January 2014 - discuss - mailman3.cels.anl.gov

Re: [mpich-discuss] build broken on 'master': 'mpitypedefs.h' file not found
by Rob Latham 06 Jan '14

06 Jan '14

On Sun, Jan 05, 2014 at 04:34:45PM -0600, Jed Brown wrote: > Robert Latham <robl(a)mcs.anl.gov> writes: > > > I'll just pull the two lines I need from mpitypedefs.h instead of including the entire file. Until I do that, delete the, #include. It's only needed on bluegene > > Where is it used (git grep is not helping)? Yeah, it's a pain to track down. cscope is more helpful. mpitypedefs.h provides these lines: /* Adding the 32-bit compute/64-bit I/O related type-casts in here as * they are not a part of the MPI standard yet. */ #define MPI_AINT_CAST_TO_VOID_PTR (void *)(MPIR_Pint) #define MPI_VOID_PTR_CAST_TO_MPI_AINT (MPI_Aint)(MPIR_Upint) #define MPI_PTR_DISP_CAST_TO_MPI_AINT (MPI_Aint)(MPIR_Pint) adio/include/adioi.h has this pointer-related lines: #define ADIOI_ENSURE_AINT_FITS_IN_PTR(aint_value) MPIR_Ext_ensure_Aint_fits_in_pointer(aint_value) and finally, MPIR_Ext_ensure_Aint_fits_in_pointer is a macro, provided by src/include/glue_romio.h.in /* a copy of MPID_Ensure_Aint_fits_in_pointer for external use */ #define MPIR_Ext_ensure_Aint_fits_in_pointer(aint) \ MPIR_Ext_assert((aint) == (MPI_Aint)(MPIR_Upint) MPI_AINT_CAST_TO_VOID_PTR(aint)); ==rob -- Rob Latham Mathematics and Computer Science Division Argonne National Lab, IL USA

1 0

Re: [mpich-discuss] MPI fail on 20 processes, but not on 10.
by Pavan Balaji 06 Jan '14

06 Jan '14

There should be automatic flow-control within the network layer to deal with this, unless the receives are not posted causing unexpected messages. Looking through your code, that seems to be at least one of the problems. You are posting one Irecv for each slave, and waiting for all of them to finish, while the slaves are sending out many messages. So you can get many messages from one slave before getting the first message from all the slaves. This creates unexpected messages. Btw, your program can be significantly simplified to a bare-bones 100-line source code, which can make debugging easier. I’d recommend doing that. — Pavan On Jan 6, 2014, at 11:14 AM, Jeff Hammond <jeff.science(a)gmail.com> wrote: > Are you blasting the server (master) with messages from N clients > (slaves)? At some point, that will overwhelm the communication > buffers and fail. Can you turn off eager using the documented > environment variable? Rendezvous-only should be much slower but not > fail. Then you can eliminate the pathological usage in your > application. > > Jeff > > On Sun, Jan 5, 2014 at 10:38 PM, Anatoly G <anatolyrishon(a)gmail.com> wrote: >> Hi. >> I have created an application. This application fails on MPI error. >> Assertion failed in file src/mpid/ch3/channels/nemesis/src/ch3_progress.c at >> line 640: pkt->type >= 0 && pkt->type < MPIDI_NEM_PKT_END >> internal ABORT - process 0 >> >> Scenario: >> Master receives messages from slaves. >> Each slave sends data using MPI_Send. >> Master receives using MPI_Irecv and MPI_Recv. >> >> There are another errors in out*.log files. >> Application doesn't fail with 10 processes, but fails with 20. >> >> execute command: >> mpiexec.hydra -genvall -f MpiConfigMachines1.txt -launcher=rsh -n 20 >> /home/anatol-g/Grape/release_under_constr_MPI_tests_quantum/bin/linux64/rhe6/g++4.4.6/debug/mpi_rcv_any_multithread >> 100000 1000000 -1 -1 1 out >> >> Please help, >> >> Regards, >> Anatoly. >> >> >> >> _______________________________________________ >> discuss mailing list discuss(a)mpich.org >> To manage subscription options or unsubscribe: >> https://lists.mpich.org/mailman/listinfo/discuss > > > > -- > Jeff Hammond > jeff.science(a)gmail.com > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0

Re: [mpich-discuss] MPICH for Windows
by Luiz Carlos da Costa Junior 06 Jan '14

06 Jan '14

Thanks very much! On 6 January 2014 15:18, Pavan Balaji <balaji(a)mcs.anl.gov> wrote: > > On Jan 6, 2014, at 11:10 AM, Luiz Carlos da Costa Junior <lcjunior(a)ufrj.br> > wrote: > > I also have the Intel MPI library. > > Is it based on MPICH? > > Do you know how active they are? > > Yes, Intel MPI is based on MPICH as well, and they are very actively > supporting Windows too. > > Both Microsoft and Intel MPI are well supported on Windows, but Intel MPI > has more MPI features, including almost all of MPI-3 in their latest 5.0 > beta release. If you don’t care about anything more than MPI-2, then > either one should be fine. I think Microsoft MPI is missing one feature in > MPI-2 (either connect/accept or spawn; can’t remember which one). > > If you are looking for > MPI-2 features (including MPI-2.2 or MPI-3), you > should also send a note to the Microsoft folks asking for these features in > Microsoft MPI. Users asking for these features increases their priority to > implement them. > > — Pavan > > -- > Pavan Balaji > http://www.mcs.anl.gov/~balaji > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss >

1 0

Re: [mpich-discuss] MPICH for Windows
by Pavan Balaji 06 Jan '14

06 Jan '14

On Jan 6, 2014, at 11:10 AM, Luiz Carlos da Costa Junior <lcjunior(a)ufrj.br> wrote: > I also have the Intel MPI library. > Is it based on MPICH? > Do you know how active they are? Yes, Intel MPI is based on MPICH as well, and they are very actively supporting Windows too. Both Microsoft and Intel MPI are well supported on Windows, but Intel MPI has more MPI features, including almost all of MPI-3 in their latest 5.0 beta release. If you don’t care about anything more than MPI-2, then either one should be fine. I think Microsoft MPI is missing one feature in MPI-2 (either connect/accept or spawn; can’t remember which one). If you are looking for > MPI-2 features (including MPI-2.2 or MPI-3), you should also send a note to the Microsoft folks asking for these features in Microsoft MPI. Users asking for these features increases their priority to implement them. — Pavan -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0

Re: [mpich-discuss] MPI fail on 20 processes, but not on 10.
by Jeff Hammond 06 Jan '14

06 Jan '14

Are you blasting the server (master) with messages from N clients (slaves)? At some point, that will overwhelm the communication buffers and fail. Can you turn off eager using the documented environment variable? Rendezvous-only should be much slower but not fail. Then you can eliminate the pathological usage in your application. Jeff On Sun, Jan 5, 2014 at 10:38 PM, Anatoly G <anatolyrishon(a)gmail.com> wrote: > Hi. > I have created an application. This application fails on MPI error. > Assertion failed in file src/mpid/ch3/channels/nemesis/src/ch3_progress.c at > line 640: pkt->type >= 0 && pkt->type < MPIDI_NEM_PKT_END > internal ABORT - process 0 > > Scenario: > Master receives messages from slaves. > Each slave sends data using MPI_Send. > Master receives using MPI_Irecv and MPI_Recv. > > There are another errors in out*.log files. > Application doesn't fail with 10 processes, but fails with 20. > > execute command: > mpiexec.hydra -genvall -f MpiConfigMachines1.txt -launcher=rsh -n 20 > /home/anatol-g/Grape/release_under_constr_MPI_tests_quantum/bin/linux64/rhe6/g++4.4.6/debug/mpi_rcv_any_multithread > 100000 1000000 -1 -1 1 out > > Please help, > > Regards, > Anatoly. > > > > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss -- Jeff Hammond jeff.science(a)gmail.com

1 0

MPI database connectivity
by bhushan sable 06 Jan '14

06 Jan '14

sir,I am doing final year project in pub-sub system using mpi cluster.For this,i want to know how to connect MPI with database?

1 0

Re: [mpich-discuss] MPICH for Windows
by Luiz Carlos da Costa Junior 06 Jan '14

06 Jan '14

Thanks all for the quick replies. I also have the Intel MPI library. Is it based on MPICH? Do you know how active they are? Which one you recommend or prefer (on Windows)? In Linux, there would be some advantage in using the Intel MPI instead of MPICH? Thanks in advance. Regards, Luiz On 6 January 2014 14:52, Lu, Huiwei <huiweilu(a)mcs.anl.gov> wrote: > > On Jan 6, 2014, at 10:47 AM, Luiz Carlos da Costa Junior <lcjunior(a)ufrj.br> > wrote: > > > I went to MPICH's download page and I understood that MPICH's group no > longer supports Windows distribution because it now points to Microsoft MPI > homepage. Is this correct? > > > > Yes, Luiz, > > We have stopped supporting MPICH on Windows due to the lack of developer > resources. > > http://wiki.mpich.org/mpich/index.php/Frequently_Asked_Questions#Q:_Why_can… > > — > Huiwei > http://www.mcs.anl.gov/~huiweilu/ > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss >

1 0

Re: [mpich-discuss] MPICH for Windows
by Lu, Huiwei 06 Jan '14

06 Jan '14

On Jan 6, 2014, at 10:47 AM, Luiz Carlos da Costa Junior <lcjunior(a)ufrj.br> wrote: > I went to MPICH's download page and I understood that MPICH's group no longer supports Windows distribution because it now points to Microsoft MPI homepage. Is this correct? > Yes, Luiz, We have stopped supporting MPICH on Windows due to the lack of developer resources. http://wiki.mpich.org/mpich/index.php/Frequently_Asked_Questions#Q:_Why_can… — Huiwei http://www.mcs.anl.gov/~huiweilu/

1 0

Re: [mpich-discuss] MPICH for Windows
by Wesley Bland 06 Jan '14

06 Jan '14

That’s correct. Microsoft makes their own version of MPI that is based on MPICH. It’s available via the website that we link to on the downloads page and they provide their own support. On Jan 6, 2014, at 10:47 AM, Luiz Carlos da Costa Junior <lcjunior(a)ufrj.br> wrote: > Hi all, > > It has been quite a long time since the last tie I updated the MPI's version of the Windows version of my software (I am using still using MPICH 1.4.1). I am having many problems regarding folder permissions and credentials using Windows 8. > > I went to MPICH's download page and I understood that MPICH's group no longer supports Windows distribution because it now points to Microsoft MPI homepage. Is this correct? > > Regards, > Luiz > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss

1 0

Re: [mpich-discuss] MPICH for Windows
by Pavan Balaji 06 Jan '14

06 Jan '14

Hi Luiz, On Jan 6, 2014, at 10:47 AM, Luiz Carlos da Costa Junior <lcjunior(a)ufrj.br> wrote: > I went to MPICH's download page and I understood that MPICH's group no longer supports Windows distribution because it now points to Microsoft MPI homepage. Is this correct? Yes, that’s correct. We recommend using Microsoft MPI, which (obviously) does support Windows. http://msdn.microsoft.com/en-us/library/bb524831%28v=vs.85%29.aspx It’s free to use and redistribute. — Pavan -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0