<div dir="ltr"><div dir="ltr">On Thu, Oct 10, 2019 at 8:41 AM Dave May <<a href="mailto:dave.mayhem23@gmail.com">dave.mayhem23@gmail.com</a>> wrote:<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"><div><div>On Thu 10. Oct 2019 at 14:34, Matthew Knepley <<a href="mailto:knepley@gmail.com" target="_blank">knepley@gmail.com</a>> wrote:<br></div></div><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">On Thu, Oct 10, 2019 at 8:31 AM Dave May <<a href="mailto:dave.mayhem23@gmail.com" target="_blank">dave.mayhem23@gmail.com</a>> wrote:<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"><div dir="ltr"><div dir="ltr">On Thu, 10 Oct 2019 at 13:21, Matthew Knepley via petsc-users <<a href="mailto:petsc-users@mcs.anl.gov" target="_blank">petsc-users@mcs.anl.gov</a>> wrote:<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"><div dir="ltr"><div dir="ltr">On Wed, Oct 9, 2019 at 5:10 PM Danyang Su via petsc-users <<a href="mailto:petsc-users@mcs.anl.gov" target="_blank">petsc-users@mcs.anl.gov</a>> wrote:<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">
<div bgcolor="#FFFFFF">
<p>Dear All,</p>
<p>I have a question regarding the maximum memory usage for the
scaling test. My code is written in Fortran with support for both
structured grid (DM) and unstructured grid (DMPlex). It looks like
memory consumption is much larger when DMPlex is used and finally
causew out_of_memory problem. <br>
</p>
<p>Below are some test using both structured grid and unstructured
grid. The memory consumption by the code is estimated based on all
allocated arrays and PETSc memory consumption is estimated based
on PetscMemoryGetMaximumUsage. <br>
</p>
<p>I just wonder why the PETSc memory consumption does not decrease
when number of processors increases. For structured grid (scenario
7-9), the memory consumption decreases as number of processors
increases. However, for unstructured grid case (scenario 14-16),
the memory for PETSc part
remains unchanged. When I run a larger case, the code crashes
because memory is ran out. The same case works on another cluster
with 480GB memory per node. Does this make sense?<br></p></div></blockquote><div>We would need a finer breakdown of where memory is being used. I did this for a paper:</div><div><br></div><div> <a href="https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/jgrb.50217" target="_blank">https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/jgrb.50217</a></div><div><br></div><div>If the subdomains, the halo sizes can overwhelm the basic storage. It looks like the subdomains are big here,</div><div>but things are not totally clear to me. It would be helpful to send the output of -log_view for each case since</div><div>PETSc tries to keep track of allocated memory. </div></div></div></blockquote><div><br></div><div>Matt - I'd guess that there is a sequential (non-partitioned) mesh hanging around in memory.</div><div>Is it possible that he's created the PLEX object which is loaded sequentially (stored and retained in memory and never released), and then afterwards distributed?</div><div>This can never happen with the DMDA and the table verifies this.</div><div>If his code using the DMDA and DMPLEX are as identical as possible (albeit the DM used), then a sequential mesh held in memory seems the likely cause.</div></div></div></blockquote><div><br></div><div>Dang it, Dave is always right.</div><div><br></div><div>How to prevent this? </div></div></div></blockquote><div dir="auto"><br></div></div><div><div dir="auto">I thought you/Lawrence/Vaclav/others... had developed and provided support for a parallel DMPLEX load via a suitably defined plex specific H5 mesh file.</div></div></blockquote><div><br></div><div>We have, but these tests looked like generated meshes.</div><div><br></div><div> Thanks,</div><div><br></div><div> Matt</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div><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 dir="ltr"><div class="gmail_quote"><div>Since it looks like you are okay with fairly regular meshes, I would construct the</div><div>coarsest mesh you can, and then use</div><div><br></div><div> -dm_refine <k></div><div><br></div><div>which is activated by DMSetFromOptions(). Make sure to call it after DMPlexDistribute(). It will regularly</div><div>refine in parallel and should show good memory scaling as Dave says.</div><div><br></div><div> Thanks,</div><div><br></div><div> Matt </div></div></div><div dir="ltr"><div class="gmail_quote"><div> </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 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 dir="ltr"><div class="gmail_quote"><div><br></div><div> Thanks,</div><div><br></div><div> Matt</div><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>
</p>
<p> </p>
<table cellspacing="0" border="0">
<colgroup width="85" span="5"></colgroup> <colgroup width="83"></colgroup>
<colgroup width="125"></colgroup> <colgroup width="211"></colgroup>
<colgroup width="125"></colgroup> <colgroup width="175"></colgroup>
<colgroup width="119"></colgroup> <tbody>
<tr>
<td height="17" align="right">scenario</td>
<td align="right">no. points</td>
<td align="right">cell type</td>
<td align="right">DMPLex</td>
<td align="right">nprocs</td>
<td align="right">no. nodes</td>
<td align="right">mem per node GB</td>
<td align="right">solver</td>
<td align="right">Rank 0 memory MB</td>
<td align="right">Rank 0 petsc memory MB</td>
<td align="right">Runtime (sec)</td>
</tr>
<tr>
<td height="17" align="right">1</td>
<td align="right">2121</td>
<td align="right">rectangle</td>
<td align="right">no</td>
<td align="right">40</td>
<td align="right">1</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right">0.21</td>
<td align="right">41.6</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right">2</td>
<td align="right">8241</td>
<td align="right">rectangle</td>
<td align="right">no</td>
<td align="right">40</td>
<td align="right">1</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right">0.59</td>
<td align="right">51.84</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right">3</td>
<td align="right">32481</td>
<td align="right">rectangle</td>
<td align="right">no</td>
<td align="right">40</td>
<td align="right">1</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right">1.95</td>
<td align="right">59.1</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right">4</td>
<td align="right">128961</td>
<td align="right">rectangle</td>
<td align="right">no</td>
<td align="right">40</td>
<td align="right">1</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right">7.05</td>
<td align="right">89.71</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right">5</td>
<td align="right">513921</td>
<td align="right">rectangle</td>
<td align="right">no</td>
<td align="right">40</td>
<td align="right">1</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right">26.76</td>
<td align="right">110.58</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right">6</td>
<td align="right">2051841</td>
<td align="right">rectangle</td>
<td align="right">no</td>
<td align="right">40</td>
<td align="right">1</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right">104.21</td>
<td align="right">232.05</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right"><b>7</b></td>
<td align="right"><b>8199681</b></td>
<td align="right"><b>rectangle</b></td>
<td align="right"><b>no</b></td>
<td align="right"><b>40</b></td>
<td align="right"><b>1</b></td>
<td align="right"><b>200</b></td>
<td align="right"><b>GMRES,Hypre preconditioner</b></td>
<td align="right"><b>411.26</b></td>
<td align="right"><b>703.27</b></td>
<td align="right"><b>140.29</b></td>
</tr>
<tr>
<td height="17" align="right"><b>8</b></td>
<td align="right"><b>8199681</b></td>
<td align="right"><b>rectangle</b></td>
<td align="right"><b>no</b></td>
<td align="right"><b>80</b></td>
<td align="right"><b>2</b></td>
<td align="right"><b>200</b></td>
<td align="right"><b>GMRES,Hypre preconditioner</b></td>
<td align="right"><b>206.6</b></td>
<td align="right"><b>387.25</b></td>
<td align="right"><b>62.04</b></td>
</tr>
<tr>
<td height="17" align="right"><b>9</b></td>
<td align="right"><b>8199681</b></td>
<td align="right"><b>rectangle</b></td>
<td align="right"><b>no</b></td>
<td align="right"><b>160</b></td>
<td align="right"><b>4</b></td>
<td align="right"><b>200</b></td>
<td align="right"><b>GMRES,Hypre preconditioner</b></td>
<td align="right"><b>104.28</b></td>
<td align="right"><b>245.3</b></td>
<td align="right"><b>32.76</b></td>
</tr>
<tr>
<td height="17" align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right">10</td>
<td align="right">2121</td>
<td align="right">triangle</td>
<td align="right">yes</td>
<td align="right">40</td>
<td align="right">1</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right">0.49</td>
<td align="right">61.78</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right">11</td>
<td align="right">15090</td>
<td align="right">triangle</td>
<td align="right">yes</td>
<td align="right">40</td>
<td align="right">1</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right">2.32</td>
<td align="right">96.61</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right">12</td>
<td align="right">59847</td>
<td align="right">triangle</td>
<td align="right">yes</td>
<td align="right">40</td>
<td align="right">1</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right">8.28</td>
<td align="right">176.14</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right">13</td>
<td align="right">238568</td>
<td align="right">triangle</td>
<td align="right">yes</td>
<td align="right">40</td>
<td align="right">1</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right">31.89</td>
<td align="right">573.73</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right"><b>14</b></td>
<td align="right"><b>953433</b></td>
<td align="right"><b>triangle</b></td>
<td align="right"><b>yes</b></td>
<td align="right"><b>40</b></td>
<td align="right"><b>1</b></td>
<td align="right"><b>200</b></td>
<td align="right"><b>GMRES,Hypre preconditioner</b></td>
<td align="right"><b>119.23</b></td>
<td align="right"><b>2102.54</b></td>
<td align="right"><b>44.11</b></td>
</tr>
<tr>
<td height="17" align="right"><b>15</b></td>
<td align="right"><b>953433</b></td>
<td align="right"><b>triangle</b></td>
<td align="right"><b>yes</b></td>
<td align="right"><b>80</b></td>
<td align="right"><b>2</b></td>
<td align="right"><b>200</b></td>
<td align="right"><b>GMRES,Hypre preconditioner</b></td>
<td align="right"><b>72.99</b></td>
<td align="right"><b>2123.8</b></td>
<td align="right"><b>24.36</b></td>
</tr>
<tr>
<td height="17" align="right"><b>16</b></td>
<td align="right"><b>953433</b></td>
<td align="right"><b>triangle</b></td>
<td align="right"><b>yes</b></td>
<td align="right"><b>160</b></td>
<td align="right"><b>4</b></td>
<td align="right"><b>200</b></td>
<td align="right"><b>GMRES,Hypre preconditioner</b></td>
<td align="right"><b>48.65</b></td>
<td align="right"><b>2076.25</b></td>
<td align="right"><b>14.87</b></td>
</tr>
<tr>
<td height="17" align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right">17</td>
<td align="right">55770</td>
<td align="right">prism</td>
<td align="right">yes</td>
<td align="right">40</td>
<td align="right">1</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right">18.46</td>
<td align="right">219.39</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right">18</td>
<td align="right">749814</td>
<td align="right">prism</td>
<td align="right">yes</td>
<td align="right">40</td>
<td align="right">1</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right">149.86</td>
<td align="right">2412.39</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right">19</td>
<td align="right">7000050</td>
<td align="right">prism</td>
<td align="right">yes</td>
<td align="right">40 to 640</td>
<td align="right">1 to 16</td>
<td align="right">200</td>
<td align="right">GMRES,Hypre preconditioner</td>
<td align="right"><br>
</td>
<td align="right">out_of_memory</td>
<td align="left"><br>
</td>
</tr>
<tr>
<td height="17" align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
<td align="right"><br>
</td>
</tr>
<tr>
<td height="17" align="right"><b>20</b></td>
<td align="right"><b>7000050</b></td>
<td align="right"><b>prism</b></td>
<td align="right"><b>yes</b></td>
<td align="right"><b>64</b></td>
<td align="right"><b>2</b></td>
<td align="right"><b>480</b></td>
<td align="right"><b>GMRES,Hypre preconditioner</b></td>
<td align="right"><b>890.92</b></td>
<td align="right"><b>17214.41</b></td>
<td align="right"><br>
</td>
</tr>
</tbody>
</table>
<p>
</p>
<p>The error information of scenario 19 is shown below:<br>
</p>
<p>kernel messages produced during job executions:<br>
[Oct 9 10:41] mpiexec.hydra invoked oom-killer: gfp_mask=0x200da,
order=0, oom_score_adj=0<br>
[ +0.010274] mpiexec.hydra cpuset=/ mems_allowed=0-1<br>
[ +0.006680] CPU: 2 PID: 144904 Comm: mpiexec.hydra Tainted:
G OE ------------ 3.10.0-862.14.4.el7.x86_64 #1<br>
[ +0.013365] Hardware name: Lenovo ThinkSystem SD530
-[7X21CTO1WW]-/-[7X21CTO1WW]-, BIOS -[TEE124N-1.40]- 06/12/2018<br>
[ +0.012866] Call Trace:<br>
[ +0.003945] [<ffffffffb3313754>] dump_stack+0x19/0x1b<br>
[ +0.006995] [<ffffffffb330e91f>] dump_header+0x90/0x229<br>
[ +0.007121] [<ffffffffb2cfa982>] ?
ktime_get_ts64+0x52/0xf0<br>
[ +0.007451] [<ffffffffb2d5141f>] ?
delayacct_end+0x8f/0xb0<br>
[ +0.007393] [<ffffffffb2d9ac94>]
oom_kill_process+0x254/0x3d0<br>
[ +0.007592] [<ffffffffb2d9a73d>] ?
oom_unkillable_task+0xcd/0x120<br>
[ +0.007978] [<ffffffffb2d9a7e6>] ?
find_lock_task_mm+0x56/0xc0<br>
[ +0.007729] [<ffffffffb2d9b4d6>] <b>out_of_memory+0x4b6/0x4f0</b><br>
[ +0.007358] [<ffffffffb330f423>]
__alloc_pages_slowpath+0x5d6/0x724<br>
[ +0.008190] [<ffffffffb2da18b5>]
__alloc_pages_nodemask+0x405/0x420</p>
<p>Thanks,</p>
<p>Danyang<br>
</p>
</div>
</blockquote></div><br clear="all"><div><br></div>-- <br><div dir="ltr"><div dir="ltr"><div><div dir="ltr"><div><div dir="ltr"><div>What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>-- Norbert Wiener</div><div><br></div><div><a href="http://www.cse.buffalo.edu/~knepley/" target="_blank">https://www.cse.buffalo.edu/~knepley/</a><br></div></div></div></div></div></div></div></div>
</blockquote></div></div>
</blockquote></div><br clear="all"><div><br></div>-- <br><div dir="ltr"><div dir="ltr"><div><div dir="ltr"><div><div dir="ltr"><div>What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>-- Norbert Wiener</div><div><br></div><div><a href="http://www.cse.buffalo.edu/~knepley/" target="_blank">https://www.cse.buffalo.edu/~knepley/</a><br></div></div></div></div></div></div></div></div>
</blockquote></div></div>
</div>
</blockquote></div><br clear="all"><div><br></div>-- <br><div dir="ltr" class="gmail_signature"><div dir="ltr"><div><div dir="ltr"><div><div dir="ltr"><div>What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>-- Norbert Wiener</div><div><br></div><div><a href="http://www.cse.buffalo.edu/~knepley/" target="_blank">https://www.cse.buffalo.edu/~knepley/</a><br></div></div></div></div></div></div></div></div>