory of P. </div>
<div><br>
</div>
<div>Fande added two new implementations for MatPtAP() to petsc-master which use much smaller and scalable memories with slightly higher computing time (faster than hypre though). You may use these new implementations if you have concern on memory scalability.
The option for these new implementation are: </div>
<div>-matptap_via allatonce<br>
</div>
<div>-matptap_via allatonce_merged<br>
</div>
<div><br>
</div>
<div>Hong</div>
<br>
<div class="gmail_quote">
<div dir="ltr" class="gmail_attr">On Mon, Apr 15, 2019 at 12:10 PM <a href="mailto:hzhang@mcs.anl.gov" target="_blank">
hzhang@mcs.anl.gov</a> <<a href="mailto:hzhang@mcs.anl.gov" target="_blank">hzhang@mcs.anl.gov</a>> wrote:<br>
</div>
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<div dir="ltr">
<div dir="ltr">Myriam:<br>
</div>
<div>Thank you very much for providing these results!</div>
<div>I have put effort to accelerate execution time and avoid using global sizes in PtAP, for which the algorithm of transpose of P_local and P_other likely doubles the memory usage. I'll try to investigate why it becomes unscalable.</div>
<div>Hong</div>
<div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<div bgcolor="#FFFFFF">
<p>Hi,</p>
<p>you'll find the new scaling attached (green line). I used the version 3.11 and the four scalability options :<br>
-matptap_via scalable<br>
-inner_diag_matmatmult_via scalable<br>
-inner_offdiag_matmatmult_via scalable<br>
-mat_freeintermediatedatastructures</p>
<p>The scaling is much better! The code even uses less memory for the smallest cases. There is still an increase for the larger one.
<br>
</p>
<p>With regard to the time scaling, I used KSPView and LogView on the two previous scalings (blue and yellow lines) but not on the last one (green line). So we can't really compare them, am I right? However, we can see that the new time scaling looks quite
good. It slightly increases from ~8s to ~27s. <br>
</p>
<p>Unfortunately, the computations are expensive so I would like to avoid re-run them if possible. How relevant would be a proper time scaling for you?
<br>
</p>
<p>Myriam<br>
</p>
<br>
<div class="gmail-m_5870970137787136754gmail-m_4593329201565690262m_-4364359315279719822gmail-m_-6245019727744503832moz-cite-prefix">
Le 04/12/19 à 18:18, Zhang, Hong a écrit :<br>
</div>
<blockquote type="cite">
<div dir="ltr">
<div dir="ltr">Myriam :<br>
</div>
<div>Thanks for your effort. It will help us improve PETSc.</div>
<div>Hong</div>
<div><br>
</div>
<div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
Hi all,<br>
<br>
I used the wrong script, that's why it diverged... Sorry about that. <br>
I tried again with the right script applied on a tiny problem (~200<br>
elements). I can see a small difference in memory usage (gain ~ 1mB).<br>
when adding the -mat_freeintermediatestructures option. I still have to<br>
execute larger cases to plot the scaling. The supercomputer I am used to<br>
run my jobs on is really busy at the moment so it takes a while. I hope<br>
I'll send you the results on Monday.<br>
<br>
Thanks everyone,<br>
<br>
Myriam<br>
<br>
<br>
Le 04/11/19 à 06:01, Jed Brown a écrit :<br>
> "Zhang, Hong" <<a href="mailto:hzhang@mcs.anl.gov" target="_blank">hzhang@mcs.anl.gov</a>> writes:<br>
><br>
>> Jed:<br>
>>>> Myriam,<br>
>>>> Thanks for the plot. '-mat_freeintermediatedatastructures' should not affect solution. It releases almost half of memory in C=PtAP if C is not reused.<br>
>>> And yet if turning it on causes divergence, that would imply a bug.<br>
>>> Hong, are you able to reproduce the experiment to see the memory<br>
>>> scaling?<br>
>> I like to test his code using an alcf machine, but my hands are full now. I'll try it as soon as I find time, hopefully next week.<br>
> I have now compiled and run her code locally.<br>
><br>
> Myriam, thanks for your last mail adding configuration and removing the<br>
> MemManager.h dependency. I ran with and without<br>
> -mat_freeintermediatedatastructures and don't see a difference in<br>
> convergence. What commands did you run to observe that difference?<br>
<br>
-- <br>
Myriam Peyrounette<br>
CNRS/IDRIS - HLST<br>
--<br>
<br>
<br>
</blockquote>
</div>
</div>
</blockquote>
<br>
<pre class="gmail-m_5870970137787136754gmail-m_4593329201565690262m_-4364359315279719822gmail-m_-6245019727744503832moz-signature" cols="72">--
Myriam Peyrounette
CNRS/IDRIS - HLST
--
</pre>
</div>
</blockquote>
</div>
</div>
</blockquote>
</div>
</div>
</div>
</div>
</div>
</body>
</html>