<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">
Hi All,</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">
Here is the updated <a href="https://drive.google.com/file/d/1VkkqAd56lgxEt3JEt-zmGR0aC1VvLTDh/view?usp=sharing" title="https://drive.google.com/file/d/1VkkqAd56lgxEt3JEt-zmGR0aC1VvLTDh/view?usp=sharing">
spreadsheet </a>for the appendix and the updated <a href="https://docs.google.com/document/d/12M6glWfNDt3WpLsf8FNug2ma0cRG2QZ_mNKnvUl19Ws/edit?usp=sharing" title="https://docs.google.com/document/d/12M6glWfNDt3WpLsf8FNug2ma0cRG2QZ_mNKnvUl19Ws/edit?usp=sharing">
text</a> for the semantic terms section based on the discussion with Ryan related to partitioned communication functions this week. Please review.
<br>
</div>
<div><br>
</div>
<div>The basic idea is to introduce a new operation type called partitioned operation. This would require the introduction of a new stage called ready stage after the starting stage. Ryan has updated the text for the ready stage and we have added a new state
transition diagram for the partitioned operations. <br>
</div>
<div><br>
</div>
<div>Thanks,</div>
<div>Puri<br>
</div>
<div>
<div id="appendonsend"></div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<hr tabindex="-1" style="display:inline-block; width:98%">
<div id="divRplyFwdMsg" dir="ltr"><font style="font-size: 11pt;" data-ogsc="" face="Calibri, sans-serif" color="#000000"><b>From:</b> mpiwg-semantic-terms <mpiwg-semantic-terms-bounces@lists.mpi-forum.org> on behalf of Bangalore, Purushotham via mpiwg-semantic-terms
<mpiwg-semantic-terms@lists.mpi-forum.org><br>
<b>Sent:</b> Tuesday, July 21, 2020 4:01 PM<br>
<b>To:</b> Rolf Rabenseifner <rabenseifner@hlrs.de><br>
<b>Cc:</b> mpiwg-semantic-terms <mpiwg-semantic-terms@lists.mpi-forum.org>; mpiwg-persistence@lists.mpi-forum.org <mpiwg-persistence@lists.mpi-forum.org><br>
<b>Subject:</b> Re: [mpiwg-semantic-terms] [mpiwg-persistence] Summary of Semantics for Partitioned Communication Operations</font>
<div> </div>
</div>
<div dir="ltr">
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">
Hi Rolf,</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">
I used the text as a placeholder. This text currently appears as an advice to implementors in the partitioned communication chapter.
<br>
</div>
<br>
<div>
<div id="x_appendonsend"></div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
I will let the partitioned communication WG decide the appropriate text for the footnote
<span id="x_π">π</span></div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<span id="x_π"><br>
</span></div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<span id="x_π">Thanks,</span></div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<span id="x_π">Puri</span><br>
</div>
<hr tabindex="-1" style="display:inline-block; width:98%">
<div id="x_divRplyFwdMsg" dir="ltr"><font style="font-size: 11pt;" data-ogsc="" face="Calibri, sans-serif" color="#000000"><b>From:</b> Rolf Rabenseifner <rabenseifner@hlrs.de><br>
<b>Sent:</b> Tuesday, July 21, 2020 3:43 PM<br>
<b>To:</b> Bangalore, Purushotham <puri@uab.edu><br>
<b>Cc:</b> mpiwg-persistence@lists.mpi-forum.org <mpiwg-persistence@lists.mpi-forum.org>; mpiwg-semantic-terms <mpiwg-semantic-terms@lists.mpi-forum.org><br>
<b>Subject:</b> Re: [mpiwg-semantic-terms] [mpiwg-persistence] Summary of Semantics for Partitioned Communication Operations</font>
<div> </div>
</div>
<div class="x_BodyFragment"><font size="2"><span style="font-size:11pt">
<div class="x_PlainText">Dear puri and all,<br>
<br>
I expect that the proposed footnote<br>
<br>
"20) Unlike MPI_SEND_INIT and MPI_RECV_INIT, MPI_PSEND_INIT and MPI_PRECV_INIT may communicate."<br>
<br>
is incorrect.<br>
<br>
MPI_SEND_INIT and MPI_RECV_INIT, MPI_PSEND_INIT and MPI_PRECV_INIT are all local <br>
procedures and we know, that local is weak local and no line in the MPI<br>
Standard can prohibit that a (weak) local is performing some "internal"<br>
communication for optimization.<br>
As long as the user's (= application programmer's) viewpoint can see<br>
only a local routine, and the freeing is also a local routine, <br>
even in the case that there was never an MPI_START, then all is fine.<br>
<br>
Maybe you meant:<br>
20) Although these are local routines, an MPI implementation may internally<br>
already start some communication of metadata, but must guarantee that the INIT and the
<br>
MPI_REQUEST_FREE are both local procedures. <br>
<br>
But such a comment should be an advice to implementors in the <br>
definition of the MPI_P..._INIT routines, and not as a footnote here.<br>
<br>
My comment is completely independent of what is written in the MPI Standard elsewhere.<br>
<br>
MPI-3.1 p 73 line 41 on persistent pt-to-pt init:<br>
"These calls involve no communication."<br>
<br>
In principle, this sentence should be substituted by <br>
"These procedures are local and do not send or receive any buffer content."<br>
<br>
But this may be a cleanup for MPI 4.1 ;-)<br>
<br>
Such a sentence should be part of the new partitioned comm. chapter!<br>
<br>
Best regards<br>
Rolf <br>
<br>
<br>
<br>
<br>
----- Original Message -----<br>
> From: "mpiwg-semantic-terms" <mpiwg-semantic-terms@lists.mpi-forum.org><br>
> To: mpiwg-persistence@lists.mpi-forum.org<br>
> Cc: "mpiwg-semantic-terms" <mpiwg-semantic-terms@lists.mpi-forum.org><br>
> Sent: Tuesday, July 21, 2020 8:34:09 PM<br>
> Subject: Re: [mpiwg-semantic-terms] [mpiwg-persistence] Summary of Semantics for Partitioned Communication Operations<br>
<br>
> Here is the updated spreadsheet based on Dan's suggestions. I would really like<br>
> the persistence WG to finalize this ASAP.<br>
> <br>
> I know Ryan would like to review the changes to the chapter before it is merged<br>
> during tomorrow's meeting. If there is time left after that, I would like to<br>
> discuss this.<br>
> <br>
> Thanks,<br>
> Puri<br>
> <br>
> <br>
> From: mpiwg-persistence <mpiwg-persistence-bounces@lists.mpi-forum.org> on<br>
> behalf of Bangalore, Purushotham via mpiwg-persistence<br>
> <mpiwg-persistence@lists.mpi-forum.org><br>
> Sent: Wednesday, July 15, 2020 12:54 PM<br>
> To: HOLMES Daniel <d.holmes@epcc.ed.ac.uk><br>
> Cc: mpiwg-semantic-terms@lists.mpi-forum.org<br>
> <mpiwg-semantic-terms@lists.mpi-forum.org>;<br>
> mpiwg-persistence@lists.mpi-forum.org <mpiwg-persistence@lists.mpi-forum.org><br>
> Subject: Re: [mpiwg-persistence] Summary of Semantics for Partitioned<br>
> Communication Operations<br>
> Here are the other footnotes (I should have pasted them near footnote 20):<br>
> <br>
> 7) Addresses are cached on the request handle.<br>
> <br>
> 9) One shall not free or deallocate the bu er before the operation is freed,<br>
> that is MPI_REQUEST_FREE returned.<br>
> <br>
> 14) Nonblocking procedure without an I pre fix.<br>
> <br>
> =====<br>
> I will move PREADY,...,PARRIVED after START... good suggestion.<br>
> <br>
> I also agree with "ic" for the PREADY, PARRIVED procedures.<br>
> <br>
> <br>
> From: HOLMES Daniel <d.holmes@epcc.ed.ac.uk><br>
> Sent: Wednesday, July 15, 2020 12:42 PM<br>
> To: Bangalore, Purushotham <puri@uab.edu><br>
> Cc: mpiwg-persistence@lists.mpi-forum.org<br>
> <mpiwg-persistence@lists.mpi-forum.org>;<br>
> mpiwg-semantic-terms@lists.mpi-forum.org<br>
> <mpiwg-semantic-terms@lists.mpi-forum.org><br>
> Subject: Re: [mpiwg-persistence] Summary of Semantics for Partitioned<br>
> Communication Operations<br>
> Hi Puri,<br>
> <br>
> Thanks for this - a great way to provoke some progress (no pun intended).<br>
> <br>
> I would suggest: move the rows dealing with MPI_PREADY and MPI_PARRIVED between<br>
> MPI_START and MPI_WAIT, i.e. in the approximate calling order.<br>
> <br>
> Also, I would add βicβ for the new MPI_PREADY, MPI_PARRIVED, etc procedures -<br>
> because the operation is definitely incomplete at the point in time they are<br>
> called, during the time interval of their entire execution, and at the point in<br>
> time they return - there is no possibility they could be completing or freeing<br>
> procedures.<br>
> <br>
> What are the other footnotes: 7, 9, and 14? (I could look them up but Iβd have<br>
> to choose the same draft version as you because the numbering changed<br>
> recently.)<br>
> <br>
> Cheers,<br>
> Dan.<br>
> β<br>
> Dr Daniel Holmes PhD<br>
> Architect (HPC Research)<br>
> [ <a href="mailto:d.holmes@epcc.ed.ac.uk">mailto:d.holmes@epcc.ed.ac.uk</a> | d.holmes@epcc.ed.ac.uk ]<br>
> Phone: +44 (0) 131 651 3465<br>
> Mobile: +44 (0) 7940 524 088<br>
> Address: Room 2.09, Bayes Centre, 47 Potterrow, Central Area, Edinburgh, EH8 9BT<br>
> β<br>
> The University of Edinburgh is a charitable body, registered in Scotland, with<br>
> registration number SC005336.<br>
> β<br>
> <br>
> <br>
> <br>
> <br>
> On 15 Jul 2020, at 18:24, Bangalore, Purushotham via mpiwg-persistence < [<br>
> <a href="mailto:mpiwg-persistence@lists.mpi-forum.org">mailto:mpiwg-persistence@lists.mpi-forum.org</a> |<br>
> mpiwg-persistence@lists.mpi-forum.org ] > wrote:<br>
> <br>
> <MPI-semantics-appendix.xlsx><br>
> <br>
> <br>
> --<br>
> mpiwg-semantic-terms mailing list<br>
> mpiwg-semantic-terms@lists.mpi-forum.org<br>
> <a href="https://lists.mpi-forum.org/mailman/listinfo/mpiwg-semantic-terms">https://lists.mpi-forum.org/mailman/listinfo/mpiwg-semantic-terms</a><br>
<br>
-- <br>
Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner@hlrs.de .<br>
High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 .<br>
University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 .<br>
Head of Dpmt Parallel Computing . . . <a href="http://www.hlrs.de/people/rabenseifner">
www.hlrs.de/people/rabenseifner</a> .<br>
Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room 1.307) .<br>
</div>
</span></font></div>
</div>
</div>
</div>
</body>
</html>