<html>
<body>
Do you have time to complete the Microsoft column?<br><br>
At 12:14 PM 4/24/2008, Erez Haba wrote:<br>
<blockquote type=cite class=cite cite="">Okay with me.<br>
<br>
<b>From:</b> mpi3-abi-bounces@lists.mpi-forum.org
[<a href="mailto:mpi3-abi-bounces@lists.mpi-forum.org" eudora="autourl">
mailto:mpi3-abi-bounces@lists.mpi-forum.org</a>] <b>On Behalf Of </b>Jeff
Brown<br>
<b>Sent:</b> Thursday, April 24, 2008 10:53 AM<br>
<b>To:</b> MPI 3.0 ABI working group; MPI 3.0 ABI working group<br>
<b>Subject:</b> Re: [Mpi3-abi] For the April MPI Forum Meeting<br>
<br>
here's what I see on the wiki:<br>
Types
HP
IBM
Microsoft MPICH2
OpenMPI
ABI <br>
MPI_Datatype struct hpmp_dtype_s*
int
int
int struct
ompi_datatype_t* TBD
<br>
MPI_Op struct
hpmp_op_s*
int
int
int struct
ompi_op_t*
TBD <br>
MPI_Comm
struct hpmp_comm_s* int
int
int struct
ompi_communicator_t* TBD <br>
MPI_Errhandler struct hpmp_err_s*
int
int
int struct
ompi_errhandler_t* TBD <br>
Examples
HP
IBM
Microsoft
MPICH2
OpenMPI
ABI <br>
Datatype MPI_CHAR
&hpmp_char
4
0x4c000101
0x4c000101
&ompi_mpi_char
TBD <br>
Op MPI_SUM
&hpmp_sum
enum 2 0x58000003
0x58000003
&ompi_mpi_op_sum TBD <br>
MPI_COMM_WORLD
&hpmp_comm_world enum
0 0x44000000
0x44000000
&ompi_mpi_comm_world TBD <br>
Compare MPI_IDENT
0
enum 0
0
0
enum
0
TBD<br>
There's a lot more detail in the spreadsheet. To do this right, we
need to cover the entire space. I'd prefer to stick with excel
(posted to the wiki) and add columns for the various
implementations.<br><br>
At 10:23 AM 4/24/2008, Erez Haba wrote:<br><br>
Jeff, are you aware that we started that table on the wiki pages? (or you
just prefer it in excel?)<br>
<a href="http://svn.mpi-forum.org/trac/mpi-forum-web/wiki/Compare_mpi_h">
http://svn.mpi-forum.org/trac/mpi-forum-web/wiki/Compare_mpi_h</a><br><br>
On that same page you can also find the various mpi.h files.<br><br>
<br>
Thanks,<br>
.Erez<br><br>
-----Original Message-----<br>
From: mpi3-abi-bounces@lists.mpi-forum.org
[<a href="mailto:mpi3-abi-bounces@lists.mpi-forum.org" eudora="autourl">
</a>
<a href="mailto:mpi3-abi-bounces@lists.mpi-forum.org" eudora="autourl">
mailto:mpi3-abi-bounces@lists.mpi-forum.org</a>] On Behalf Of Jeff
Brown<br>
Sent: Thursday, April 24, 2008 9:18 AM<br>
To: MPI 3.0 ABI working group; MPI 3.0 ABI working group<br>
Subject: Re: [Mpi3-abi] For the April MPI Forum Meeting<br><br>
Just talked to our phone folks. Our trunks are down in New Mexico
-<br>
so this is a bust. We are sort of a third world country out
here.<br><br>
I don't think we have time to reschedule at this point.<br><br>
So ... if folks have the time please populate the matrix with your<br>
favorite MPI implementation and distribute to the group. We'll
get<br>
into the guts of all this at the meeting.<br><br>
For my 5 minute briefing, I'll just show folks where we are and give<br>
a glimpse into the details.<br><br>
see you all at the meeting<br><br>
Jeff<br><br>
At 10:01 AM 4/24/2008, Terry Dontje wrote:<br>
>Am I the only one getting an "all circuits are busy"
message from the<br>
>number below?<br>
><br>
>--td<br>
><br>
>Jeff Brown wrote:<br>
> > all,<br>
> ><br>
> > I scheduled a telecon to discuss status and get somewhat
organized<br>
> > for the meeting:<br>
> ><br>
> > Thursday April 24, 10:00 MDT<br>
> > local number:
606-1201(6-1201)<br>
> > toll free number: 888 343-0702.<br>
> ><br>
> > I'll send out some slides for the 5 minute briefing for the
group.<br>
> ><br>
> > I'm having a hard time finding time to devote to this, but I'll
have<br>
> > a cut at the OpenMPI and LAMPI analysis prior to the
telecon. We<br>
> > need someone to look at MPICH, and the vendor implementations
need to<br>
> > be posted.<br>
> ><br>
> > Jeff<br>
> ><br>
> ><br>
> ><br>
> > At 10:03 AM 4/16/2008, Jeff Brown wrote:<br>
> ><br>
> >> Yes, it's time to put some cycles toward this. Let's
start<br>
> >> populating the matrix and have a telecon toward the end of
next<br>
> >> week. I'll schedule a WG working session at the
meeting.<br>
> >><br>
> >> I'll take a look at OpenMPI and LAMPI, the two primary
MPI<br>
> >> implementations we use at LANL, and post to the wiki by the
end of<br>
> >> the week. Others, please do the same for your MPI
implementation<br>
> >> (especially the vendors). Overlap is OK.<br>
> >><br>
> >> I'll send out specifics on the telecon. Let's shoot
for Thursday<br>
> >> April 24, 9:00 A.M. MST.<br>
> >><br>
> >> Jeff<br>
> >><br>
> >> At 09:51 AM 4/16/2008, Narasimhan, Kannan wrote:<br>
> >><br>
> >>> Folks,<br>
> >>><br>
> >>> Are we planning on a WG update to report at the April
28-30 Forum<br>
> >>> meeting? We have started the process of identifying the
mpi.h<br>
> >>> differences, but I dont think we have synthesized the
data yet, or<br>
> >>> come to any conclusions/next steps... Or did I miss
something here?<br>
> >>><br>
> >>> Thanx!<br>
> >>> Kannan<br>
> >>><br>
> >>> -----Original Message-----<br>
> >>> From: mpi3-abi-bounces@lists.mpi-forum.org<br>
> >>>
[<a href="mailto:mpi3-abi-bounces@lists.mpi-forum.org" eudora="autourl">
</a>
<a href="mailto:mpi3-abi-bounces@lists.mpi-forum.org" eudora="autourl">
mailto:mpi3-abi-bounces@lists.mpi-forum.org</a>] On Behalf Of Edric
Ellis<br>
> >>> Sent: Monday, March 17, 2008 4:18 AM<br>
> >>> To: MPI 3.0 ABI working group<br>
> >>> Subject: Re: [Mpi3-abi] Meeting notes from 10th
March<br>
> >>><br>
> >>><br>
> >>> I'm not sure how best to express this, but there are a
couple of<br>
> >>> things that occur to me that might be important:<br>
> >>><br>
> >>> 1. The size of the handle types (cf. size of a pointer
perhaps?)<br>
> >>><br>
> >>> 2. should we add some sort of table describing the
current situation<br>
> >>> as to how applications pick up the value of e.g.
MPI_COMM_WORLD? E.g.<br>
> >>> MPICH2 uses "#define MPI_COMM_WORLD
0x44000000", so that value is<br>
> >>> burned into the binary; whereas OpenMPI uses extern
pointers - i.e.<br>
> >>> ompi_mpi_comm_world is in the initialized data section
of libmpi.so,<br>
> >>> and the value resolved at (dynamic) link time.<br>
> >>><br>
> >>> Cheers,<br>
> >>><br>
> >>> Edric.<br>
> >>><br>
> >>><br>
> >>>> -----Original Message-----<br>
> >>>> From: mpi3-abi-bounces@lists.mpi-forum.org<br>
> >>>><br>
> >>> [<<a href="mailto:mpi3-abi-">mailto:mpi3-abi-</a>
><a href="mailto:mpi3-abi" eudora="autourl">mailto:mpi3-abi</a>-<br>
> >>><br>
> >>>> bounces@lists.mpi-forum.org] On Behalf Of Jeff
Brown<br>
> >>>> Sent: Thursday, March 13, 2008 10:11 PM<br>
> >>>> To: MPI 3.0 ABI working group;
mpi3-abi@lists.mpi-forum.org<br>
> >>>> Subject: Re: [Mpi3-abi] Meeting notes from 10th
March<br>
> >>>><br>
> >>>> I propose a way we can make progress ...<br>
> >>>><br>
> >>>> Let's start populating a matrix (excel spreadsheet)
with a column for<br>
> >>>> each MPI implementation, and rows for the various
MPI datatypes,<br>
> >>>> constants, etc. where the internal implementations
varys. I'll kick<br>
> >>>> it off for OpenMPI and send out.<br>
> >>>><br>
> >>>> The last column of the matrix can be
"ABI" where we propose a common<br>
> >>>> approach across the implementations.<br>
> >>>><br>
> >>>> A couple of driving principles:<br>
> >>>> 1. the ABI solution shouldn't negatively impact
quality of<br>
> >>>><br>
> >>> implementation<br>
> >>><br>
> >>>> 2. minimize platform specific solutions<br>
> >>>><br>
> >>>> I'd like to see if we can produce a single ABI that
spans platforms.<br>
> >>>><br>
> >>>> comments?<br>
> >>>><br>
> >>>> Jeff<br>
> >>>><br>
> >>>><br>
> >>>>
_______________________________________________<br>
> >>>> mpi3-abi mailing list<br>
> >>>> mpi3-abi@lists.mpi-forum.org<br>
> >>>>
<a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi">
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi</a><br>
> >>>><br>
> >>> _______________________________________________<br>
> >>> mpi3-abi mailing list<br>
> >>> mpi3-abi@lists.mpi-forum.org<br>
> >>>
<a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi">
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi</a><br>
> >>><br>
> >>> _______________________________________________<br>
> >>> mpi3-abi mailing list<br>
> >>> mpi3-abi@lists.mpi-forum.org<br>
> >>>
<a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi">
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi</a><br>
> >>><br>
> >> _______________________________________________<br>
> >> mpi3-abi mailing list<br>
> >> mpi3-abi@lists.mpi-forum.org<br>
> >>
<a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi">
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi</a><br>
> >><br>
> ><br>
> ><br>
> > _______________________________________________<br>
> > mpi3-abi mailing list<br>
> > mpi3-abi@lists.mpi-forum.org<br>
> >
<a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi">
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi</a><br>
> ><br>
><br>
>_______________________________________________<br>
>mpi3-abi mailing list<br>
>mpi3-abi@lists.mpi-forum.org<br>
><a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi" eudora="autourl">
</a><a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi" eudora="autourl">http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi</a><br><br>
<br>
_______________________________________________<br>
mpi3-abi mailing list<br>
mpi3-abi@lists.mpi-forum.org<br>
<a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi">
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi</a><br><br>
_______________________________________________<br>
mpi3-abi mailing list<br>
mpi3-abi@lists.mpi-forum.org<br>
<a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi">
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi</a><br>
_______________________________________________<br>
mpi3-abi mailing list<br>
mpi3-abi@lists.mpi-forum.org<br>
<a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi" eudora="autourl">
http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-abi</a></blockquote>
</body>
</html>