May 2014 - mpiwg-rma - mailman3.cels.anl.gov

Re: [mpiwg-rma] #429 - Re: Problems with RMA synchronization in combination with load/store shared memory accesses
by Jim Dinan 22 May '14

22 May '14

Hi Rolf, You suggested adding the following text to the description of active target assertions: Note that in shared memory windows (allocated with MPI_WIN_ALLOCATE_SHARED), there is no difference between remote store accesses and local store accesses to the window, and the same for remote and local loads. I'm not that comfortable with this text. We definitely do have to handle local and remote stores differently within the MPI implementation in order to ensure that they become visible according to the active target synchronization model. I think we may need to update the NOPRECEDE, NOSUCCEED, and NOPUT assertions to apply to both RMA updates and updates to shared memory windows. For example, we could update the text as follows: MPI_MODE_NOPRECEDE - the fence does not complete any sequence of locally issued RMA calls or shared memory window updates. If this assertion is given by any process in the window group, then it must be given by all processes in the group. ~Jim. On Mon, May 19, 2014 at 1:59 PM, Rolf Rabenseifner <rabenseifner(a)hlrs.de>wrote: > Jim, > > I mnade #429. > It is reported by me, but not owned, i.e., the RMA WG should take it. > If the proposed text is okay for you (or with minor updates) > then you can vote on it. > > I'll put it on the errata reading+voting list. > > It is the minimal text you can do that users and implementers > see, that shared Memory Windows are Special for assertions. > Therefore also the change-log, because implementors > had and may also in future oversee this Problem. > > Best regards > Rolf > > ----- Original Message ----- > > From: "Rolf Rabenseifner" <rabenseifner(a)hlrs.de> > > To: "MPI WG Remote Memory Access working group" < > mpiwg-rma(a)lists.mpi-forum.org> > > Sent: Monday, May 19, 2014 7:16:44 PM > > Subject: Re: [mpiwg-rma] Problems with RMA synchronization in > combination with load/store shared memory accesses > > > > Jim and RMA WG, > > > > There are now two questions: > > > > Jim asked: > > > Question to WG: Do we need to update the fence assertions to better > > > define interaction with local load/store accesses and remote > > > stores? > > > > > > > Rolf asked: > > > Additionally, I would recommend that we add after MPI-3.0 p451:33 > > > > > > Note that in shared memory windows (allocated with > > > MPI_WIN_ALLOCATE_SHARED), there is no difference > > > between remote store accesses and local store accesses > > > to the window. > > > > > > This would help to understand that "the local window > > > was not updated by stores" does not mean "by local stores", > > > see p452:1 and p452:9. > > > > For me, it is important to understand the meaning of the > > current assertions if they are used in a shared memory window. > > Therefore my proposal above as erratum to MPI-3.0. > > > > In MPI-3.1 and 4.0, you may want to add additional assertions. > > > > Your analysis below, will also show that mpich implements > > Post-Start-Complete-Wait synchronization in a wrong way, > > if there are no calls to RMA routines. > > > > Best regards > > Rolf > > > > ----- Original Message ----- > > > From: "Jim Dinan" <james.dinan(a)gmail.com> > > > To: "MPI WG Remote Memory Access working group" > > > <mpiwg-rma(a)lists.mpi-forum.org> > > > Sent: Thursday, May 15, 2014 4:06:08 PM > > > Subject: Re: [mpiwg-rma] Problems with RMA synchronization in > > > combination with load/store shared memory accesses > > > > > > > > > > > > Rolf, > > > > > > > > > Here is an attempt to simplify your example for discussion. Given > > > a > > > shared memory window, shr_mem_win, with buffer, shr_mem_buf: > > > > > > MPI_Win_fence(MPI_MODE_NOSTORE | MPI_MODE_NOPUT | > > > MPI_MODE_NOPRECEDE > > > | MPI_MODE_NOSUCCEED, shr_mem_win ); > > > > > > shr_mem_buf[...] = ...; > > > > > > MPI_Win_fence(MPI_MODE_NOPUT | MPI_MODE_NOPRECEDE | > > > MPI_MODE_NOSUCCEED, shr_mem_win); > > > > > > > > > Right now, Fence assertions don't say anything special about shared > > > memory windows: > > > > > > > > > Inline image 1 > > > > > > > > > NOPRECEDE/SUCCEED are defined in terms of MPI RMA function calls, > > > and > > > do not cover load/store. Thus, Rolf's usage appears to be correct > > > per the current text. In the MPICH fence implementation, > > > src/mpid/ch3/src/ch3u_rma_sync.c:935 we have: > > > > > > if (!(assert & MPI_MODE_NOSUCCEED)) win_ptr->fence_issued = 1; > > > > > > Because of this check, we don't actually start an active target > > > epoch > > > on the first fence in the example above. On the second fence, we > > > therefore don't perform the necessary synchronization, leading to > > > incorrect output in Rolf's example. > > > > > > > > > Question to WG: Do we need to update the fence assertions to better > > > define interaction with local load/store accesses and remote > > > stores? > > > > > > > > > If not, then Rolf's code is correct and we need to modify the check > > > above in MPICH to something like: > > > > > > > > > if (!(assert & MPI_MODE_NOSUCCEED) || win_ptr->create_ flavor == > > > MPI_WIN_FLAVOR_SHARED ) > > > win_ptr->fence_issued = 1; > > > > > > > > > ~Jim. > > > > > > > > > > > > > > > On Tue, Apr 8, 2014 at 12:02 PM, Rolf Rabenseifner < > > > rabenseifner(a)hlrs.de > wrote: > > > > > > > > > Jim, > > > > > > I'm now sure, that mpich has a bug with assertions on shared memory > > > windows. > > > > > > in the example, rcv_buf_left and rcv_buf_right are the windows. > > > the only accesses to these rcv_buf_... are stores from remote > > > and fully local loads. > > > Both accesses are done in different epochs surrounded by > > > MPI_Win_fence. > > > > > > According to your interpretation (which is really okay) > > > all fences can use all possible assertions (!!!), > > > except after the remote stores, MPI_MODE_NOSTORE cannot be used. > > > > > > I updated the example and mpich is executing it wrong. > > > > > > Please check it yourself on your installation: > > > halo_1sided_store_win_alloc_shared_w-a-2-cray.c > > > > > > Without the assertions, all works: > > > halo_1sided_store_win_alloc_shared_w-a-2NO-cray.c > > > > > > Could you verify that mpich has a bug? > > > > > > Additionally, I would recommend that we add after MPI-3.0 p451:33 > > > > > > Note that in shared memory windows (allocated with > > > MPI_WIN_ALLOCATE_SHARED), there is no difference > > > between remote store accesses and local store accesses > > > to the window. > > > > > > This would help to understand that "the local window > > > was not updated by stores" does not mean "by local stores", > > > see p452:1 and p452:9. > > > > > > Is it a good idea? > > > > > > Best regards > > > Rolf > > > > > > > > > > > > ----- Original Message ----- > > > > From: "Jim Dinan" < james.dinan(a)gmail.com > > > > > To: "MPI WG Remote Memory Access working group" < > > > > mpiwg-rma(a)lists.mpi-forum.org > > > > > Sent: Friday, March 21, 2014 8:14:22 PM > > > > Subject: Re: [mpiwg-rma] Problems with RMA synchronization in > > > > combination with load/store shared memory accesses > > > > > > > > > > > > > > > > > > > Rolf, > > > > > > > > > > > > This line is incorrect: MPI_Win_fence(MPI_MODE_NOSTORE + > > > > MPI_MODE_NOPRECEDE, win_ rcv_buf _left ); > > > > > > > > > > > > > > > > > > You need to do a bitwise OR of the assertions (MPI_MODE_NOSTORE | > > > > MPI_MODE_NOPRECEDE). > > > > > > > > In halo_1sided_store_win_alloc_shared.c, you are doing stores > > > > within > > > > the epoch, so MPI_MODE_NOSTORE looks like an incorrect assertion > > > > on > > > > the closing fence. > > > > > > > > Following the Fence epoch, you are reading from the left/right > > > > recv > > > > buffers. That also needs to be done within an RMA epoch, if you > > > > are > > > > reading non-local data. > > > > > > > > ~Jim. > > > > > > > > > > > > > > > > On Fri, Feb 21, 2014 at 6:07 AM, Rolf Rabenseifner < > > > > rabenseifner(a)hlrs.de > wrote: > > > > > > > > > > > > Dear member of the RMA group and especially the mpich developers, > > > > > > > > I have real problems with the new shared memory in MPI-3.0, > > > > i.e., the load/stores together with the RMA synchronization > > > > causes wrong execution results. > > > > > > > > The attached > > > > 1sided_halo_C_mpich_problems_rabenseifner.tar.gz or .zip > > > > contains > > > > > > > > - 1sided/halo_1sided_put_win_alloc.c > > > > > > > > The basis that works. It uses MPI_Put and MPI_Win_fence for > > > > duplex left/right halo communication. > > > > > > > > - 1sided/halo_1sided_store_win_alloc_shared.c > > > > > > > > This is the same, but a shared memory window is used and > > > > the MPU_Put is substituted by storing the data in the > > > > neighbors window. Same MPI_Win_fence with same assertions. > > > > > > > > This does not work, although I'm sure that my assertions are > > > > correct. > > > > > > > > Known possibilities: > > > > - I'm wrong and was not able to understand the assertions > > > > on MPI-3.0 p452:8-19. > > > > - I'm wrong because it is invalid to use the MPI_Win_fence > > > > together with the shared memory windows. > > > > - mpich has a bug. > > > > (The first two possibilities are the reason, why I use this > > > > Forum email list) > > > > > > > > - 1sided/halo_1sided_store_win_alloc_shared_w-a-cray.c > > > > > > > > This is a work-around-for Cray that works on our Cray > > > > and does not use MPI_MODE_NOPRECEDE and MPI_MODE_NOSUCCEED. > > > > It also runs on another mpich installation. > > > > > > > > - 1sided/halo_1sided_store_win_alloc_shared_pscw.c > > > > > > > > Here, MPI_Win_fence is substituted by Post-Start-Complete-Wait > > > > and it does not work for any assertions. > > > > > > > > Same possibilities as above. > > > > > > > > - 1sided/halo_1sided_store_win_alloc_shared_query.c > > > > - 1sided/halo_1sided_store_win_alloc_shared_query_w-a-cray.c > > > > > > > > Same as halo_1sided_store_win_alloc_shared.c > > > > but non-contigues windows are used. > > > > Same problems as above. > > > > > > > > - 1sided/halo_1sided_store_win_alloc_shared_othersync.c > > > > > > > > This version uses the synchronization according to > > > > #413 and it is tested and works on two platforms. > > > > > > > > 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-rma mailing list > > > > mpiwg-rma(a)lists.mpi-forum.org > > > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > > > > > > > > > > > _______________________________________________ > > > > mpiwg-rma mailing list > > > > mpiwg-rma(a)lists.mpi-forum.org > > > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > > > > > -- > > > 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-rma mailing list > > > mpiwg-rma(a)lists.mpi-forum.org > > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > > > > > > > > _______________________________________________ > > > mpiwg-rma mailing list > > > mpiwg-rma(a)lists.mpi-forum.org > > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > > > -- > > 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) > > > > -- > 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-rma mailing list > mpiwg-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma

1 0

#429 - Re: Problems with RMA synchronization in combination with load/store shared memory accesses
by Rolf Rabenseifner 19 May '14

19 May '14

Jim, I mnade #429. It is reported by me, but not owned, i.e., the RMA WG should take it. If the proposed text is okay for you (or with minor updates) then you can vote on it. I'll put it on the errata reading+voting list. It is the minimal text you can do that users and implementers see, that shared Memory Windows are Special for assertions. Therefore also the change-log, because implementors had and may also in future oversee this Problem. Best regards Rolf ----- Original Message ----- > From: "Rolf Rabenseifner" <rabenseifner(a)hlrs.de> > To: "MPI WG Remote Memory Access working group" <mpiwg-rma(a)lists.mpi-forum.org> > Sent: Monday, May 19, 2014 7:16:44 PM > Subject: Re: [mpiwg-rma] Problems with RMA synchronization in combination with load/store shared memory accesses > > Jim and RMA WG, > > There are now two questions: > > Jim asked: > > Question to WG: Do we need to update the fence assertions to better > > define interaction with local load/store accesses and remote > > stores? > > > > Rolf asked: > > Additionally, I would recommend that we add after MPI-3.0 p451:33 > > > > Note that in shared memory windows (allocated with > > MPI_WIN_ALLOCATE_SHARED), there is no difference > > between remote store accesses and local store accesses > > to the window. > > > > This would help to understand that "the local window > > was not updated by stores" does not mean "by local stores", > > see p452:1 and p452:9. > > For me, it is important to understand the meaning of the > current assertions if they are used in a shared memory window. > Therefore my proposal above as erratum to MPI-3.0. > > In MPI-3.1 and 4.0, you may want to add additional assertions. > > Your analysis below, will also show that mpich implements > Post-Start-Complete-Wait synchronization in a wrong way, > if there are no calls to RMA routines. > > Best regards > Rolf > > ----- Original Message ----- > > From: "Jim Dinan" <james.dinan(a)gmail.com> > > To: "MPI WG Remote Memory Access working group" > > <mpiwg-rma(a)lists.mpi-forum.org> > > Sent: Thursday, May 15, 2014 4:06:08 PM > > Subject: Re: [mpiwg-rma] Problems with RMA synchronization in > > combination with load/store shared memory accesses > > > > > > > > Rolf, > > > > > > Here is an attempt to simplify your example for discussion. Given > > a > > shared memory window, shr_mem_win, with buffer, shr_mem_buf: > > > > MPI_Win_fence(MPI_MODE_NOSTORE | MPI_MODE_NOPUT | > > MPI_MODE_NOPRECEDE > > | MPI_MODE_NOSUCCEED, shr_mem_win ); > > > > shr_mem_buf[...] = ...; > > > > MPI_Win_fence(MPI_MODE_NOPUT | MPI_MODE_NOPRECEDE | > > MPI_MODE_NOSUCCEED, shr_mem_win); > > > > > > Right now, Fence assertions don't say anything special about shared > > memory windows: > > > > > > Inline image 1 > > > > > > NOPRECEDE/SUCCEED are defined in terms of MPI RMA function calls, > > and > > do not cover load/store. Thus, Rolf's usage appears to be correct > > per the current text. In the MPICH fence implementation, > > src/mpid/ch3/src/ch3u_rma_sync.c:935 we have: > > > > if (!(assert & MPI_MODE_NOSUCCEED)) win_ptr->fence_issued = 1; > > > > Because of this check, we don't actually start an active target > > epoch > > on the first fence in the example above. On the second fence, we > > therefore don't perform the necessary synchronization, leading to > > incorrect output in Rolf's example. > > > > > > Question to WG: Do we need to update the fence assertions to better > > define interaction with local load/store accesses and remote > > stores? > > > > > > If not, then Rolf's code is correct and we need to modify the check > > above in MPICH to something like: > > > > > > if (!(assert & MPI_MODE_NOSUCCEED) || win_ptr->create_ flavor == > > MPI_WIN_FLAVOR_SHARED ) > > win_ptr->fence_issued = 1; > > > > > > ~Jim. > > > > > > > > > > On Tue, Apr 8, 2014 at 12:02 PM, Rolf Rabenseifner < > > rabenseifner(a)hlrs.de > wrote: > > > > > > Jim, > > > > I'm now sure, that mpich has a bug with assertions on shared memory > > windows. > > > > in the example, rcv_buf_left and rcv_buf_right are the windows. > > the only accesses to these rcv_buf_... are stores from remote > > and fully local loads. > > Both accesses are done in different epochs surrounded by > > MPI_Win_fence. > > > > According to your interpretation (which is really okay) > > all fences can use all possible assertions (!!!), > > except after the remote stores, MPI_MODE_NOSTORE cannot be used. > > > > I updated the example and mpich is executing it wrong. > > > > Please check it yourself on your installation: > > halo_1sided_store_win_alloc_shared_w-a-2-cray.c > > > > Without the assertions, all works: > > halo_1sided_store_win_alloc_shared_w-a-2NO-cray.c > > > > Could you verify that mpich has a bug? > > > > Additionally, I would recommend that we add after MPI-3.0 p451:33 > > > > Note that in shared memory windows (allocated with > > MPI_WIN_ALLOCATE_SHARED), there is no difference > > between remote store accesses and local store accesses > > to the window. > > > > This would help to understand that "the local window > > was not updated by stores" does not mean "by local stores", > > see p452:1 and p452:9. > > > > Is it a good idea? > > > > Best regards > > Rolf > > > > > > > > ----- Original Message ----- > > > From: "Jim Dinan" < james.dinan(a)gmail.com > > > > To: "MPI WG Remote Memory Access working group" < > > > mpiwg-rma(a)lists.mpi-forum.org > > > > Sent: Friday, March 21, 2014 8:14:22 PM > > > Subject: Re: [mpiwg-rma] Problems with RMA synchronization in > > > combination with load/store shared memory accesses > > > > > > > > > > > > > > Rolf, > > > > > > > > > This line is incorrect: MPI_Win_fence(MPI_MODE_NOSTORE + > > > MPI_MODE_NOPRECEDE, win_ rcv_buf _left ); > > > > > > > > > > > > > You need to do a bitwise OR of the assertions (MPI_MODE_NOSTORE | > > > MPI_MODE_NOPRECEDE). > > > > > > In halo_1sided_store_win_alloc_shared.c, you are doing stores > > > within > > > the epoch, so MPI_MODE_NOSTORE looks like an incorrect assertion > > > on > > > the closing fence. > > > > > > Following the Fence epoch, you are reading from the left/right > > > recv > > > buffers. That also needs to be done within an RMA epoch, if you > > > are > > > reading non-local data. > > > > > > ~Jim. > > > > > > > > > > > > On Fri, Feb 21, 2014 at 6:07 AM, Rolf Rabenseifner < > > > rabenseifner(a)hlrs.de > wrote: > > > > > > > > > Dear member of the RMA group and especially the mpich developers, > > > > > > I have real problems with the new shared memory in MPI-3.0, > > > i.e., the load/stores together with the RMA synchronization > > > causes wrong execution results. > > > > > > The attached > > > 1sided_halo_C_mpich_problems_rabenseifner.tar.gz or .zip > > > contains > > > > > > - 1sided/halo_1sided_put_win_alloc.c > > > > > > The basis that works. It uses MPI_Put and MPI_Win_fence for > > > duplex left/right halo communication. > > > > > > - 1sided/halo_1sided_store_win_alloc_shared.c > > > > > > This is the same, but a shared memory window is used and > > > the MPU_Put is substituted by storing the data in the > > > neighbors window. Same MPI_Win_fence with same assertions. > > > > > > This does not work, although I'm sure that my assertions are > > > correct. > > > > > > Known possibilities: > > > - I'm wrong and was not able to understand the assertions > > > on MPI-3.0 p452:8-19. > > > - I'm wrong because it is invalid to use the MPI_Win_fence > > > together with the shared memory windows. > > > - mpich has a bug. > > > (The first two possibilities are the reason, why I use this > > > Forum email list) > > > > > > - 1sided/halo_1sided_store_win_alloc_shared_w-a-cray.c > > > > > > This is a work-around-for Cray that works on our Cray > > > and does not use MPI_MODE_NOPRECEDE and MPI_MODE_NOSUCCEED. > > > It also runs on another mpich installation. > > > > > > - 1sided/halo_1sided_store_win_alloc_shared_pscw.c > > > > > > Here, MPI_Win_fence is substituted by Post-Start-Complete-Wait > > > and it does not work for any assertions. > > > > > > Same possibilities as above. > > > > > > - 1sided/halo_1sided_store_win_alloc_shared_query.c > > > - 1sided/halo_1sided_store_win_alloc_shared_query_w-a-cray.c > > > > > > Same as halo_1sided_store_win_alloc_shared.c > > > but non-contigues windows are used. > > > Same problems as above. > > > > > > - 1sided/halo_1sided_store_win_alloc_shared_othersync.c > > > > > > This version uses the synchronization according to > > > #413 and it is tested and works on two platforms. > > > > > > 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-rma mailing list > > > mpiwg-rma(a)lists.mpi-forum.org > > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > > > > > > > > _______________________________________________ > > > mpiwg-rma mailing list > > > mpiwg-rma(a)lists.mpi-forum.org > > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > > > -- > > 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-rma mailing list > > mpiwg-rma(a)lists.mpi-forum.org > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > > > > > _______________________________________________ > > mpiwg-rma mailing list > > mpiwg-rma(a)lists.mpi-forum.org > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > -- > 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) > -- 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: [mpiwg-rma] Problems with RMA synchronization in combination with load/store shared memory accesses
by Rolf Rabenseifner 19 May '14

19 May '14

Jim and RMA WG, There are now two questions: Jim asked: > Question to WG: Do we need to update the fence assertions to better > define interaction with local load/store accesses and remote stores? > Rolf asked: > Additionally, I would recommend that we add after MPI-3.0 p451:33 > > Note that in shared memory windows (allocated with > MPI_WIN_ALLOCATE_SHARED), there is no difference > between remote store accesses and local store accesses > to the window. > > This would help to understand that "the local window > was not updated by stores" does not mean "by local stores", > see p452:1 and p452:9. For me, it is important to understand the meaning of the current assertions if they are used in a shared memory window. Therefore my proposal above as erratum to MPI-3.0. In MPI-3.1 and 4.0, you may want to add additional assertions. Your analysis below, will also show that mpich implements Post-Start-Complete-Wait synchronization in a wrong way, if there are no calls to RMA routines. Best regards Rolf ----- Original Message ----- > From: "Jim Dinan" <james.dinan(a)gmail.com> > To: "MPI WG Remote Memory Access working group" <mpiwg-rma(a)lists.mpi-forum.org> > Sent: Thursday, May 15, 2014 4:06:08 PM > Subject: Re: [mpiwg-rma] Problems with RMA synchronization in combination with load/store shared memory accesses > > > > Rolf, > > > Here is an attempt to simplify your example for discussion. Given a > shared memory window, shr_mem_win, with buffer, shr_mem_buf: > > MPI_Win_fence(MPI_MODE_NOSTORE | MPI_MODE_NOPUT | MPI_MODE_NOPRECEDE > | MPI_MODE_NOSUCCEED, shr_mem_win ); > > shr_mem_buf[...] = ...; > > MPI_Win_fence(MPI_MODE_NOPUT | MPI_MODE_NOPRECEDE | > MPI_MODE_NOSUCCEED, shr_mem_win); > > > Right now, Fence assertions don't say anything special about shared > memory windows: > > > Inline image 1 > > > NOPRECEDE/SUCCEED are defined in terms of MPI RMA function calls, and > do not cover load/store. Thus, Rolf's usage appears to be correct > per the current text. In the MPICH fence implementation, > src/mpid/ch3/src/ch3u_rma_sync.c:935 we have: > > if (!(assert & MPI_MODE_NOSUCCEED)) win_ptr->fence_issued = 1; > > Because of this check, we don't actually start an active target epoch > on the first fence in the example above. On the second fence, we > therefore don't perform the necessary synchronization, leading to > incorrect output in Rolf's example. > > > Question to WG: Do we need to update the fence assertions to better > define interaction with local load/store accesses and remote stores? > > > If not, then Rolf's code is correct and we need to modify the check > above in MPICH to something like: > > > if (!(assert & MPI_MODE_NOSUCCEED) || win_ptr->create_ flavor == > MPI_WIN_FLAVOR_SHARED ) > win_ptr->fence_issued = 1; > > > ~Jim. > > > > > On Tue, Apr 8, 2014 at 12:02 PM, Rolf Rabenseifner < > rabenseifner(a)hlrs.de > wrote: > > > Jim, > > I'm now sure, that mpich has a bug with assertions on shared memory > windows. > > in the example, rcv_buf_left and rcv_buf_right are the windows. > the only accesses to these rcv_buf_... are stores from remote > and fully local loads. > Both accesses are done in different epochs surrounded by > MPI_Win_fence. > > According to your interpretation (which is really okay) > all fences can use all possible assertions (!!!), > except after the remote stores, MPI_MODE_NOSTORE cannot be used. > > I updated the example and mpich is executing it wrong. > > Please check it yourself on your installation: > halo_1sided_store_win_alloc_shared_w-a-2-cray.c > > Without the assertions, all works: > halo_1sided_store_win_alloc_shared_w-a-2NO-cray.c > > Could you verify that mpich has a bug? > > Additionally, I would recommend that we add after MPI-3.0 p451:33 > > Note that in shared memory windows (allocated with > MPI_WIN_ALLOCATE_SHARED), there is no difference > between remote store accesses and local store accesses > to the window. > > This would help to understand that "the local window > was not updated by stores" does not mean "by local stores", > see p452:1 and p452:9. > > Is it a good idea? > > Best regards > Rolf > > > > ----- Original Message ----- > > From: "Jim Dinan" < james.dinan(a)gmail.com > > > To: "MPI WG Remote Memory Access working group" < > > mpiwg-rma(a)lists.mpi-forum.org > > > Sent: Friday, March 21, 2014 8:14:22 PM > > Subject: Re: [mpiwg-rma] Problems with RMA synchronization in > > combination with load/store shared memory accesses > > > > > > > > > Rolf, > > > > > > This line is incorrect: MPI_Win_fence(MPI_MODE_NOSTORE + > > MPI_MODE_NOPRECEDE, win_ rcv_buf _left ); > > > > > > > > You need to do a bitwise OR of the assertions (MPI_MODE_NOSTORE | > > MPI_MODE_NOPRECEDE). > > > > In halo_1sided_store_win_alloc_shared.c, you are doing stores > > within > > the epoch, so MPI_MODE_NOSTORE looks like an incorrect assertion on > > the closing fence. > > > > Following the Fence epoch, you are reading from the left/right recv > > buffers. That also needs to be done within an RMA epoch, if you > > are > > reading non-local data. > > > > ~Jim. > > > > > > > > On Fri, Feb 21, 2014 at 6:07 AM, Rolf Rabenseifner < > > rabenseifner(a)hlrs.de > wrote: > > > > > > Dear member of the RMA group and especially the mpich developers, > > > > I have real problems with the new shared memory in MPI-3.0, > > i.e., the load/stores together with the RMA synchronization > > causes wrong execution results. > > > > The attached > > 1sided_halo_C_mpich_problems_rabenseifner.tar.gz or .zip > > contains > > > > - 1sided/halo_1sided_put_win_alloc.c > > > > The basis that works. It uses MPI_Put and MPI_Win_fence for > > duplex left/right halo communication. > > > > - 1sided/halo_1sided_store_win_alloc_shared.c > > > > This is the same, but a shared memory window is used and > > the MPU_Put is substituted by storing the data in the > > neighbors window. Same MPI_Win_fence with same assertions. > > > > This does not work, although I'm sure that my assertions are > > correct. > > > > Known possibilities: > > - I'm wrong and was not able to understand the assertions > > on MPI-3.0 p452:8-19. > > - I'm wrong because it is invalid to use the MPI_Win_fence > > together with the shared memory windows. > > - mpich has a bug. > > (The first two possibilities are the reason, why I use this > > Forum email list) > > > > - 1sided/halo_1sided_store_win_alloc_shared_w-a-cray.c > > > > This is a work-around-for Cray that works on our Cray > > and does not use MPI_MODE_NOPRECEDE and MPI_MODE_NOSUCCEED. > > It also runs on another mpich installation. > > > > - 1sided/halo_1sided_store_win_alloc_shared_pscw.c > > > > Here, MPI_Win_fence is substituted by Post-Start-Complete-Wait > > and it does not work for any assertions. > > > > Same possibilities as above. > > > > - 1sided/halo_1sided_store_win_alloc_shared_query.c > > - 1sided/halo_1sided_store_win_alloc_shared_query_w-a-cray.c > > > > Same as halo_1sided_store_win_alloc_shared.c > > but non-contigues windows are used. > > Same problems as above. > > > > - 1sided/halo_1sided_store_win_alloc_shared_othersync.c > > > > This version uses the synchronization according to > > #413 and it is tested and works on two platforms. > > > > 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-rma mailing list > > mpiwg-rma(a)lists.mpi-forum.org > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > > > > > _______________________________________________ > > mpiwg-rma mailing list > > mpiwg-rma(a)lists.mpi-forum.org > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > -- > 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-rma mailing list > mpiwg-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > > _______________________________________________ > mpiwg-rma mailing list > mpiwg-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma -- 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: [mpiwg-rma] Problems with RMA synchronization in combination with load/store shared memory accesses
by Jim Dinan 15 May '14

15 May '14

Rolf, Here is an attempt to simplify your example for discussion. Given a shared memory window, shr_mem_win, with buffer, shr_mem_buf: MPI_Win_fence(MPI_MODE_NOSTORE | MPI_MODE_NOPUT | MPI_MODE_NOPRECEDE | MPI_MODE_NOSUCCEED, shr_mem_win ); shr_mem_buf[...] = ...; MPI_Win_fence(MPI_MODE_NOPUT | MPI_MODE_NOPRECEDE | MPI_MODE_NOSUCCEED, shr_mem_win); Right now, Fence assertions don't say anything special about shared memory windows: [image: Inline image 1] NOPRECEDE/SUCCEED are defined in terms of MPI RMA function calls, and do not cover load/store. Thus, Rolf's usage appears to be correct per the current text. In the MPICH fence implementation, src/mpid/ch3/src/ch3u_rma_sync.c:935 we have: if (!(assert & MPI_MODE_NOSUCCEED)) win_ptr->fence_issued = 1; Because of this check, we don't actually start an active target epoch on the first fence in the example above. On the second fence, we therefore don't perform the necessary synchronization, leading to incorrect output in Rolf's example. Question to WG: Do we need to update the fence assertions to better define interaction with local load/store accesses and remote stores? If not, then Rolf's code is correct and we need to modify the check above in MPICH to something like: if (!(assert & MPI_MODE_NOSUCCEED) || win_ptr->create_flavor == MPI_WIN_FLAVOR_SHARED) win_ptr->fence_issued = 1; ~Jim. On Tue, Apr 8, 2014 at 12:02 PM, Rolf Rabenseifner <rabenseifner(a)hlrs.de>wrote: > Jim, > > I'm now sure, that mpich has a bug with assertions on shared memory > windows. > > in the example, rcv_buf_left and rcv_buf_right are the windows. > the only accesses to these rcv_buf_... are stores from remote > and fully local loads. > Both accesses are done in different epochs surrounded by MPI_Win_fence. > > According to your interpretation (which is really okay) > all fences can use all possible assertions (!!!), > except after the remote stores, MPI_MODE_NOSTORE cannot be used. > > I updated the example and mpich is executing it wrong. > > Please check it yourself on your installation: > halo_1sided_store_win_alloc_shared_w-a-2-cray.c > > Without the assertions, all works: > halo_1sided_store_win_alloc_shared_w-a-2NO-cray.c > > Could you verify that mpich has a bug? > > Additionally, I would recommend that we add after MPI-3.0 p451:33 > > Note that in shared memory windows (allocated with > MPI_WIN_ALLOCATE_SHARED), there is no difference > between remote store accesses and local store accesses > to the window. > > This would help to understand that "the local window > was not updated by stores" does not mean "by local stores", > see p452:1 and p452:9. > > Is it a good idea? > > Best regards > Rolf > > > ----- Original Message ----- > > From: "Jim Dinan" <james.dinan(a)gmail.com> > > To: "MPI WG Remote Memory Access working group" < > mpiwg-rma(a)lists.mpi-forum.org> > > Sent: Friday, March 21, 2014 8:14:22 PM > > Subject: Re: [mpiwg-rma] Problems with RMA synchronization in > combination with load/store shared memory accesses > > > > > > > > Rolf, > > > > > > This line is incorrect: MPI_Win_fence(MPI_MODE_NOSTORE + > > MPI_MODE_NOPRECEDE, win_ rcv_buf _left ); > > > > > > You need to do a bitwise OR of the assertions (MPI_MODE_NOSTORE | > > MPI_MODE_NOPRECEDE). > > > > In halo_1sided_store_win_alloc_shared.c, you are doing stores within > > the epoch, so MPI_MODE_NOSTORE looks like an incorrect assertion on > > the closing fence. > > > > Following the Fence epoch, you are reading from the left/right recv > > buffers. That also needs to be done within an RMA epoch, if you are > > reading non-local data. > > > > ~Jim. > > > > > > > > On Fri, Feb 21, 2014 at 6:07 AM, Rolf Rabenseifner < > > rabenseifner(a)hlrs.de > wrote: > > > > > > Dear member of the RMA group and especially the mpich developers, > > > > I have real problems with the new shared memory in MPI-3.0, > > i.e., the load/stores together with the RMA synchronization > > causes wrong execution results. > > > > The attached > > 1sided_halo_C_mpich_problems_rabenseifner.tar.gz or .zip > > contains > > > > - 1sided/halo_1sided_put_win_alloc.c > > > > The basis that works. It uses MPI_Put and MPI_Win_fence for > > duplex left/right halo communication. > > > > - 1sided/halo_1sided_store_win_alloc_shared.c > > > > This is the same, but a shared memory window is used and > > the MPU_Put is substituted by storing the data in the > > neighbors window. Same MPI_Win_fence with same assertions. > > > > This does not work, although I'm sure that my assertions are > > correct. > > > > Known possibilities: > > - I'm wrong and was not able to understand the assertions > > on MPI-3.0 p452:8-19. > > - I'm wrong because it is invalid to use the MPI_Win_fence > > together with the shared memory windows. > > - mpich has a bug. > > (The first two possibilities are the reason, why I use this > > Forum email list) > > > > - 1sided/halo_1sided_store_win_alloc_shared_w-a-cray.c > > > > This is a work-around-for Cray that works on our Cray > > and does not use MPI_MODE_NOPRECEDE and MPI_MODE_NOSUCCEED. > > It also runs on another mpich installation. > > > > - 1sided/halo_1sided_store_win_alloc_shared_pscw.c > > > > Here, MPI_Win_fence is substituted by Post-Start-Complete-Wait > > and it does not work for any assertions. > > > > Same possibilities as above. > > > > - 1sided/halo_1sided_store_win_alloc_shared_query.c > > - 1sided/halo_1sided_store_win_alloc_shared_query_w-a-cray.c > > > > Same as halo_1sided_store_win_alloc_shared.c > > but non-contigues windows are used. > > Same problems as above. > > > > - 1sided/halo_1sided_store_win_alloc_shared_othersync.c > > > > This version uses the synchronization according to > > #413 and it is tested and works on two platforms. > > > > 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-rma mailing list > > mpiwg-rma(a)lists.mpi-forum.org > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > > > > > _______________________________________________ > > mpiwg-rma mailing list > > mpiwg-rma(a)lists.mpi-forum.org > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > -- > 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-rma mailing list > mpiwg-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma >

1 0

Fwd: MPI Forum / RMA WG
by Rajeev Thakur 06 May '14

06 May '14

How much working group or plenary time do we need at the June meeting? Rajeev Begin forwarded message: > From: "Schulz, Martin" <schulz6(a)llnl.gov> > Subject: MPI Forum / RMA WG > Date: May 6, 2014 2:29:47 AM CDT > To: William Gropp <wgropp(a)illinois.edu>, Rajeev Thakur <thakur(a)mcs.anl.gov> > > Hi Bill and Rajeev, > > Do you have any time requests (WG or plenary) for the June forum meeting? > > Thanks! > > Martin >

1 0