<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">There was talk of having Erik present
something on benchmarks on one of these calls. I would like to
request that we *not* do it on Nov 3, as neither Frank nor I can
be there.<br>
<br>
Cheers,<br>
Steve<br>
<br>
On 10/28/2013 06:40 PM, Roland Haas wrote:<br>
</div>
<blockquote cite="mid:526EF5D2.4030609@physics.gatech.edu"
type="cite">
<pre wrap="">Present: Erik, Frank, Roland, Ian, Steve
Release date:
* run testsuite by week of Nov 11th
* aim for release before Supercomputing conference
* Erik will start the tests on all machines, will then assign failing
tests to maintainers
* ask if Bruno Mundim would want to run on their machines in Frankfurt
Restriction in Carpet:
* commit patch to disable restriction into outer boundaries (which can
never be fully filled via restriction) after release
* add TODO item to optimize communication algorithm
* deferred tickets
** #590 deferred to next release
** #1425 not release critical, remove milestone
** #980 deferred until after release
** #1429 patch applied, asked Bernard for exact intel compiler
identification to add #ifdef to TP
** 1317 deferred until after release
** 809 deferred until after release
** 1251 deferred until after release, assign to Steve
* those with patches will be added to trunk very shortly after the release
* tickets for release
** Frank: apply conservative patch and disable for all intel 13
compilers starting from version known to fail. Backport more restrictive
exclusion once we know of a known good version
** Roland: check if vectorization patch in PUGH needs to depend on
Vectors (# 1411)
** Roland: apply first patch in #1395
** Roland: remove OPTIONAL_IFACTIVE from #1411
** Philipp: #1409 will include but add comment that Avec does not
currently work
** Erik: #1360 (check if they even build, defer until after release)
** Ian: #626, #1075, #1387: delete faulty parfiles
** #1364 closed, suggest to require simfactories "make" option to be an
executable with options. Needs to be documented.
Running tests with different numbers of processors:
* see comment in #1075 for decision on how to proceed if a test case
asks for an explicit number of processes
Yours,
Roland
</pre>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
Users mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Users@einsteintoolkit.org">Users@einsteintoolkit.org</a>
<a class="moz-txt-link-freetext" href="http://lists.einsteintoolkit.org/mailman/listinfo/users">http://lists.einsteintoolkit.org/mailman/listinfo/users</a>
</pre>
</blockquote>
<br>
</body>
</html>