August 2012 - Mpich2-core - mailman3.cels.anl.gov

automake-1.12.3
by Pavan Balaji 31 Aug '12

31 Aug '12

I've removed the pgcc patch that we maintain internally as it's been absorbed into automake-1.12.3. I've added a dependency on that version as well. -- Pavan -- Pavan Balaji http://www.mcs.anl.gov/~balaji

3 4

Fwd: [mcs-systems #127883] more mpich2 trac spam
by Dave Goodell 29 Aug '12

29 Aug '12

FYI, email-based registration has been activated for the MPICH2 trac page, and anonymous ticket creation has been disabled. -Dave Begin forwarded message: > From: Ken Raffenetti <raffenet(a)mcs.anl.gov> > Subject: Re: [mcs-systems #127883] more mpich2 trac spam > Date: August 28, 2012 1:34:10 PM CDT > To: Dave Goodell <goodell(a)mcs.anl.gov> > Cc: systems(a)mcs.anl.gov > > Dave, > > This should be all set. I removed the ticket create/edit perms from > anonymous users. Now only authenticated should work. > > Ken Raffenetti > Systems Administration Associate > MCS Division - Argonne National Laboratory > > On 08/28/2012 01:16 PM, Dave Goodell wrote: >> How about this instead: >> >> WARNING! You need to register a trac account with an email address (or use an existing MCS SVN login) before you can create or edit tickets. >> IMPORTANT NOTE: For queries on problems that you are facing with your >> MPICH2 installation or usage, please send it to the >> mpich-discuss(a)mcs.anl.gov mailing list. You need to subscribe to this >> mailing list here before sending an email. If you have found a bug in >> MPICH2 and want to report it, then fill out the form below to open the >> ticket. >> >> -Dave >> >> On Aug 28, 2012, at 1:10 PM CDT, Ken Raffenetti wrote: >> >>> Dave, >>> >>> Sure. Do you want to make updates to this text on the New Ticket page? >>> I have to edit that text on the server. If you give me a new version I >>> can put both in place at the same time. >>> >>> WARNING! You need to enter your email address here before you can create >>> or edit tickets. Otherwise the ticket will get treated as spam. >>> IMPORTANT NOTE: For queries on problems that you are facing with your >>> MPICH2 installation or usage, please send it to the >>> mpich-discuss(a)mcs.anl.gov mailing list. You need to subscribe to this >>> mailing list here before sending an email. If you have found a bug in >>> MPICH2 and want to report it, then fill out the form below to open the >>> ticket. >>> >>> Ken Raffenetti >>> Systems Administration Associate >>> MCS Division - Argonne National Laboratory >>> >>> On 08/28/2012 01:07 PM, Dave Goodell wrote: >>>> Sorry for the delayed response, I was on vacation. This plugin seems to work well for me. Can we enable it for the mpich2 trac page as well? >>>> >>>> -Dave >>>> >>>> On Aug 17, 2012, at 2:38 PM CDT, Ken Raffenetti wrote: >>>> >>>>> Dave, >>>>> >>>>> I managed to get a newer version of the AccountManager plugin installed at >>>>> https://trac.mcs.anl.gov/projects/roberts-test. It's got email confirmation >>>>> working and without the bug I saw earlier. When you get a chance, check it out. >>>>> >>>>> Ken >>>>> >>>>> On Tue Aug 14 13:22:13 2012, raffenet wrote: >>>>>> Okay, I'll take a crack at that. When I first enabled it, it had a >>>>>> bug >>>>>> that made it unclear why you weren't able to access the site (just >>>>>> displaying an error). Supposedly a newer version fixes this. >>>>>> >>>>>> Ken Raffenetti >>>>>> Systems Administration Associate >>>>>> MCS Division - Argonne National Laboratory >>>>>> >>>>>> On 08/14/2012 12:26 PM, Dave Goodell wrote: >>>>>>> I just played with it and I think we will need the email >>>>>> confirmation feature, otherwise the spammers won't really be slowed >>>>>> down. >>>>>>> >>>>>>> -Dave >>>>>>> >>>>>>> On Aug 14, 2012, at 9:36 AM CDT, Ken Raffenetti wrote: >>>>>>> >>>>>>>> Dave, >>>>>>>> >>>>>>>> I have setup the user registration plugin in a test area for you to >>>>>> evaluate. >>>>>>>> Please check out http://trac.mcs.anl.gov/projects/roberts-test. I >>>>>> gave your >>>>>>>> user TRAC_ADMIN privileges. Notice that when not logged in, a >>>>>> "Register" link >>>>>>>> is visible in the top-level menu. Once a user is registered and >>>>>> logged in, they >>>>>>>> have 'authenticated' level privileges, which we would require for >>>>>> TICKET_CREATE >>>>>>>> and the like. >>>>>>>> >>>>>>>> For now I have email confirmation turned off during registration, >>>>>> but we can >>>>>>>> investigate using it if you guys want. Let me know what you think. >>>>>>>> >>>>>>>> Ken >>>>>>>> >>>>>>>> On Fri Aug 10 11:05:13 2012, goodell wrote: >>>>>>>>> I think that would be worth pursuing. >>>>>>>>> >>>>>>>>> -Dave >>>>>>>>> >>>>>>>>> On Aug 10, 2012, at 9:43 AM CDT, Ken Raffenetti wrote: >>>>>>>>> >>>>>>>>>> I believe there is a user registration plugin for trac we could >>>>>>>>> enable >>>>>>>>>> to require users to register and login before submitting a >>>>>> ticket. >>>>>>>>> That >>>>>>>>>> might help cut down on spam. >>>>>>>>>> >>>>>>>>>> Ken Raffenetti >>>>>>>>>> Systems Administration Associate >>>>>>>>>> MCS Division - Argonne National Laboratory >>>>>>>>>> >>>>>>>>>> On 08/10/2012 09:05 AM, Dave Goodell wrote: >>>>>>>>>>> Systems, >>>>>>>>>>> >>>>>>>>>>> We got around 20 spam comments on our trac page overnight, an >>>>>>>>> example of which is forwarded below. It's very tedious and a big >>>>>>>>> waste of time to go in and delete each of these comments. Is there >>>>>>>>> anything that can be done to improve the spam filtering? >>>>>>>>>>> >>>>>>>>>>> Note, I have deleted all of the spam comments, so you'll have to >>>>>>>>> look at the mpich2-maint archives (or whatever list is being used >>>>>>>>> for mail distribution) for more details. >>>>>>>>>>> >>>>>>>>>>> -Dave >>>>>>>>>>> >>>>>>>>>>> Begin forwarded message: >>>>>>>>>>> >>>>>>>>>>>> From: "mpich2" <noreply(a)mcs.anl.gov> >>>>>>>>>>>> Subject: Re: [mpich2-maint] #1648: error invalid no_port >>>>>>>>>>>> Date: August 9, 2012 9:03:33 PM CDT >>>>>>>>>>>> To: undisclosed-recipients:; >>>>>>>>>>>> Reply-To: noreply(a)mcs.anl.gov >>>>>>>>>>>> >>>>>>>>>>>> #1648: error invalid no_port >>>>>>>>>>>> >>>>>>>>> >>>>>> -----------------------------------+---------------------------------------- >>>>>>>>>>>> Reporter: huki_imoet@… | Type: bug >>>>>>>>>>>> Status: closed | Priority: major >>>>>>>>>>>> Milestone: | Component: mpich2 >>>>>>>>>>>> Resolution: wontfix | Keywords: >>>>>>>>>>>> >>>>>>>>> >>>>>> -----------------------------------+---------------------------------------- >>>>>>>>>>>> >>>>>>>>>>>> Comment(by xiaoliuquan@…): >>>>>>>>>>>> >>>>>>>>>>>> Thank you for [http://www.ukchanelbags.co.uk/ chanel bags] >>>>>> another >>>>>>>>> great >>>>>>>>>>>> article. Where else could anyone get that kind of information >>>>>> in >>>>>>>>> such a >>>>>>>>>>>> perfect way of writing? >>>>>>>>>>>> >>>>>>>>>>>> -- >>>>>>>>>>>> Ticket URL: >>>>>>>>> <http://trac.mcs.anl.gov/projects/mpich2/ticket/1648#comment:3> >>>>>>>>>>>> mpich2 <http://www.mcs.anl.gov/research/projects/mpich2> >>>>>>>>>>>> MPICH2 is a high-performance and widely portable implementation >>>>>> of >>>>>>>>> MPI >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>> >>>>> >>>> >>

4 4

Disp_unit in dynamic windows should not be signed
by Jim Dinan 28 Aug '12

28 Aug '12

Hi All, The MPI spec states that the disp_unit argument to RMA calls must be a non-negative integer (of type MPI_Aint). For dynamic windows, the displacement is the address relative to MPI_BOTTOM (i.e. NULL), which can be in the negative range when interpreted as a signed integer. We currently implement MPI_Aint as a signed integer. Is this correct? If so, we need to make changes to the RMA chapter to account for disp_units on dynamic windows (e.g. to allow us to ignore signedness of disp_unit for windows of this flavor). ~Jim.

2 6

Fwd: New Kepler GPU Greases the Wheels for MPI Applications
by Rajeev Thakur 24 Aug '12

24 Aug '12

From today's hpcwire http://www.hpcwire.com/hpcwire/2012-08-23/new_kepler_gpu_greases_the_wheels…

1 0

MPID_Comm_fns
by Jim Dinan 20 Aug '12

20 Aug '12

Hi All, While reading code in the MPICH trunk, I came across (mpiimpl.h): typedef struct MPID_CommOps { int (*split_type)(MPID_Comm *, int, int, MPID_Info *, MPID_Comm **); } MPID_CommOps; extern struct MPID_CommOps *MPID_Comm_fns; /* Communicator creation functions */ Is MPI_Comm_split_type the only communicator creation routine that the device can override? If so, is there another vtable that this can fit in? Right now, this code looks like a special/isolated case. There also appears to be a naming collision with MPIDI_Comm_ops_t in CH3, which is used for the pt-to-pt functions in the VC. ~Jim.

2 6

Fwd: [mpich-discuss] Spawn MPI-Process
by Rajeev Thakur 20 Aug '12

20 Aug '12

Why is universe_size 1? I printed out universe size on my laptop and on vanquish, and it is 1 in both cases. Rajeev Begin forwarded message: > From: Silvan Brändli <silvan.braendli(a)tu-harburg.de> > Date: August 20, 2012 4:39:52 AM CDT > To: mpich-discuss(a)mcs.anl.gov > Subject: [mpich-discuss] Spawn MPI-Process > Reply-To: mpich-discuss(a)mcs.anl.gov > > Dear all > > I try to start an application using the MPI_Comm_spawn command. However, when I execute my program I get a segmentation fault. I guess, this might be because universe_size is one. As I couldn't find helpful information on setting universe_size I hope some of you could help me. > > Find the source code below or in the attached files. I compile the code using > mpicxx main.cpp -o main > mpicxx hi.cpp -o hi > > and execute the case with > > mpiexec -prepend-rank -np 1 ./main > > The output is > > $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ > > [0] world_size 1 > [0] universe size 1 > [0] try to spawn > > ===================================================================================== > = BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES > = EXIT CODE: 11 > = CLEANING UP REMAINING PROCESSES > = YOU CAN IGNORE THE BELOW CLEANUP MESSAGES > ===================================================================================== > APPLICATION TERMINATED WITH THE EXIT STRING: Segmentation fault (signal 11) > > $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ > > Best regards > Silvan > > $$$$==main.cpp==$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ > > #include <sstream> > #include <mpi.h>; > > using std::string; > > int main(int argc, char *argv[]) > { > int myrank; > if (MPI_Init(&argc,&argv)!=MPI_SUCCESS) > { > printf("MPI_Init failed"); > } > MPI_Comm_rank(MPI_COMM_WORLD, &myrank); > > int world_size, universe_size, *universe_sizep, flag; > MPI_Comm_size(MPI_COMM_WORLD, &world_size); > printf("world_size %d\n",world_size); > > MPI_Attr_get(MPI_COMM_WORLD, MPI_UNIVERSE_SIZE, > &universe_sizep, &flag); > > if (!flag) > { > printf("This MPI does not support UNIVERSE_SIZE. How many processes total?"); > scanf("%d", &universe_size); > } else universe_size = *universe_sizep; > printf("universe size %d\n",universe_size); > > char* hiargv[1]; > char* m0=" "; > hiargv[0]=m0; > > MPI_Comm childComm; > int spawnerror; > printf("try to spawn\n"); > MPI_Comm_spawn("./hi",hiargv, universe_size-1, MPI_INFO_NULL, myrank, MPI_COMM_SELF, &childComm, &spawnerror); > printf("after spawn\n"); > MPI_Finalize(); > return 0; > } > > $$$$==hi.cpp==$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ > > #include <mpi.h>; > int main(int argc, char** argv) { > MPI_Init(NULL, NULL); > > int world_size; > MPI_Comm_size(MPI_COMM_WORLD, &world_size); > > int world_rank; > MPI_Comm_rank(MPI_COMM_WORLD, &world_rank); > > printf("Hello world from, rank %d" > " out of %d processors\n", > world_rank, world_size); > > MPI_Finalize(); > return 0; > } > > $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ > -- > Dipl.-Ing. Silvan Brändli > Numerische Strukturanalyse mit Anwendungen in der Schiffstechnik (M-10) > > Technische Universität Hamburg-Harburg > Schwarzenbergstraße 95c > 21073 Hamburg > > Tel. : +49 (0)40 42878 - 6187 > Fax. : +49 (0)40 42878 - 6090 > e-mail: silvan.braendli(a)tu-harburg.de > www : http://www.tuhh.de/skf > > _______________________________________________ > mpich-discuss mailing list mpich-discuss(a)mcs.anl.gov > To manage subscription options or unsubscribe: > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss

2 3

Re: [mpich2-core] [mpich-ibm] pamid code contribution - based on mpich2 1.5b2
by William Gropp 19 Aug '12

19 Aug '12

So we're not going to fix the other blockers first, like #1679 or #1659 ? Bill William Gropp Director, Parallel Computing Institute Deputy Director for Research Institute for Advanced Computing Applications and Technologies Paul and Cynthia Saylor Professor of Computer Science University of Illinois Urbana-Champaign On Aug 20, 2012, at 10:37 AM, Pavan Balaji wrote: > Jeff -- this is the only thing pending for 1.5rc2. As soon as we integrate the patch, 1.5rc2 will be released.

2 2

cutting the mpich2-1.5rc1 release today
by Dave Goodell 14 Aug '12

14 Aug '12

The nightly tests are clean enough right now, and all of the critical features are in. I'm on vacation starting Thursday, so I need to cut this either today or tomorrow, or someone else will need to do it while I'm gone. The rough plan is to cut the final release at the end of the month, after I come back on the 28th. I am planning to hold off on cutting a branch for 1.5, but if someone wants to make more disruptive changes in the near term then I'm happy to cut one. If anyone has an objection to this schedule, let me know *soon*. -Dave

2 2

Re: [mpich2-core] [mpich2-maint] #1158: Fortran-Binding for MPI_Comm_get_attr(MPI_APPNUM) broken on bigendian/64 bits
by Rajeev Thakur 14 Aug '12

14 Aug '12

Bill is probably referring to the changes made by IBM for the BG/P. Rajeev On Aug 14, 2012, at 12:55 PM, mpich2 wrote: > #1158: Fortran-Binding for MPI_Comm_get_attr(MPI_APPNUM) broken on bigendian/64 > bits > ------------------------------------+--------------------------------------- > Reporter: nicolai.stange@… | Owner: gropp > Type: bug | Status: reopened > Priority: major | Milestone: mpich2-1.5 > Component: mpich2 | Resolution: > Keywords: | > ------------------------------------+--------------------------------------- > > Comment(by gropp): > > Well before r6346 - the current code confuses int and MPI_Fint and > MPIR_Pint and MPIR_FAint, since the integer values are for the Fortran > 77/9x interface respectively. The code is demonstrably incorrect if > MPI_Fint is not the same as int, which is valid, we can handle, and some > users have requested. A much earlier change removed the distinction > between Fortran and C attributes in the code and introduce the problems > that r6346 fixed. I'll look into designing and documenting a fix that > will handle the various cases (there is an attribute test that Jeff > Squyres wrote that checks for the various interlanguage cases). > > -- > Ticket URL: <http://trac.mcs.anl.gov/projects/mpich2/ticket/1158#comment:16> > mpich2 <http://www.mcs.anl.gov/research/projects/mpich2> > MPICH2 is a high-performance and widely portable implementation of MPI

1 0

Re: [mpich2-core] [mcs-systems #127883] more mpich2 trac spam
by Ken Raffenetti 10 Aug '12

10 Aug '12

I believe there is a user registration plugin for trac we could enable to require users to register and login before submitting a ticket. That might help cut down on spam. Ken Raffenetti Systems Administration Associate MCS Division - Argonne National Laboratory On 08/10/2012 09:05 AM, Dave Goodell wrote: > Systems, > > We got around 20 spam comments on our trac page overnight, an example of which is forwarded below. It's very tedious and a big waste of time to go in and delete each of these comments. Is there anything that can be done to improve the spam filtering? > > Note, I have deleted all of the spam comments, so you'll have to look at the mpich2-maint archives (or whatever list is being used for mail distribution) for more details. > > -Dave > > Begin forwarded message: > >> From: "mpich2" <noreply(a)mcs.anl.gov> >> Subject: Re: [mpich2-maint] #1648: error invalid no_port >> Date: August 9, 2012 9:03:33 PM CDT >> To: undisclosed-recipients:; >> Reply-To: noreply(a)mcs.anl.gov >> >> #1648: error invalid no_port >> -----------------------------------+---------------------------------------- >> Reporter: huki_imoet@… | Type: bug >> Status: closed | Priority: major >> Milestone: | Component: mpich2 >> Resolution: wontfix | Keywords: >> -----------------------------------+---------------------------------------- >> >> Comment(by xiaoliuquan@…): >> >> Thank you for [http://www.ukchanelbags.co.uk/ chanel bags] another great >> article. Where else could anyone get that kind of information in such a >> perfect way of writing? >> >> -- >> Ticket URL: <http://trac.mcs.anl.gov/projects/mpich2/ticket/1648#comment:3> >> mpich2 <http://www.mcs.anl.gov/research/projects/mpich2> >> MPICH2 is a high-performance and widely portable implementation of MPI > >

2 1