<html><body>
<p><font size="2" face="sans-serif">Bronis,</font><br>
<br>
<tt><font size="2">>> >> Page 12 lines 46-48, I think MPI_THREAD_FUNNELED should have the same note <br>
>> >> about excepting helper threads calls? I think it can still be allowed, or <br>
>> >> might be desirable, to use helper threads in this case?<br>
>> <br>
>> I do not understand this comment. Which note? This confusion<br>
>> is probably an issue over which version of Pavan's document<br>
>> you used for page/line numbers. However, I think this comment<br>
>> and all remaining ones only pertain to the version with more<br>
>> significant changes (i.e., the helper threads and shared memory<br>
>> proposals). I don't intend to integrate them into the branch<br>
>> with the small changes yet so I will stop here. Please let<br>
>> me know if I have misinterpreted something.</font></tt><br>
<br>
<font size="2" face="sans-serif">I was referring to the note on MPI_THREAD_SERIALIZED that said that MPI_Helper_* calls were an exception to the rule that only one thread makes MPI calls at a time. I think (maybe) MPI_THREAD_FUNNELED should have the same/similar exception. We might have already discussed that, but I don't recall if there were compelling reasons to leave FUNNELED out.</font><br>
<br>
<br>
<font size="2" face="sans-serif">_______________________________________________<br>
Douglas Miller BlueGene Messaging Development<br>
IBM Corp., Rochester, MN USA Bldg 030-2 A410<br>
dougmill@us.ibm.com Douglas Miller/Rochester/IBM</font><br>
<br>
<img width="16" height="16" src="cid:1__=09BBFD46DFC2026F8f9e8a93df938@us.ibm.com" border="0" alt="Inactive hide details for "Bronis R. de Supinski" ---11/05/2010 06:53:54 PM---Pavan:"><font size="2" color="#424282" face="sans-serif">"Bronis R. de Supinski" ---11/05/2010 06:53:54 PM---Pavan:</font><br>
<br>
<table width="100%" border="0" cellspacing="0" cellpadding="0">
<tr valign="top"><td style="background-image:url(cid:2__=09BBFD46DFC2026F8f9e8a93df938@us.ibm.com); background-repeat: no-repeat; " width="40%">
<ul style="padding-left: 72pt"><font size="1" face="sans-serif"><b>"Bronis R. de Supinski" <bronis@llnl.gov></b></font><font size="1" face="sans-serif"> </font><br>
<font size="1" face="sans-serif">Sent by: mpi3-hybridpm-bounces@lists.mpi-forum.org</font>
<p><font size="1" face="sans-serif">11/05/2010 06:52 PM</font>
<table border="1">
<tr valign="top"><td width="168" bgcolor="#FFFFFF"><div align="center"><font size="1" face="sans-serif">Please respond to<br>
"Bronis R. de Supinski" <bronis@llnl.gov>; Please respond to<br>
mpi3-hybridpm@lists.mpi-forum.org</font></div></td></tr>
</table>
</ul>
</td><td width="60%">
<table width="100%" border="0" cellspacing="0" cellpadding="0">
<tr valign="top"><td width="1%"><img width="58" height="1" src="cid:3__=09BBFD46DFC2026F8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<div align="right"><font size="1" face="sans-serif">To</font></div></td><td width="100%"><img width="1" height="1" src="cid:3__=09BBFD46DFC2026F8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<ul style="padding-left: 7pt"><font size="1" face="sans-serif">"mpi3-hybridpm@lists.mpi-forum.org" <mpi3-hybridpm@lists.mpi-forum.org></font></ul>
</td></tr>
<tr valign="top"><td width="1%"><img width="58" height="1" src="cid:3__=09BBFD46DFC2026F8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<div align="right"><font size="1" face="sans-serif">cc</font></div></td><td width="100%"><img width="1" height="1" src="cid:3__=09BBFD46DFC2026F8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
</td></tr>
<tr valign="top"><td width="1%"><img width="58" height="1" src="cid:3__=09BBFD46DFC2026F8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<div align="right"><font size="1" face="sans-serif">Subject</font></div></td><td width="100%"><img width="1" height="1" src="cid:3__=09BBFD46DFC2026F8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<ul style="padding-left: 7pt"><font size="1" face="sans-serif">Re: [Mpi3-hybridpm] External interfaces chapter updates</font></ul>
</td></tr>
</table>
<table border="0" cellspacing="0" cellpadding="0">
<tr valign="top"><td width="58"><img width="1" height="1" src="cid:3__=09BBFD46DFC2026F8f9e8a93df938@us.ibm.com" border="0" alt=""></td><td width="336"><img width="1" height="1" src="cid:3__=09BBFD46DFC2026F8f9e8a93df938@us.ibm.com" border="0" alt=""></td></tr>
</table>
</td></tr>
</table>
<br>
<tt><font size="2"><br>
Pavan:<br>
<br>
Sorry I had to drop out of the call early today but<br>
I had to be somewhere at noon.<br>
<br>
Anyway, I am looking over Doug's comments now.<br>
<br>
Pavan and Doug:<br>
<br>
Re:<br>
> Pavan:<br>
><br>
> Doug's first issue indicates that you need to update<br>
> your PDF again. I have fixed that issue (which you<br>
> caught earlier) in the current draft in the MPI-3.0<br>
> trunk. Please start with the version that is now in<br>
> the MPI-3.0-2010-11-draft branch as I have made the<br>
> other minor corrections that we have discussed in<br>
> that section.<br>
<br>
Looks good now. Thanks.<br>
<br>
> I will look over Doug's other points to see which<br>
> should be included in that draft as well as your<br>
> proposed version once you update it.<br>
><br>
> Thanks,<br>
><br>
> Bronis<br>
><br>
><br>
> On Mon, 1 Nov 2010, Douglas Miller wrote:<br>
><br>
>> Just proof reading again...<br>
>> <br>
>> Page 11 lines 32-34, this paragraph still says "Advice to implementers" <br>
>> twice.<br>
<br>
This issue should now be fixed in all versions.<br>
<br>
>> Page 12 lines 3-4, missing close-paren at "async-signal-safe".<br>
<br>
Good catch. fixing it requires that we move the period outside<br>
of the quotation marks so it can follow the close-paren. I have<br>
made this change and commited it to MPI-3.0-2010-11-draft branch<br>
in which we are storing changes that will not go out in the<br>
first release draft since they have not been discussed/seen by<br>
the overall Forum (I know this seems like it should be OK to<br>
include in that draft but I am just following my understanding<br>
of the rules; it should get into the next release draft).<br>
<br>
>> Page 12 lines 46-48, I think MPI_THREAD_FUNNELED should have the same note <br>
>> about excepting helper threads calls? I think it can still be allowed, or <br>
>> might be desirable, to use helper threads in this case?<br>
<br>
I do not understand this comment. Which note? This confusion<br>
is probably an issue over which version of Pavan's document<br>
you used for page/line numbers. However, I think this comment<br>
and all remaining ones only pertain to the version with more<br>
significant changes (i.e., the helper threads and shared memory<br>
proposals). I don't intend to integrate them into the branch<br>
with the small changes yet so I will stop here. Please let<br>
me know if I have misinterpreted something.<br>
<br>
<br>
Pavan:<br>
<br>
Where are you keeping the version with the more significant<br>
changes? Did you have Jeff cut you a branch? We should definitely<br>
do something to keep those working changes available. It would<br>
probably be good if both of us had write access to them in<br>
case something happens and you need someone else to pick it up.<br>
<br>
Bronis<br>
<br>
<br>
<br>
<br>
>> Page 14 lines 42.5-48, the wording sounds a little soft, as if the only <br>
>> goal is to pass the linking phase. should it additionally say something <br>
>> like "must return a meaningful and accurate value"?<br>
>> <br>
>> Pages 16-17, should we add an "advice to users" to recommend/remind that <br>
>> all communications between JOIN and LEAVE be self-completing? What I mean <br>
>> is that if a thread does an MPI_ISEND between JOIN and LEAVE, that is also <br>
>> does an MPI_WAIT (or equiv.) on that ISEND before LEAVE? Since JOIN and <br>
>> LEAVE have no knowledge of requests, etc, isn't that prudent or even <br>
>> necessary?<br>
>> <br>
>> Page 17, section 12.5 Shared Memory. Rather than be collective, could these <br>
>> calls reflect the API of something like shm_open() whereby they have a <br>
>> "key" parameter that uniquely identifies the segment of shared memory? Our <br>
>> experience with DCMF (where we did all shmem allocations in a ordered, <br>
>> synchronized "collective" manner) was that it is fraught with problems and <br>
>> restrictions. We're moving to using an API that takes a string "key" so <br>
>> that we need not force such semantics. Are there any OS shmem APIs that <br>
>> require ordered, collective allocation? I know UPC does not use a "key", <br>
>> but wouldn't this allow for better implementations? Are there platforms <br>
>> where these semantics would NOT work? [probably a topic for our meeting]<br>
>> <br>
>> [also another topic for the meeting] Should we say something about how to <br>
>> get a communicator of appropriate ranks for shmem allocation? Many <br>
>> platforms do not support global shared memory (only shmem local to a node), <br>
>> and I don't think there are any MPI mechanisms for testing or selecting <br>
>> ranks that are node-local.<br>
>> <br>
>> Thanks, Pavan, for doing this integration, by the way.<br>
>> Myself, I don't know LaTeX and can't afford the learning curve right now, <br>
>> so you're really helping out.<br>
>> <br>
>> _______________________________________________<br>
>> Douglas Miller BlueGene Messaging Development<br>
>> IBM Corp., Rochester, MN USA Bldg 030-2 A410<br>
>> dougmill@us.ibm.com Douglas Miller/Rochester/IBM<br>
>> <br>
>> [</font></tt><tt><font size="2"><a href="cid:1__=09BBFD5DDFD0140C8f9e8a93df938@us.ibm.com">cid:1__=09BBFD5DDFD0140C8f9e8a93df938@us.ibm.com</a></font></tt><tt><font size="2">]Pavan Balaji <br>
>> ---10/30/2010 07:37:24 PM---On 10/30/2010 01:49 PM, Bronis R. de Supinski <br>
>> wrote:<br>
>> <br>
>> <br>
>> Pavan Balaji <balaji@mcs.anl.gov><br>
>> Sent by: mpi3-hybridpm-bounces@lists.mpi-forum.org<br>
>> <br>
>> 10/30/2010 07:36 PM<br>
>> Please respond to<br>
>> mpi3-hybridpm@lists.mpi-forum.org<br>
>> <br>
>> <br>
>> <br>
>> <br>
>> To<br>
>> <br>
>> "Bronis R. de Supinski" <bronis@llnl.gov>, <br>
>> mpi3-hybridpm@lists.mpi-forum.org<br>
>> <br>
>> cc<br>
>> <br>
>> <br>
>> <br>
>> Subject<br>
>> <br>
>> Re: [Mpi3-hybridpm] External interfaces chapter updates<br>
>> <br>
>> <br>
>> <br>
>> <br>
>> On 10/30/2010 01:49 PM, Bronis R. de Supinski wrote:<br>
>>>> I think it's time for us to start dumping in text into the chapter and<br>
>>>> start discussing the exact wording. I've included the helper threads and<br>
>>>> shared memory extensions proposals into the chapter and uploaded it to<br>
>>>> the wiki<br>
>>>> (</font></tt><tt><font size="2"><a href="https://svn.mpi-forum.org/trac/mpi-forum-web/attachment/wiki/MPI3Hybrid/ei-2-v0.1.pdf">https://svn.mpi-forum.org/trac/mpi-forum-web/attachment/wiki/MPI3Hybrid/ei-2-v0.1.pdf</a></font></tt><tt><font size="2">).<br>
>>>> Please take a look at it and let me know your comments.<br>
>>> <br>
>>> I'll try to make a detailed reading next week. In the<br>
>> <br>
>> I had initially incorrectly uploaded a change with respect to the<br>
>> MPI_THREAD_SERIALIZED semantics which we decided to drop last time. I've<br>
>> now uploaded v0.2 of the document with this correction as well as the<br>
>> other changes suggested by Bronis.<br>
>> <br>
>> -- Pavan<br>
>> <br>
>> --<br>
>> Pavan Balaji<br>
>> </font></tt><tt><font size="2"><a href="http://www.mcs.anl.gov/~balaji">http://www.mcs.anl.gov/~balaji</a></font></tt><tt><font size="2"><br>
>> _______________________________________________<br>
>> Mpi3-hybridpm mailing list<br>
>> Mpi3-hybridpm@lists.mpi-forum.org<br>
>> </font></tt><tt><font size="2"><a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-hybridpm">http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-hybridpm</a></font></tt><tt><font size="2"><br>
>> <br>
>> <br>
><br>
_______________________________________________<br>
Mpi3-hybridpm mailing list<br>
Mpi3-hybridpm@lists.mpi-forum.org<br>
</font></tt><tt><font size="2"><a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-hybridpm">http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-hybridpm</a></font></tt><tt><font size="2"><br>
</font></tt><br>
</body></html>