Send discuss mailing list submissions to
discuss@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@mpich.org
You can reach the person managing the list at
discuss-owner@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@hawk.iit.edu>
To: discuss@mpich.org
Subject: [mpich-discuss] beginner for remote mem access
Message-ID:
<CAFNNHkwxqXUB_+b8_tZ=L5K7VcDVhsz4ZMDDMGF55v-FAhgeqg@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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/discuss/attachments/20130703/91c4ec09/attachment-0001.html>
------------------------------
Message: 2
Date: Wed, 3 Jul 2013 13:18:26 -0500
From: Yi Gu <gyi@mtu.edu>
To: discuss@mpich.org
Subject: Re: [mpich-discuss] beginner for remote mem access
Message-ID:
<CAE5iO_SNJKyC9sG3XbTu4vbS=boDrOEV-Y69jUXZR6MQDanUZA@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"
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@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@mpich.org
> To manage subscription options or unsubscribe:
> https://lists.mpich.org/mailman/listinfo/discuss
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/discuss/attachments/20130703/4414c5cd/attachment-0001.html>
------------------------------
Message: 3
Date: Wed, 3 Jul 2013 13:26:29 -0500
From: Yi Gu <gyi@mtu.edu>
To: discuss@mpich.org
Subject: Re: [mpich-discuss] beginner for remote mem access
Message-ID:
<CAE5iO_Rj7uTaVsEmJ1LBHe+OpTCJjLqpMmvBmjZRd=6WLWC=EQ@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"
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@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@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@mpich.org
>> To manage subscription options or unsubscribe:
>> https://lists.mpich.org/mailman/listinfo/discuss
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/discuss/attachments/20130703/362a0cda/attachment-0001.html>
------------------------------
Message: 4
Date: Wed, 3 Jul 2013 13:28:31 -0500
From: Jeff Hammond <jeff.science@gmail.com>
To: discuss@mpich.org
Subject: Re: [mpich-discuss] beginner for remote mem access
Message-ID:
<CAGKz=uLwki+cCY2oL_2vFwkwCi=A1Qd9VV1qXHno20gHxqY6iw@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
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@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@mpich.org
> To manage subscription options or unsubscribe:
> https://lists.mpich.org/mailman/listinfo/discuss
--
Jeff Hammond
jeff.science@gmail.com
------------------------------
_______________________________________________
discuss mailing list
discuss@mpich.org
https://lists.mpich.org/mailman/listinfo/discuss
End of discuss Digest, Vol 9, Issue 5
*************************************