<div dir="ltr">Hi Ou and Patrick,<br><br>Thanks for your quick replies. I&#39;ve been modifying data.exch2 (i.e. uncommenting and commenting out the relevant sections) with respect to the different core numbers, so hopefully that part is okay. <br><br>Here&#39;s one of the STDOUT files from a run with 360 cores:<br><br><a href="https://raw.githubusercontent.com/DanJonesOcean/myscripts/master/STDOUT.0000">https://raw.githubusercontent.com/DanJonesOcean/myscripts/master/STDOUT.0000</a><br><br>Thanks,<br>Dan<br><br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Wed, Dec 30, 2020 at 5:58 PM &lt;<a href="mailto:ecco-support-request@mit.edu">ecco-support-request@mit.edu</a>&gt; 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">Send ecco-support mailing list submissions to<br>
        <a href="mailto:ecco-support@mit.edu" target="_blank">ecco-support@mit.edu</a><br>
<br>
To subscribe or unsubscribe via the World Wide Web, visit<br>
        <a href="http://mailman.mit.edu/mailman/listinfo/ecco-support" rel="noreferrer" target="_blank">http://mailman.mit.edu/mailman/listinfo/ecco-support</a><br>
or, via email, send a message with subject or body &#39;help&#39; to<br>
        <a href="mailto:ecco-support-request@mit.edu" target="_blank">ecco-support-request@mit.edu</a><br>
<br>
You can reach the person managing the list at<br>
        <a href="mailto:ecco-support-owner@mit.edu" target="_blank">ecco-support-owner@mit.edu</a><br>
<br>
When replying, please edit your Subject line so it is more specific<br>
than &quot;Re: Contents of ecco-support digest...&quot;<br>
<br>
<br>
Today&#39;s Topics:<br>
<br>
   1. Re: Cost function calculation problems when changing number<br>
      of cores in ECCOv4-r4 (Dan Jones)<br>
   2. Re: [EXTERNAL] Re: Cost function calculation problems when<br>
      changing number of cores in ECCOv4-r4 (Wang, Ou (US 329B))<br>
   3. Re: [EXTERNAL] Re: Cost function calculation problems when<br>
      changing number of cores in ECCOv4-r4 (Patrick Heimbach)<br>
<br>
<br>
----------------------------------------------------------------------<br>
<br>
Message: 1<br>
Date: Wed, 30 Dec 2020 17:27:39 +0000<br>
From: Dan Jones &lt;<a href="mailto:dcjones.work@gmail.com" target="_blank">dcjones.work@gmail.com</a>&gt;<br>
Subject: Re: [ecco-support] Cost function calculation problems when<br>
        changing number of cores in ECCOv4-r4<br>
To: <a href="mailto:ecco-support@mit.edu" target="_blank">ecco-support@mit.edu</a><br>
Message-ID:<br>
        &lt;CAPj3iHT55=rFszYS1E0rF4gmc=<a href="mailto:B39EbiEMQSD%2BwWMe9Bf3E0OQ@mail.gmail.com" target="_blank">B39EbiEMQSD+wWMe9Bf3E0OQ@mail.gmail.com</a>&gt;<br>
Content-Type: text/plain; charset=&quot;utf-8&quot;<br>
<br>
Hi Hong,<br>
<br>
So, turning off the &quot;profiles&quot; package helped with some of the runs, but<br>
not with others.<br>
<br>
In some of the runs, I noticed that the &quot;sshv4-mdt&quot; term of the cost<br>
function is NaN. Does that term also depend on the grid? It doesn&#39;t cause<br>
problems with the 96 core version, but it does cause an issue with the 192<br>
and 360 core versions.<br>
<br>
Best,<br>
Dan<br>
-------------- next part --------------<br>
An HTML attachment was scrubbed...<br>
URL: <a href="http://mailman.mit.edu/pipermail/ecco-support/attachments/20201230/2cbaf1d6/attachment-0001.html" rel="noreferrer" target="_blank">http://mailman.mit.edu/pipermail/ecco-support/attachments/20201230/2cbaf1d6/attachment-0001.html</a><br>
<br>
------------------------------<br>
<br>
Message: 2<br>
Date: Wed, 30 Dec 2020 17:46:51 +0000<br>
From: &quot;Wang, Ou (US 329B)&quot; &lt;<a href="mailto:ou.wang@jpl.nasa.gov" target="_blank">ou.wang@jpl.nasa.gov</a>&gt;<br>
Subject: Re: [ecco-support] [EXTERNAL] Re: Cost function calculation<br>
        problems when changing number of cores in ECCOv4-r4<br>
To: &quot;ECCO support list, wider membership&quot; &lt;<a href="mailto:ecco-support@mit.edu" target="_blank">ecco-support@mit.edu</a>&gt;<br>
Message-ID: &lt;<a href="mailto:AAF31FEB-9653-4CA8-A603-EE365D4C4958@jpl.nasa.gov" target="_blank">AAF31FEB-9653-4CA8-A603-EE365D4C4958@jpl.nasa.gov</a>&gt;<br>
Content-Type: text/plain; charset=&quot;utf-8&quot;<br>
<br>
Dan,<br>
<br>
The ?ssh-mdt? term of the cost function shouldn?t depend on the grid. Could you post your STDOUT.0000 so I can take a look at it?<br>
<br>
Thanks,<br>
Ou<br>
<br>
From: &lt;<a href="mailto:ecco-support-bounces@mit.edu" target="_blank">ecco-support-bounces@mit.edu</a>&gt; on behalf of Dan Jones &lt;<a href="mailto:dcjones.work@gmail.com" target="_blank">dcjones.work@gmail.com</a>&gt;<br>
Reply-To: &quot;ECCO support list, wider membership&quot; &lt;<a href="mailto:ecco-support@mit.edu" target="_blank">ecco-support@mit.edu</a>&gt;<br>
Date: Wednesday, December 30, 2020 at 9:28 AM<br>
To: &quot;<a href="mailto:ecco-support@mit.edu" target="_blank">ecco-support@mit.edu</a>&quot; &lt;<a href="mailto:ecco-support@mit.edu" target="_blank">ecco-support@mit.edu</a>&gt;<br>
Subject: [EXTERNAL] Re: [ecco-support] Cost function calculation problems when changing number of cores in ECCOv4-r4<br>
<br>
Hi Hong,<br>
<br>
So, turning off the &quot;profiles&quot; package helped with some of the runs, but not with others.<br>
<br>
In some of the runs, I noticed that the &quot;sshv4-mdt&quot; term of the cost function is NaN. Does that term also depend on the grid? It doesn&#39;t cause problems with the 96 core version, but it does cause an issue with the 192 and 360 core versions.<br>
<br>
Best,<br>
Dan<br>
-------------- next part --------------<br>
An HTML attachment was scrubbed...<br>
URL: <a href="http://mailman.mit.edu/pipermail/ecco-support/attachments/20201230/6d13a65a/attachment-0001.html" rel="noreferrer" target="_blank">http://mailman.mit.edu/pipermail/ecco-support/attachments/20201230/6d13a65a/attachment-0001.html</a><br>
<br>
------------------------------<br>
<br>
Message: 3<br>
Date: Wed, 30 Dec 2020 17:58:18 +0000<br>
From: Patrick Heimbach &lt;<a href="mailto:heimbach@mit.edu" target="_blank">heimbach@mit.edu</a>&gt;<br>
Subject: Re: [ecco-support] [EXTERNAL] Re: Cost function calculation<br>
        problems when changing number of cores in ECCOv4-r4<br>
To: ecco-support &lt;<a href="mailto:ecco-support@mit.edu" target="_blank">ecco-support@mit.edu</a>&gt;<br>
Message-ID: &lt;<a href="mailto:9A726223-81E7-4146-898B-F56C690AF505@mit.edu" target="_blank">9A726223-81E7-4146-898B-F56C690AF505@mit.edu</a>&gt;<br>
Content-Type: text/plain; charset=&quot;utf-8&quot;<br>
<br>
You may want to check the blanklist in data.exch2 which is specific to the tiling.<br>
p.<br>
<br>
&gt; On Dec 30, 2020, at 11:46 AM, Wang, Ou (US 329B) &lt;<a href="mailto:ou.wang@jpl.nasa.gov" target="_blank">ou.wang@jpl.nasa.gov</a>&gt; wrote:<br>
&gt; <br>
&gt; Dan,<br>
&gt;  <br>
&gt; The ?ssh-mdt? term of the cost function shouldn?t depend on the grid. Could you post your STDOUT.0000 so I can take a look at it?<br>
&gt;  <br>
&gt; Thanks,<br>
&gt; Ou<br>
&gt;  <br>
&gt; From: &lt;<a href="mailto:ecco-support-bounces@mit.edu" target="_blank">ecco-support-bounces@mit.edu</a>&gt; on behalf of Dan Jones &lt;<a href="mailto:dcjones.work@gmail.com" target="_blank">dcjones.work@gmail.com</a>&gt;<br>
&gt; Reply-To: &quot;ECCO support list, wider membership&quot; &lt;<a href="mailto:ecco-support@mit.edu" target="_blank">ecco-support@mit.edu</a>&gt;<br>
&gt; Date: Wednesday, December 30, 2020 at 9:28 AM<br>
&gt; To: &quot;<a href="mailto:ecco-support@mit.edu" target="_blank">ecco-support@mit.edu</a>&quot; &lt;<a href="mailto:ecco-support@mit.edu" target="_blank">ecco-support@mit.edu</a>&gt;<br>
&gt; Subject: [EXTERNAL] Re: [ecco-support] Cost function calculation problems when changing number of cores in ECCOv4-r4<br>
&gt;  <br>
&gt; Hi Hong, <br>
&gt;  <br>
&gt; So, turning off the &quot;profiles&quot; package helped with some of the runs, but not with others. <br>
&gt; <br>
&gt; In some of the runs, I noticed that the &quot;sshv4-mdt&quot; term of the cost function is NaN. Does that term also depend on the grid? It doesn&#39;t cause problems with the 96 core version, but it does cause an issue with the 192 and 360 core versions.<br>
&gt; <br>
&gt; Best,<br>
&gt; Dan<br>
&gt; _______________________________________________<br>
&gt; ecco-support mailing list<br>
&gt; <a href="mailto:ecco-support@mit.edu" target="_blank">ecco-support@mit.edu</a><br>
&gt; <a href="http://mailman.mit.edu/mailman/listinfo/ecco-support" rel="noreferrer" target="_blank">http://mailman.mit.edu/mailman/listinfo/ecco-support</a><br>
<br>
-------------- next part --------------<br>
A non-text attachment was scrubbed...<br>
Name: smime.p7s<br>
Type: application/pkcs7-signature<br>
Size: 1865 bytes<br>
Desc: not available<br>
Url : <a href="http://mailman.mit.edu/pipermail/ecco-support/attachments/20201230/3ce86058/attachment.bin" rel="noreferrer" target="_blank">http://mailman.mit.edu/pipermail/ecco-support/attachments/20201230/3ce86058/attachment.bin</a><br>
<br>
------------------------------<br>
<br>
_______________________________________________<br>
ecco-support mailing list<br>
<a href="mailto:ecco-support@mit.edu" target="_blank">ecco-support@mit.edu</a><br>
<a href="http://mailman.mit.edu/mailman/listinfo/ecco-support" rel="noreferrer" target="_blank">http://mailman.mit.edu/mailman/listinfo/ecco-support</a><br>
<br>
<br>
End of ecco-support Digest, Vol 59, Issue 3<br>
*******************************************<br>
</blockquote></div>