- mpi-comments - mailman3.cels.anl.gov

Re: [Mpi-comments] [mpiwg-p2p] Errata to MPI-3.0 and MPI-3.1 about MPI_WAIT
by Thakur, Rajeev 28 Dec '15

28 Dec '15

This was reported before. See http://lists.mpi-forum.org/mpiwg-p2p/2015/06/0122.php Rajeev > On Dec 28, 2015, at 5:50 AM, Rolf Rabenseifner <rabenseifner(a)hlrs.de> wrote: > > Dear pt-2-pt WG and all, > > In the description of MPI_WAIT, > MPI-3.1 page 53 line 24-25 and > MPI-3.0 page 53 line 27-28 reads > > Any other type of request is > and the request handle is set to MPI_REQUEST_NULL. > > but it should read > > Any other type of request is deallocated > and the request handle is set to MPI_REQUEST_NULL. > > This corrected text was voted into MPI-3.0 through ticket 321. > The missing "deallocated" is a "typo" probably during > the implementation of this ticket before MPI-3.0 was published. > > Ticket 321: > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/321 > and > https://svn.mpi-forum.org/trac/mpi-forum-web/attachment/ticket/321/ticket32… > page 54, lines 14-18. > > This is a request for an erratum for MPI-3.1, > which also applies to MPI-3.0. > > Best regards > Rolf > > > > -- > Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)hlrs.de > High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 > University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 > Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner > Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room 1.307) > _______________________________________________ > mpiwg-p2p mailing list > mpiwg-p2p(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p

1 0

Re: [Mpi-comments] [mpiwg-p2p] Errata to MPI-3.0 and MPI-3.1 about MPI_WAIT
by Rolf Rabenseifner 28 Dec '15

28 Dec '15

My understanding of English is that > Any other type of request is > and the request handle is set to MPI_REQUEST_NULL. is a broken sentence. The Erratum should only correct the wrong implementation of ticket 321. It has nothing to do with any discussion of whether the current description of MPI_WAIT (including the missing word "deallocated") is perfect. I would say, that any further change of the wording would be another discussion for the next version of the standard, or for another erratum. Rolf ----- Original Message ----- > From: "Richard Graham" <richardg(a)mellanox.com> > To: "MPI WG Point To Point Communications working group" <mpiwg-p2p(a)lists.mpi-forum.org> > Cc: "mpi-comments" <mpi-comments(a)mpi-forum.org> > Sent: Monday, December 28, 2015 1:32:46 PM > Subject: Re: [mpiwg-p2p] Errata to MPI-3.0 and MPI-3.1 about MPI_WAIT > Just saw this, but not sure I agree with the correction. What actually happens > is implementation dependent, but the handle the user sees is indeed > MPI_REQUEST_NULL. > > Rich > > -----Original Message----- > From: mpiwg-p2p [mailto:[email protected]] On Behalf Of Rolf > Rabenseifner > Sent: Monday, December 28, 2015 6:12 AM > To: MPI WG Point To Point Communications working group > <mpiwg-p2p(a)lists.mpi-forum.org> > Cc: mpi-comments <mpi-comments(a)mpi-forum.org> > Subject: [mpiwg-p2p] Errata to MPI-3.0 and MPI-3.1 about MPI_WAIT > > Dear pt-2-pt WG and all, > > In the description of MPI_WAIT, > MPI-3.1 page 53 line 24-25 and > MPI-3.0 page 53 line 27-28 reads > > Any other type of request is > and the request handle is set to MPI_REQUEST_NULL. > > but it should read > > Any other type of request is deallocated > and the request handle is set to MPI_REQUEST_NULL. > > This corrected text was voted into MPI-3.0 through ticket 321. > The missing "deallocated" is a "typo" probably during the implementation of this > ticket before MPI-3.0 was published. > > Ticket 321: > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/321 > and > https://svn.mpi-forum.org/trac/mpi-forum-web/attachment/ticket/321/ticket32… > page 54, lines 14-18. > > This is a request for an erratum for MPI-3.1, which also applies to MPI-3.0. > > Best regards > Rolf > > > > -- > Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)hlrs.de High > Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 University of > Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 Head of Dpmt Parallel > Computing . . . www.hlrs.de/people/rabenseifner Nobelstr. 19, D-70550 > Stuttgart, Germany . . . . (Office: Room 1.307) > _______________________________________________ > mpiwg-p2p mailing list > mpiwg-p2p(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p > _______________________________________________ > mpiwg-p2p mailing list > mpiwg-p2p(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p -- Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)hlrs.de High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room 1.307)

1 0

Errata to MPI-3.0 and MPI-3.1 about MPI_WAIT
by Rolf Rabenseifner 28 Dec '15

28 Dec '15

Dear pt-2-pt WG and all, In the description of MPI_WAIT, MPI-3.1 page 53 line 24-25 and MPI-3.0 page 53 line 27-28 reads Any other type of request is and the request handle is set to MPI_REQUEST_NULL. but it should read Any other type of request is deallocated and the request handle is set to MPI_REQUEST_NULL. This corrected text was voted into MPI-3.0 through ticket 321. The missing "deallocated" is a "typo" probably during the implementation of this ticket before MPI-3.0 was published. Ticket 321: https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/321 and https://svn.mpi-forum.org/trac/mpi-forum-web/attachment/ticket/321/ticket32… page 54, lines 14-18. This is a request for an erratum for MPI-3.1, which also applies to MPI-3.0. Best regards Rolf -- Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)hlrs.de High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room 1.307)

1 0

Re: [Mpi-comments] Small typos in MPI 3.1
by William Gropp 31 Oct '15

31 Oct '15

Thanks - I’ve added these to our issue tracking for the next version. Bill William Gropp Director, Parallel Computing Institute Thomas M. Siebel Chair in Computer Science Chief Scientist, NCSA University of Illinois Urbana-Champaign On Oct 30, 2015, at 2:26 AM, Mikhail Kurnosov <mkurnosov(a)gmail.com> wrote: > Hello Forum, > > I found small typos in MPI 3.1 Standard (http://www.mpi-forum.org/docs/mpi-3.1/mpi31-report.pdf) > > * Page 176 / line 46: no comma after "MPI_DOUBLE_PRECISION". > * Page 176 / line 47: no comma after "MPI_LONG_DOUBLE". > * Page 177 / line 4: no space before "MPI_C_BOOL". > > WBR, > Mikhail Kurnosov > --- > Computer Systems Department > Siberian State University of Telecommunications and Information Sciences > Email: mkurnosov(a)gmail.com > WWW: www.mkurnosov.net > > _______________________________________________ > mpi-comments mailing list > mpi-comments(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-comments

1 0

Small typos in MPI 3.1
by Mikhail Kurnosov 30 Oct '15

30 Oct '15

Hello Forum, I found small typos in MPI 3.1 Standard (http://www.mpi-forum.org/docs/mpi-3.1/mpi31-report.pdf) * Page 176 / line 46: no comma after "MPI_DOUBLE_PRECISION". * Page 176 / line 47: no comma after "MPI_LONG_DOUBLE". * Page 177 / line 4: no space before "MPI_C_BOOL". WBR, Mikhail Kurnosov --- Computer Systems Department Siberian State University of Telecommunications and Information Sciences Email: mkurnosov(a)gmail.com WWW: www.mkurnosov.net

1 0

Re: [Mpi-comments] Knowing differences between MPI versions
by Bland, Wesley 04 Aug '15

04 Aug '15

It’s not in version 1.3, but for all other versions, you can look at the changelog (Annex B) in the PDF for differences from one version to another. On Aug 4, 2015, at 11:51 AM, Both, Michael (GE Intelligent Platforms, US) <michael.both(a)ge.com<mailto:[email protected]>> wrote: Hi, How do I determine the changes between minor versions? For examples, what’s the difference between version 1.2 and 1.3? I created a 1.2 implementation about 10 years ago, and I’d like to know specifically what I need to add to make it 1.3 compliant. I looked at the 1.3 MPI document but it does not appear to have a section that shows the differences from the 1.2 version. Also, If I intend to update the implementation to 3.1, what is the best way to know the differences between 3.1 and 1.3? Thanks, Michael _______________________________________________ mpi-comments mailing list mpi-comments(a)lists.mpi-forum.org<mailto:[email protected]> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-comments

1 0

Knowing differences between MPI versions
by Both, Michael (GE Intelligent Platforms, US) 04 Aug '15

04 Aug '15

Hi, How do I determine the changes between minor versions? For examples, what's the difference between version 1.2 and 1.3? I created a 1.2 implementation about 10 years ago, and I'd like to know specifically what I need to add to make it 1.3 compliant. I looked at the 1.3 MPI document but it does not appear to have a section that shows the differences from the 1.2 version. Also, If I intend to update the implementation to 3.1, what is the best way to know the differences between 3.1 and 1.3? Thanks, Michael

1 0

Fortran MPI
by Bhanu Pratap 09 Jan '15

09 Jan '15

I am a research student working on a CFD problem using Fortran. The code has to simulate flow of bubbles over a cylindrical body. So bubbles are being injected in a flow field moving in the direction of cylinder. The motion of the bubbles have to be tracked as well as other parameters like change in diameter and local pressureare to be measured. Since a large number of bubbles have to be injected, therefore parallel processing is required using MPI. PROBLEM I am facing right now is that when I am running the code using only one processor it's working correctly but when I increase the number of processors than it starts writing wrong values. If anyone have faced anything like this before or have any experience in this PLEASE HELP. P.s If you need to look into the code I can upload that too.

1 0

(no subject)
by mustafa mohamed 19 Dec '14

19 Dec '14

Hello there ..i already do lab for mpich2 and connect all computers i have and i do cpi program and its work correctly .. but when i execute my program its give a message : the process stop working because one process exiting without calling MPI_Init .. what i can do to make it works.. ??

1 0

Re: [Mpi-comments] non-blocking communication deadlock possible?
by Maciej Szpindler 19 Dec '14

19 Dec '14

Dear Mustafa Mahamed, It sounds like your program is broken but I believe that this list is not the proper place for such issues. Try posting your question to stackoverflow.com instead. Regards, Maciej W dniu 18.12.2014 o 14:05, mustafa mohamed pisze: > Hello there ..i already do lab for mpich2 and connect all computers i have and i do cpi program and its work correctly .. but when i execute my program its give a message : the process stop working because one process exiting without calling MPI_Init .. what i can do to make it works.. ?? > > > On Wednesday, December 10, 2014 4:33 PM, Rolf Rabenseifner <rabenseifner(a)hlrs.de> wrote: > > > Dear Mr. Hehn, > > as long as you start your example with exactly two MPI processes, > your example is guaranteed to run without deadlock > based on the wording on nonblocking progress MPI-3.0 page 56, lines 30-35. > > Therefore, it is still deadlock-free if you Substitute your > >> MPI_Waitall(2,r,s); > > by > > MPI_Wait(&r[0],s[0]); > MPI_Wait(&r[1],s[1]); > > Thank you for your request on whether the MPI standard is > completely describing the nonblocking functionality. > > Best regards > Rolf > > > ----- Original Message ----- >> From: "Andreas Hehn" <hehn(a)phys.ethz.ch> >> To: mpi-comments(a)lists.mpi-forum.org >> Sent: Wednesday, December 10, 2014 12:15:28 PM >> Subject: [Mpi-comments] non-blocking communication deadlock possible? >> >> Dear MPI comments readers, >> >> I am not 100% sure if the following piece of code could cause a deadlock >> in a standard-compliant MPI implementation or not. >> >> #include <mpi.h> >> >> int main(int argc, char** argv) { >> MPI_Status s[2]; >> int num; >> >> MPI_Init(&argc, &argv); >> MPI_Comm_rank(MPI_COMM_WORLD,&num); >> >> double ds=5.4; // to send >> double dr; // to receive >> int tag=99; >> >> MPI_Request r[2]; >> if(num==0) { >> MPI_Isend(&ds,1,MPI_DOUBLE,1,tag,MPI_COMM_WORLD,&r[0]); >> MPI_Irecv (&dr,1,MPI_DOUBLE,1,tag,MPI_COMM_WORLD,&r[1]); >> } >> else { >> MPI_Isend(&ds,1,MPI_DOUBLE,0,tag,MPI_COMM_WORLD,&r[0]); >> MPI_Irecv (&dr,1,MPI_DOUBLE,0,tag,MPI_COMM_WORLD,&r[1]); >> } >> >> MPI_Waitall(2,r,s); >> >> MPI_Finalize(); >> return 0; >> } >> >> It is not entirely clear to me if the order of MPI_Isend and MPI_Irecv >> matters (as it does for blocking communication). Both function calls >> should return immediately according to section 3.7 of the standard - >> even if MPI_Isend uses synchronous send. >> So at this point there shouldn't be any problem. >> However, will MPI_Waitall() return for any order of the filed requests? >> The standard says it is equivalent to MPI_Wait() for the number of >> requests in "some arbitrary order" (line 28 of page 59). >> I suppose this means MPI_Waitall() is not supposed to mind the order of >> the requests in any way. Therefore it could not deadlock. >> Is this correct and is the line of argument sufficient to exclude the >> possibility of a deadlock? >> >> Best regards, >> >> Andreas Hehn >> _______________________________________________ >> mpi-comments mailing list >> mpi-comments(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-comments >> > > > > _______________________________________________ > mpi-comments mailing list > mpi-comments(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-comments >

1 0