From jsquyres at [hidden]  Fri Feb  5 10:09:03 2010
From: jsquyres at [hidden] (Jeff Squyres)
Date: Fri, 5 Feb 2010 11:09:03 -0500
Subject: [Mpi3-bwcompat] ABI compat list from Intel?
Message-ID: <4F4BC2C7-9FCB-4133-91D6-7CC518728C88@cisco.com>



Alexander --

I seem to recall you mentioning that you had a list of 10-12 items that affect ABI from the MPI-2.x series and/or MPI-3 proposals.  Did you ever share that list with the group?


-- 
Jeff Squyres <jsquyres_at_[hidden]>
Cisco.com - http://www.cisco.com
For corporate legal information go to:
http://www.cisco.com/web/about/doing_business/legal/cri/



From jsquyres at [hidden]  Fri Feb  5 12:44:33 2010
From: jsquyres at [hidden] (Jeff Squyres)
Date: Fri, 5 Feb 2010 13:44:33 -0500
Subject: [Mpi3-bwcompat] Notes from ATL Forum meeting
In-Reply-To: <79404EC9-50F2-46CD-B384-1397E57CC597@cisco.com>
Message-ID: <8A883C17-A507-41A3-9828-2E88BD2C2488@cisco.com>



On Jan 22, 2010, at 1:50 PM, Jeff Squyres (jsquyres) wrote:

>  * What to do about the proposed MPI-3 explicit Fortran bindings (the
>    so-called "F03 bindings")?  We could use function overloading and
>    forego the suffix approach that is necessary in C.  But this was
>    deemed to be a bad idea -- it seemed better to keep function name
>    parity with C.  So have both an MPI_File_write with int count and
>    MPI_File_writeL with MPI_Count count in the "F03" bindings.

One more point here: there should be no "integer" version of MPI_FILE_WRITE in the F03 bindings because they're new.  This follows the same precedent of "no int versions of non-blocking collectives".  There should only be MPI_FILE_WRITE<suffix> in the F03 bindings that takes an argument of the MPI_COUNT variety (exact type tbd; probably a KIND sort of thing).


-- 
Jeff Squyres <jsquyres_at_[hidden]>
Cisco.com - http://www.cisco.com
For corporate legal information go to:
http://www.cisco.com/web/about/doing_business/legal/cri/



From david.solt at [hidden]  Tue Feb  9 11:31:25 2010
From: david.solt at [hidden] (Solt, David George)
Date: Tue, 9 Feb 2010 17:31:25 +0000
Subject: [Mpi3-bwcompat] Notes from last meeting + draft of standard text
In-Reply-To: <F14CDB5DAAD62742A2D7F497F9DA7E053217734784@GVW0547EXC.americas.hpqcorp.net>
Message-ID: <F14CDB5DAAD62742A2D7F497F9DA7E05321961FA91@GVW0547EXC.americas.hpqcorp.net>



I added a few notes to the meeting note page based on our meeting last week.   I also put forth a rough draft of our explanation of versioning and an introduction to the MPI_Count concept.  

https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/DraftMpiCountText

I plan to add a presentation of the associated MPI datatype for MPI_Count later this week.   Feel free to update, modify, criticize, praise, etc, etc.

I also added a placeholder on the main page for our recommendations to authors, but have not created the link yet.  If anyone wants to beat me to it, feel free to have at it.  Otherwise I will take a stab at it later in the week.

Thanks,
Dave


From thakur at [hidden]  Tue Feb  9 12:47:39 2010
From: thakur at [hidden] (Rajeev Thakur)
Date: Tue, 9 Feb 2010 12:47:39 -0600
Subject: [Mpi3-bwcompat] Notes from last meeting + draft of standard text
In-Reply-To: <F14CDB5DAAD62742A2D7F497F9DA7E05321961FA91@GVW0547EXC.americas.hpqcorp.net>
Message-ID: <593DD7ECE63C4BD6BCE104FA104785CC@thakurlaptop>



So MPI_Count is not required to 64-bit?

Rjaeev 

> -----Original Message-----
> From: mpi3-bwcompat-bounces_at_[hidden] 
> [mailto:mpi3-bwcompat-bounces_at_[hidden]] On Behalf Of 
> Solt, David George
> Sent: Tuesday, February 09, 2010 11:31 AM
> To: MPI-3 backwards compatability WG
> Subject: [Mpi3-bwcompat] Notes from last meeting + draft of 
> standard text
> 
> I added a few notes to the meeting note page based on our 
> meeting last week.   I also put forth a rough draft of our 
> explanation of versioning and an introduction to the 
> MPI_Count concept.  
> 
> https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/DraftMpiCountText
> 
> I plan to add a presentation of the associated MPI datatype 
> for MPI_Count later this week.   Feel free to update, modify, 
> criticize, praise, etc, etc.
> 
> I also added a placeholder on the main page for our 
> recommendations to authors, but have not created the link 
> yet.  If anyone wants to beat me to it, feel free to have at 
> it.  Otherwise I will take a stab at it later in the week.
> 
> Thanks,
> Dave
> 
> 
> _______________________________________________
> Mpi3-bwcompat mailing list
> Mpi3-bwcompat_at_[hidden]
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-bwcompat
> 


From david.solt at [hidden]  Tue Feb  9 13:20:11 2010
From: david.solt at [hidden] (Solt, David George)
Date: Tue, 9 Feb 2010 19:20:11 +0000
Subject: [Mpi3-bwcompat] Notes from last meeting + draft of standard text
In-Reply-To: <593DD7ECE63C4BD6BCE104FA104785CC@thakurlaptop>
Message-ID: <F14CDB5DAAD62742A2D7F497F9DA7E05321961FBDB@GVW0547EXC.americas.hpqcorp.net>



I believe that is a mistake on my part.   Yes, for this proposal I believe the intent is to require it to be 64-bit.  I will update.

Dave

-----Original Message-----
From: mpi3-bwcompat-bounces_at_[hidden] [mailto:mpi3-bwcompat-bounces_at_[hidden]] On Behalf Of Rajeev Thakur
Sent: Tuesday, February 09, 2010 12:48 PM
To: 'MPI-3 backwards compatability WG'
Subject: Re: [Mpi3-bwcompat] Notes from last meeting + draft of standard text

So MPI_Count is not required to 64-bit?

Rjaeev 

> -----Original Message-----
> From: mpi3-bwcompat-bounces_at_[hidden] 
> [mailto:mpi3-bwcompat-bounces_at_[hidden]] On Behalf Of 
> Solt, David George
> Sent: Tuesday, February 09, 2010 11:31 AM
> To: MPI-3 backwards compatability WG
> Subject: [Mpi3-bwcompat] Notes from last meeting + draft of 
> standard text
> 
> I added a few notes to the meeting note page based on our 
> meeting last week.   I also put forth a rough draft of our 
> explanation of versioning and an introduction to the 
> MPI_Count concept.  
> 
> https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/DraftMpiCountText
> 
> I plan to add a presentation of the associated MPI datatype 
> for MPI_Count later this week.   Feel free to update, modify, 
> criticize, praise, etc, etc.
> 
> I also added a placeholder on the main page for our 
> recommendations to authors, but have not created the link 
> yet.  If anyone wants to beat me to it, feel free to have at 
> it.  Otherwise I will take a stab at it later in the week.
> 
> Thanks,
> Dave
> 
> 
> _______________________________________________
> Mpi3-bwcompat mailing list
> Mpi3-bwcompat_at_[hidden]
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-bwcompat
> 

_______________________________________________
Mpi3-bwcompat mailing list
Mpi3-bwcompat_at_[hidden]
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-bwcompat