July 2013 - discuss - mailman3.cels.anl.gov

Running configure on Suse Enterprise 10
by anton.s.murfin＠sellafieldsites.com 05 Jul '13

05 Jul '13

1 0

Re: [mpich-discuss] beginner for remote mem access
by Jeff Hammond 03 Jul '13

03 Jul '13

1) When you reply to the digest emails, change the subject to that of the thread you care to respond to. That will make it much easier for others to track the thread. 2) MPI_Win_fence is not similar to MPI_Barrier in some cases. The MPICH implementation can turn some calls into no-ops and others into MPI_Reduce_scatter. 3) Regarding "use MPI window as thread level", I really don't understand your thinking at all. MPI RMA is not shared memory nor is MPI a threading model. Jeff On Wed, Jul 3, 2013 at 2:46 PM, Sufeng Niu <sniu(a)hawk.iit.edu> wrote: > > Thanks a lot for your guys reply. I should initial b array, I figure out the > MPI_Win_fence should be called by all processes, which will solve the > assertion failed issue. > > it seems that MPI_Win_fence is very similar to MPI_Barrier, Can I say > MPI_Win_fence is the RMA version of MPI_Barrier? > > sorry about so many questions. Can I use MPI window as thread level? let's > say I have 4 processes, each process has 8 threads. if only one thread is > used to create mem access window, other process copy it and internal 8 > threads use shared memory to perform some operations. Is that ok? > > Thanks a lot! > > > > On Wed, Jul 3, 2013 at 1:28 PM, <discuss-request(a)mpich.org> wrote: >> >> Send discuss mailing list submissions to >> discuss(a)mpich.org >> >> To subscribe or unsubscribe via the World Wide Web, visit >> https://lists.mpich.org/mailman/listinfo/discuss >> or, via email, send a message with subject or body 'help' to >> discuss-request(a)mpich.org >> >> You can reach the person managing the list at >> discuss-owner(a)mpich.org >> >> When replying, please edit your Subject line so it is more specific >> than "Re: Contents of discuss digest..." >> >> >> Today's Topics: >> >> 1. beginner for remote mem access (Sufeng Niu) >> 2. Re: beginner for remote mem access (Yi Gu) >> 3. Re: beginner for remote mem access (Yi Gu) >> 4. Re: beginner for remote mem access (Jeff Hammond) >> >> >> ---------------------------------------------------------------------- >> >> Message: 1 >> Date: Wed, 3 Jul 2013 13:12:46 -0500 >> From: Sufeng Niu <sniu(a)hawk.iit.edu> >> To: discuss(a)mpich.org >> Subject: [mpich-discuss] beginner for remote mem access >> Message-ID: >> >> <CAFNNHkwxqXUB_+b8_tZ=L5K7VcDVhsz4ZMDDMGF55v-FAhgeqg(a)mail.gmail.com> >> Content-Type: text/plain; charset="iso-8859-1" >> >> Hi, >> >> I am a beginner and just try to use remote memory access, and I wrote a >> simple program to test it: >> >> #include "mpi.h" >> #include <stdio.h> >> #define SIZE 8 >> >> int main(int argc, char *argv[]) >> { >> int numtasks, rank, source=0, dest, tag=1, i; >> float a[64] = >> { >> 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, >> 12.0, 13.0, 14.0, 15.0, 16.0, >> 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, >> 12.0, 13.0, 14.0, 15.0, 16.0, >> 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, >> 12.0, 13.0, 14.0, 15.0, 16.0, >> 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, >> 12.0, 13.0, 14.0, 15.0, 16.0, >> }; >> float b[SIZE]; >> >> MPI_Status stat; >> >> MPI_Init(&argc,&argv); >> MPI_Comm_rank(MPI_COMM_WORLD, &rank); >> MPI_Comm_size(MPI_COMM_WORLD, &numtasks); >> >> MPI_Win win; >> >> // check processor rank >> char processor_name[MPI_MAX_PROCESSOR_NAME]; >> int name_len; >> MPI_Get_processor_name(processor_name, &name_len); >> printf("-- processor %s, rank %d out of %d processors\n", >> processor_name, rank, numtasks); >> >> MPI_Barrier(MPI_COMM_WORLD); >> >> if (numtasks == 4) { >> if (rank == 0) { >> printf("create window \n"); >> MPI_Win_create(a, 8*sizeof(float), sizeof(float), >> MPI_INFO_NULL, MPI_COMM_WORLD, &win); >> >> } >> else { >> MPI_Win_create(MPI_BOTTOM, 0, sizeof(float), >> MPI_INFO_NULL, MPI_COMM_WORLD, &win); >> } >> >> MPI_Win_fence(0, win); >> >> if (rank == 1){ >> MPI_Get(b, SIZE, MPI_FLOAT, 0, 8, SIZE, MPI_FLOAT, >> win); >> >> MPI_Win_fence(0, win); >> } >> >> printf("rank= %d b= %3.1f %3.1f %3.1f %3.1f %3.1f %3.1f >> %3.1f %3.1f\n", rank,b[0],b[1],b[2],b[3],b[4],b[5],b[6],b[7]); >> } >> else >> printf("Must specify %d processors. Terminating.\n",SIZE); >> >> MPI_Win_free(&win); >> MPI_Finalize(); >> } >> >> However the terminal gives some odd results: >> rank= 0 b= 0.0 0.0 0.0 0.0 0.0 0.0 -71847793475452928.0 0.0 >> rank= 2 b= 0.0 0.0 0.0 0.0 0.0 0.0 222086852849451401216.0 0.0 >> rank= 3 b= 0.0 0.0 0.0 0.0 0.0 0.0 -74882.4 0.0 >> rank= 1 b= 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 >> the rank 1 is correct results, but others should be all zero. >> terminal also give some comments: "Assertion failed in file >> src/mpid/ch3/src/ch3u_rma_sync.c at line 5061: win_ptr->my_counter >= 0 >> internal ABORT - process 0" >> >> another question is if I use remote memory access. all process which does >> not create window for share must add additional line: >> MPI_Win_create(MPI_BOTTOM, 0, data_type, MPI_INFO_NULL, MPI_COMM_WORLD, >> &win); correct? >> >> Thanks a lot! >> >> -- >> Best Regards, >> Sufeng Niu >> ECASP lab, ECE department, Illinois Institute of Technology >> Tel: 312-731-7219 >>

1 0

Re: [mpich-discuss] discuss Digest, Vol 9, Issue 5
by Sufeng Niu 03 Jul '13

03 Jul '13

Thanks a lot for your guys reply. I should initial b array, I figure out the MPI_Win_fence should be called by all processes, which will solve the assertion failed issue. it seems that MPI_Win_fence is very similar to MPI_Barrier, Can I say MPI_Win_fence is the RMA version of MPI_Barrier? sorry about so many questions. Can I use MPI window as thread level? let's say I have 4 processes, each process has 8 threads. if only one thread is used to create mem access window, other process copy it and internal 8 threads use shared memory to perform some operations. Is that ok? Thanks a lot! On Wed, Jul 3, 2013 at 1:28 PM, <discuss-request(a)mpich.org> wrote: > Send discuss mailing list submissions to > discuss(a)mpich.org > > To subscribe or unsubscribe via the World Wide Web, visit > https://lists.mpich.org/mailman/listinfo/discuss > or, via email, send a message with subject or body 'help' to > discuss-request(a)mpich.org > > You can reach the person managing the list at > discuss-owner(a)mpich.org > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of discuss digest..." > > > Today's Topics: > > 1. beginner for remote mem access (Sufeng Niu) > 2. Re: beginner for remote mem access (Yi Gu) > 3. Re: beginner for remote mem access (Yi Gu) > 4. Re: beginner for remote mem access (Jeff Hammond) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Wed, 3 Jul 2013 13:12:46 -0500 > From: Sufeng Niu <sniu(a)hawk.iit.edu> > To: discuss(a)mpich.org > Subject: [mpich-discuss] beginner for remote mem access > Message-ID: > <CAFNNHkwxqXUB_+b8_tZ= > L5K7VcDVhsz4ZMDDMGF55v-FAhgeqg(a)mail.gmail.com> > Content-Type: text/plain; charset="iso-8859-1" > > Hi, > > I am a beginner and just try to use remote memory access, and I wrote a > simple program to test it: > > #include "mpi.h" > #include <stdio.h> > #define SIZE 8 > > int main(int argc, char *argv[]) > { > int numtasks, rank, source=0, dest, tag=1, i; > float a[64] = > { > 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, > 12.0, 13.0, 14.0, 15.0, 16.0, > 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, > 12.0, 13.0, 14.0, 15.0, 16.0, > 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, > 12.0, 13.0, 14.0, 15.0, 16.0, > 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, > 12.0, 13.0, 14.0, 15.0, 16.0, > }; > float b[SIZE]; > > MPI_Status stat; > > MPI_Init(&argc,&argv); > MPI_Comm_rank(MPI_COMM_WORLD, &rank); > MPI_Comm_size(MPI_COMM_WORLD, &numtasks); > > MPI_Win win; > > // check processor rank > char processor_name[MPI_MAX_PROCESSOR_NAME]; > int name_len; > MPI_Get_processor_name(processor_name, &name_len); > printf("-- processor %s, rank %d out of %d processors\n", > processor_name, rank, numtasks); > > MPI_Barrier(MPI_COMM_WORLD); > > if (numtasks == 4) { > if (rank == 0) { > printf("create window \n"); > MPI_Win_create(a, 8*sizeof(float), sizeof(float), > MPI_INFO_NULL, MPI_COMM_WORLD, &win); > > } > else { > MPI_Win_create(MPI_BOTTOM, 0, sizeof(float), > MPI_INFO_NULL, MPI_COMM_WORLD, &win); > } > > MPI_Win_fence(0, win); > > if (rank == 1){ > MPI_Get(b, SIZE, MPI_FLOAT, 0, 8, SIZE, MPI_FLOAT, > win); > > MPI_Win_fence(0, win); > } > > printf("rank= %d b= %3.1f %3.1f %3.1f %3.1f %3.1f %3.1f > %3.1f %3.1f\n", rank,b[0],b[1],b[2],b[3],b[4],b[5],b[6],b[7]); > } > else > printf("Must specify %d processors. Terminating.\n",SIZE); > > MPI_Win_free(&win); > MPI_Finalize(); > } > > However the terminal gives some odd results: > rank= 0 b= 0.0 0.0 0.0 0.0 0.0 0.0 -71847793475452928.0 0.0 > rank= 2 b= 0.0 0.0 0.0 0.0 0.0 0.0 222086852849451401216.0 0.0 > rank= 3 b= 0.0 0.0 0.0 0.0 0.0 0.0 -74882.4 0.0 > rank= 1 b= 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 > the rank 1 is correct results, but others should be all zero. > terminal also give some comments: "Assertion failed in file > src/mpid/ch3/src/ch3u_rma_sync.c at line 5061: win_ptr->my_counter >= 0 > internal ABORT - process 0" > > another question is if I use remote memory access. all process which does > not create window for share must add additional line: > MPI_Win_create(MPI_BOTTOM, 0, data_type, MPI_INFO_NULL, MPI_COMM_WORLD, > &win); correct? > > Thanks a lot! > > -- > Best Regards, > Sufeng Niu > ECASP lab, ECE department, Illinois Institute of Technology > Tel: 312-731-7219 >

1 0

Re: [mpich-discuss] beginner for remote mem access
by Jeff Hammond 03 Jul '13

03 Jul '13

Look at the MPICH test suite for numerous examples of correct RMA programs. Jeff On Wed, Jul 3, 2013 at 1:12 PM, Sufeng Niu <sniu(a)hawk.iit.edu> wrote: > Hi, > > I am a beginner and just try to use remote memory access, and I wrote a > simple program to test it: > > #include "mpi.h" > #include <stdio.h> > #define SIZE 8 > > int main(int argc, char *argv[]) > { > int numtasks, rank, source=0, dest, tag=1, i; > float a[64] = > { > 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, > 12.0, 13.0, 14.0, 15.0, 16.0, > 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, > 12.0, 13.0, 14.0, 15.0, 16.0, > 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, > 12.0, 13.0, 14.0, 15.0, 16.0, > 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, > 12.0, 13.0, 14.0, 15.0, 16.0, > }; > float b[SIZE]; > > MPI_Status stat; > > MPI_Init(&argc,&argv); > MPI_Comm_rank(MPI_COMM_WORLD, &rank); > MPI_Comm_size(MPI_COMM_WORLD, &numtasks); > > MPI_Win win; > > // check processor rank > char processor_name[MPI_MAX_PROCESSOR_NAME]; > int name_len; > MPI_Get_processor_name(processor_name, &name_len); > printf("-- processor %s, rank %d out of %d processors\n", > processor_name, rank, numtasks); > > MPI_Barrier(MPI_COMM_WORLD); > > if (numtasks == 4) { > if (rank == 0) { > printf("create window \n"); > MPI_Win_create(a, 8*sizeof(float), sizeof(float), > MPI_INFO_NULL, MPI_COMM_WORLD, &win); > > } > else { > MPI_Win_create(MPI_BOTTOM, 0, sizeof(float), > MPI_INFO_NULL, MPI_COMM_WORLD, &win); > } > > MPI_Win_fence(0, win); > > if (rank == 1){ > MPI_Get(b, SIZE, MPI_FLOAT, 0, 8, SIZE, MPI_FLOAT, > win); > > MPI_Win_fence(0, win); > } > > printf("rank= %d b= %3.1f %3.1f %3.1f %3.1f %3.1f %3.1f > %3.1f %3.1f\n", rank,b[0],b[1],b[2],b[3],b[4],b[5],b[6],b[7]); > } > else > printf("Must specify %d processors. Terminating.\n",SIZE); > > MPI_Win_free(&win); > MPI_Finalize(); > } > > However the terminal gives some odd results: > rank= 0 b= 0.0 0.0 0.0 0.0 0.0 0.0 -71847793475452928.0 0.0 > rank= 2 b= 0.0 0.0 0.0 0.0 0.0 0.0 222086852849451401216.0 0.0 > rank= 3 b= 0.0 0.0 0.0 0.0 0.0 0.0 -74882.4 0.0 > rank= 1 b= 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 > the rank 1 is correct results, but others should be all zero. > terminal also give some comments: "Assertion failed in file > src/mpid/ch3/src/ch3u_rma_sync.c at line 5061: win_ptr->my_counter >= 0 > internal ABORT - process 0" > > another question is if I use remote memory access. all process which does > not create window for share must add additional line: > MPI_Win_create(MPI_BOTTOM, 0, data_type, MPI_INFO_NULL, MPI_COMM_WORLD, > &win); correct? > > Thanks a lot! > > -- > Best Regards, > Sufeng Niu > ECASP lab, ECE department, Illinois Institute of Technology > Tel: 312-731-7219 > > _______________________________________________ > 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

Re: [mpich-discuss] beginner for remote mem access
by Yi Gu 03 Jul '13

03 Jul '13

Hi, sufeng: I am not sure the second problem is, but I think MPI_Win_create(MPI_BOTTOM, 0, sizeof(float), MPI_INFO_NULL, MPI_COMM_WORLD, &win); should be MPI_Win_create(MPI_BOTTOM, 0, 1, MPI_INFO_NULL, MPI_COMM_WORLD, &win); For the third question, if they need to access, yes. Yi On Wed, Jul 3, 2013 at 1:18 PM, Yi Gu <gyi(a)mtu.edu> wrote: > Hi, sufeng: > > I think you may first initialize array first, since you print out b > without initialization, > it could print out anything. > > Yi > > > On Wed, Jul 3, 2013 at 1:12 PM, Sufeng Niu <sniu(a)hawk.iit.edu> wrote: > >> Hi, >> >> I am a beginner and just try to use remote memory access, and I wrote a >> simple program to test it: >> >> #include "mpi.h" >> #include <stdio.h> >> #define SIZE 8 >> >> int main(int argc, char *argv[]) >> { >> int numtasks, rank, source=0, dest, tag=1, i; >> float a[64] = >> { >> 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, >> 12.0, 13.0, 14.0, 15.0, 16.0, >> 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, >> 12.0, 13.0, 14.0, 15.0, 16.0, >> 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, >> 12.0, 13.0, 14.0, 15.0, 16.0, >> 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, >> 12.0, 13.0, 14.0, 15.0, 16.0, >> }; >> float b[SIZE]; >> >> MPI_Status stat; >> >> MPI_Init(&argc,&argv); >> MPI_Comm_rank(MPI_COMM_WORLD, &rank); >> MPI_Comm_size(MPI_COMM_WORLD, &numtasks); >> >> MPI_Win win; >> >> // check processor rank >> char processor_name[MPI_MAX_PROCESSOR_NAME]; >> int name_len; >> MPI_Get_processor_name(processor_name, &name_len); >> printf("-- processor %s, rank %d out of %d processors\n", >> processor_name, rank, numtasks); >> >> MPI_Barrier(MPI_COMM_WORLD); >> >> if (numtasks == 4) { >> if (rank == 0) { >> printf("create window \n"); >> MPI_Win_create(a, 8*sizeof(float), sizeof(float), >> MPI_INFO_NULL, MPI_COMM_WORLD, &win); >> >> } >> else { >> MPI_Win_create(MPI_BOTTOM, 0, sizeof(float), >> MPI_INFO_NULL, MPI_COMM_WORLD, &win); >> } >> >> MPI_Win_fence(0, win); >> >> if (rank == 1){ >> MPI_Get(b, SIZE, MPI_FLOAT, 0, 8, SIZE, >> MPI_FLOAT, win); >> >> MPI_Win_fence(0, win); >> } >> >> printf("rank= %d b= %3.1f %3.1f %3.1f %3.1f %3.1f %3.1f >> %3.1f %3.1f\n", rank,b[0],b[1],b[2],b[3],b[4],b[5],b[6],b[7]); >> } >> else >> printf("Must specify %d processors. Terminating.\n",SIZE); >> >> MPI_Win_free(&win); >> MPI_Finalize(); >> } >> >> However the terminal gives some odd results: >> rank= 0 b= 0.0 0.0 0.0 0.0 0.0 0.0 -71847793475452928.0 0.0 >> rank= 2 b= 0.0 0.0 0.0 0.0 0.0 0.0 222086852849451401216.0 0.0 >> rank= 3 b= 0.0 0.0 0.0 0.0 0.0 0.0 -74882.4 0.0 >> rank= 1 b= 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 >> the rank 1 is correct results, but others should be all zero. >> terminal also give some comments: "Assertion failed in file >> src/mpid/ch3/src/ch3u_rma_sync.c at line 5061: win_ptr->my_counter >= 0 >> internal ABORT - process 0" >> >> another question is if I use remote memory access. all process which does >> not create window for share must add additional line: >> MPI_Win_create(MPI_BOTTOM, 0, data_type, MPI_INFO_NULL, MPI_COMM_WORLD, >> &win); correct? >> >> Thanks a lot! >> >> -- >> Best Regards, >> Sufeng Niu >> ECASP lab, ECE department, Illinois Institute of Technology >> Tel: 312-731-7219 >> >> _______________________________________________ >> 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] beginner for remote mem access
by Yi Gu 03 Jul '13

03 Jul '13

Hi, sufeng: I think you may first initialize array first, since you print out b without initialization, it could print out anything. Yi On Wed, Jul 3, 2013 at 1:12 PM, Sufeng Niu <sniu(a)hawk.iit.edu> wrote: > Hi, > > I am a beginner and just try to use remote memory access, and I wrote a > simple program to test it: > > #include "mpi.h" > #include <stdio.h> > #define SIZE 8 > > int main(int argc, char *argv[]) > { > int numtasks, rank, source=0, dest, tag=1, i; > float a[64] = > { > 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, > 12.0, 13.0, 14.0, 15.0, 16.0, > 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, > 12.0, 13.0, 14.0, 15.0, 16.0, > 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, > 12.0, 13.0, 14.0, 15.0, 16.0, > 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, > 12.0, 13.0, 14.0, 15.0, 16.0, > }; > float b[SIZE]; > > MPI_Status stat; > > MPI_Init(&argc,&argv); > MPI_Comm_rank(MPI_COMM_WORLD, &rank); > MPI_Comm_size(MPI_COMM_WORLD, &numtasks); > > MPI_Win win; > > // check processor rank > char processor_name[MPI_MAX_PROCESSOR_NAME]; > int name_len; > MPI_Get_processor_name(processor_name, &name_len); > printf("-- processor %s, rank %d out of %d processors\n", > processor_name, rank, numtasks); > > MPI_Barrier(MPI_COMM_WORLD); > > if (numtasks == 4) { > if (rank == 0) { > printf("create window \n"); > MPI_Win_create(a, 8*sizeof(float), sizeof(float), > MPI_INFO_NULL, MPI_COMM_WORLD, &win); > > } > else { > MPI_Win_create(MPI_BOTTOM, 0, sizeof(float), > MPI_INFO_NULL, MPI_COMM_WORLD, &win); > } > > MPI_Win_fence(0, win); > > if (rank == 1){ > MPI_Get(b, SIZE, MPI_FLOAT, 0, 8, SIZE, MPI_FLOAT, > win); > > MPI_Win_fence(0, win); > } > > printf("rank= %d b= %3.1f %3.1f %3.1f %3.1f %3.1f %3.1f > %3.1f %3.1f\n", rank,b[0],b[1],b[2],b[3],b[4],b[5],b[6],b[7]); > } > else > printf("Must specify %d processors. Terminating.\n",SIZE); > > MPI_Win_free(&win); > MPI_Finalize(); > } > > However the terminal gives some odd results: > rank= 0 b= 0.0 0.0 0.0 0.0 0.0 0.0 -71847793475452928.0 0.0 > rank= 2 b= 0.0 0.0 0.0 0.0 0.0 0.0 222086852849451401216.0 0.0 > rank= 3 b= 0.0 0.0 0.0 0.0 0.0 0.0 -74882.4 0.0 > rank= 1 b= 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 > the rank 1 is correct results, but others should be all zero. > terminal also give some comments: "Assertion failed in file > src/mpid/ch3/src/ch3u_rma_sync.c at line 5061: win_ptr->my_counter >= 0 > internal ABORT - process 0" > > another question is if I use remote memory access. all process which does > not create window for share must add additional line: > MPI_Win_create(MPI_BOTTOM, 0, data_type, MPI_INFO_NULL, MPI_COMM_WORLD, > &win); correct? > > Thanks a lot! > > -- > Best Regards, > Sufeng Niu > ECASP lab, ECE department, Illinois Institute of Technology > Tel: 312-731-7219 > > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss >

1 0

beginner for remote mem access
by Sufeng Niu 03 Jul '13

03 Jul '13

Hi, I am a beginner and just try to use remote memory access, and I wrote a simple program to test it: #include "mpi.h" #include <stdio.h> #define SIZE 8 int main(int argc, char *argv[]) { int numtasks, rank, source=0, dest, tag=1, i; float a[64] = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0, }; float b[SIZE]; MPI_Status stat; MPI_Init(&argc,&argv); MPI_Comm_rank(MPI_COMM_WORLD, &rank); MPI_Comm_size(MPI_COMM_WORLD, &numtasks); MPI_Win win; // check processor rank char processor_name[MPI_MAX_PROCESSOR_NAME]; int name_len; MPI_Get_processor_name(processor_name, &name_len); printf("-- processor %s, rank %d out of %d processors\n", processor_name, rank, numtasks); MPI_Barrier(MPI_COMM_WORLD); if (numtasks == 4) { if (rank == 0) { printf("create window \n"); MPI_Win_create(a, 8*sizeof(float), sizeof(float), MPI_INFO_NULL, MPI_COMM_WORLD, &win); } else { MPI_Win_create(MPI_BOTTOM, 0, sizeof(float), MPI_INFO_NULL, MPI_COMM_WORLD, &win); } MPI_Win_fence(0, win); if (rank == 1){ MPI_Get(b, SIZE, MPI_FLOAT, 0, 8, SIZE, MPI_FLOAT, win); MPI_Win_fence(0, win); } printf("rank= %d b= %3.1f %3.1f %3.1f %3.1f %3.1f %3.1f %3.1f %3.1f\n", rank,b[0],b[1],b[2],b[3],b[4],b[5],b[6],b[7]); } else printf("Must specify %d processors. Terminating.\n",SIZE); MPI_Win_free(&win); MPI_Finalize(); } However the terminal gives some odd results: rank= 0 b= 0.0 0.0 0.0 0.0 0.0 0.0 -71847793475452928.0 0.0 rank= 2 b= 0.0 0.0 0.0 0.0 0.0 0.0 222086852849451401216.0 0.0 rank= 3 b= 0.0 0.0 0.0 0.0 0.0 0.0 -74882.4 0.0 rank= 1 b= 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 the rank 1 is correct results, but others should be all zero. terminal also give some comments: "Assertion failed in file src/mpid/ch3/src/ch3u_rma_sync.c at line 5061: win_ptr->my_counter >= 0 internal ABORT - process 0" another question is if I use remote memory access. all process which does not create window for share must add additional line: MPI_Win_create(MPI_BOTTOM, 0, data_type, MPI_INFO_NULL, MPI_COMM_WORLD, &win); correct? Thanks a lot! -- Best Regards, Sufeng Niu ECASP lab, ECE department, Illinois Institute of Technology Tel: 312-731-7219

1 0

Re: [mpich-discuss] MPI_Send and recv
by Wesley Bland 03 Jul '13

03 Jul '13

On Jul 2, 2013, at 9:03 PM, Matthieu Dorier <matthieu.dorier(a)irisa.fr> wrote: > Hi, > > These messages piqued my curiosity; > > If I have two processes, process 0 does an MPI_Send to process 1 and eventually, process 1 posts a matching MPI_Recv or MPI_Irecv; does process 0 waits for the MPI_Recv/Irecv to be posted before returning from the MPI_Send? (in other words if process 0 does 2 MPI_Send and process 1 does 2 MPI_Recv, is the linearization "Send; Send; Recv; Recv" possible?) It is possible that process 0 could return from the MPI_Send before process 1 posts the corresponding MPI_Recv, but it's not guaranteed. The only thing that is guaranteed from the standard is that when MPI_Send returns, the buffer is available to use again. It is valid to wait for the MPI_Recv to post as well. So to answer your example, S - S - R - R is valid, as is S - R - S - R. > > Second question: in the situation described by Sufeng, where a process does a MPI_Send to itself, MPI has enough information to see that the process is sending to itself while it has not posted any matching receive, why would MPI block instead of printing an error and aborting? Is it because in a multithreaded program, the matching receive could be posted by another thread? (Yet in that case I always thought it was unsafe to use the same communicator from multiple threads…) There are probably many reasons why MPI doesn't try to do internal error checking. The first is performance. If a program is correct, then checking for all of the cases that might cause a deadlock would be not only very difficult, but also expensive. Therefore, we just say that the user needs to write a correct program and MPI makes no guarantees that anything smart will happen in the background. It's possible that someone could write Nice-MPI that does lots of error checking and verification, but I would think that most users wouldn't want it as it would add overhead. Second, it's not always unsafe to use the same communicator in multiple threads. If you use different tags for each thread, then it shouldn't be a problem to communicate between them. Wesley > > Thanks, > > Matthieu > > De: "Sufeng Niu" <sniu(a)hawk.iit.edu> > À: "Pavan Balaji" <balaji(a)mcs.anl.gov> > Cc: discuss(a)mpich.org > Envoyé: Mardi 2 Juillet 2013 17:20:25 > Objet: Re: [mpich-discuss] MPI_Send and recv > > Thank you so much! Now I make sense > > Sufeng > > > On Tue, Jul 2, 2013 at 5:19 PM, Pavan Balaji <balaji(a)mcs.anl.gov> wrote: > > On 07/02/2013 05:18 PM, Sufeng Niu wrote: > Thanks a lot, that is most confusing part for me in MPI. I would like to > clear this type of problems. > > So the blocking begins to send message to itself, but there is no > receive posted. Thus, the MPI block it. Am I right here? > > Correct. > > > Thus, I need a nonblocking receive to wait the message here. Can I use > blocking receive? > > If you use blocking receive, you'll block in the receive; how will you reach the send? > > -- Pavan > > > -- > Pavan Balaji > http://www.mcs.anl.gov/~balaji > > > > -- > Best Regards, > Sufeng Niu > ECASP lab, ECE department, Illinois Institute of Technology > Tel: 312-731-7219 > > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss > > _______________________________________________ > 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] MPI_Send and recv
by Matthieu Dorier 02 Jul '13

02 Jul '13

Hi, These messages piqued my curiosity; If I have two processes, process 0 does an MPI_Send to process 1 and eventually, process 1 posts a matching MPI_Recv or MPI_Irecv; does process 0 waits for the MPI_Recv/Irecv to be posted before returning from the MPI_Send? (in other words if process 0 does 2 MPI_Send and process 1 does 2 MPI_Recv, is the linearization "Send; Send; Recv; Recv" possible?) Second question: in the situation described by Sufeng, where a process does a MPI_Send to itself, MPI has enough information to see that the process is sending to itself while it has not posted any matching receive, why would MPI block instead of printing an error and aborting? Is it because in a multithreaded program, the matching receive could be posted by another thread? (Yet in that case I always thought it was unsafe to use the same communicator from multiple threads...) Thanks, Matthieu ----- Mail original ----- > De: "Sufeng Niu" <sniu(a)hawk.iit.edu> > À: "Pavan Balaji" <balaji(a)mcs.anl.gov> > Cc: discuss(a)mpich.org > Envoyé: Mardi 2 Juillet 2013 17:20:25 > Objet: Re: [mpich-discuss] MPI_Send and recv > Thank you so much! Now I make sense > Sufeng > On Tue, Jul 2, 2013 at 5:19 PM, Pavan Balaji < balaji(a)mcs.anl.gov > > wrote: > > On 07/02/2013 05:18 PM, Sufeng Niu wrote: > > > > Thanks a lot, that is most confusing part for me in MPI. I would > > > like > > > to > > > > > > clear this type of problems. > > > > > > So the blocking begins to send message to itself, but there is no > > > > > > receive posted. Thus, the MPI block it. Am I right here? > > > > > Correct. > > > > Thus, I need a nonblocking receive to wait the message here. Can > > > I > > > use > > > > > > blocking receive? > > > > > If you use blocking receive, you'll block in the receive; how will > > you reach the send? > > > -- Pavan > > > -- > > > Pavan Balaji > > > http://www.mcs.anl.gov/~balaji > > -- > Best Regards, > Sufeng Niu > ECASP lab, ECE department, Illinois Institute of Technology > Tel: 312-731-7219 > _______________________________________________ > 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] MPI_Send and recv
by Sufeng Niu 02 Jul '13

02 Jul '13

Thank you so much! Now I make sense Sufeng On Tue, Jul 2, 2013 at 5:19 PM, Pavan Balaji <balaji(a)mcs.anl.gov> wrote: > > On 07/02/2013 05:18 PM, Sufeng Niu wrote: > >> Thanks a lot, that is most confusing part for me in MPI. I would like to >> clear this type of problems. >> >> So the blocking begins to send message to itself, but there is no >> receive posted. Thus, the MPI block it. Am I right here? >> > > Correct. > > > Thus, I need a nonblocking receive to wait the message here. Can I use >> blocking receive? >> > > If you use blocking receive, you'll block in the receive; how will you > reach the send? > > -- Pavan > > > -- > Pavan Balaji > http://www.mcs.anl.gov/~balaji > -- Best Regards, Sufeng Niu ECASP lab, ECE department, Illinois Institute of Technology Tel: 312-731-7219

1 0