<html dir="ltr">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<style type="text/css" id="owaParaStyle"></style><style type="text/css"></style><style type="text/css"></style>
</head>
<body dir="auto" fpstyle="1" ocsi="0">
<div style="direction: ltr;font-family: Tahoma;color: #000000;font-size: 10pt;">The idea behind dynamic windows was to make them possible to implement today and push the vendors to support a more rational programming model tomorrow. Like Jim, I'm not sure
things are bleak enough that we should change.
<div><br>
</div>
<div>In OFED, I think it would be very difficult today. But with Mellanox's work to support dynamic registration cache in hardware, I think things are trending in the right direction there.</div>
<div><br>
</div>
<div>For Portals 4, if an implementation supports the BIND_INACCESSIBLE feature, and I think most implementations will, supporting dynamic windows is straight-forward.</div>
<div><br>
</div>
<div>PAMI and DMAPP are probably more difficult, but a firehose-like setup should be possible until hardware catches up.</div>
<div><br>
</div>
<div>Brian<br>
<div><br>
<div><font size="2" face="Tahoma">-- </font></div>
<div><font size="2" face="tahoma"> Brian W. Barrett</font></div>
<div><font size="2" face="tahoma"> Scalable System Software Group</font></div>
<div><font size="2" face="tahoma"> Sandia National Laboratories</font></div>
</div>
<div style="font-family: Times New Roman; color: #000000; font-size: 16px">
<hr tabindex="-1">
<div id="divRpF513773" style="direction: ltr;"><font face="Tahoma" size="2" color="#000000"><b>From:</b> mpiwg-rma [mpiwg-rma-bounces@lists.mpi-forum.org] on behalf of Jeff Hammond [jeff.science@gmail.com]<br>
<b>Sent:</b> Sunday, December 01, 2013 10:16 AM<br>
<b>To:</b> MPI WG Remote Memory Access working group<br>
<b>Cc:</b> MPI WG Remote Memory Access working group<br>
<b>Subject:</b> [EXTERNAL] Re: [mpiwg-rma] making dynamic windows suck less<br>
</font><br>
</div>
<div></div>
<div>
<div>How do you do pure (i.e. active-message and rendezvous-free) RDMA with just a virtual address (MPI_Aint) in (1) PAMI, (2) OFED, (3) DMAPP, (4) Portals4?</div>
<div><br>
</div>
<div>Jeff<br>
<br>
Sent from my iPhone</div>
<div><br>
On Dec 1, 2013, at 10:40 AM, Jim Dinan <<a href="mailto:james.dinan@gmail.com" target="_blank">james.dinan@gmail.com</a>> wrote:<br>
<br>
</div>
<blockquote type="cite">
<div>
<div dir="ltr">Can the MPI_Win_memobj be an MPI_Aint?
<div><br>
</div>
<div>I don't share this pessimism with respect to dynamic windows; I think we need to prove that these issues exist before we try to fix them.
<div>
<div>
<div>
<div><br>
</div>
<div> ~Jim.</div>
</div>
</div>
</div>
</div>
</div>
<div class="gmail_extra"><br>
<br>
<div class="gmail_quote">On Wed, Nov 27, 2013 at 5:46 PM, Jeff Hammond <span dir="ltr">
<<a href="mailto:jeff.science@gmail.com" target="_blank">jeff.science@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex; border-left:1px #ccc solid; padding-left:1ex">
Lots of people think - I take position on their correctness - that<br>
dynamic windows are doomed to substandard performance. I can<br>
certainly see some potential for this on networks that require memory<br>
registration.<br>
<br>
Can we try to add an opaque object that can encapsulate memory<br>
registration such that dynamic windows 2.0 would not have so many<br>
problems? If one were to send-recv not just the virtual address but<br>
MPI_Win_memobj (to be named properly later), then maybe we could deal<br>
with the shortcomings of IB, etc. w.r.t. RDMA and memory registration.<br>
<br>
This is really just a thought. I have no concrete proposal. Maybe<br>
people will have ideas in December.<br>
<span class="HOEnZb"><font color="#888888"><br>
Jeff<br>
<br>
--<br>
Jeff Hammond<br>
<a href="mailto:jeff.science@gmail.com" target="_blank">jeff.science@gmail.com</a><br>
_______________________________________________<br>
mpiwg-rma mailing list<br>
<a href="mailto:mpiwg-rma@lists.mpi-forum.org" target="_blank">mpiwg-rma@lists.mpi-forum.org</a><br>
<a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma" target="_blank">http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma</a><br>
</font></span></blockquote>
</div>
<br>
</div>
</div>
</blockquote>
<blockquote type="cite">
<div><span>_______________________________________________</span><br>
<span>mpiwg-rma mailing list</span><br>
<span><a href="mailto:mpiwg-rma@lists.mpi-forum.org" target="_blank">mpiwg-rma@lists.mpi-forum.org</a></span><br>
<span><a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma" target="_blank">http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma</a></span></div>
</blockquote>
</div>
</div>
</div>
</div>
</body>
</html>