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

Re: [mpich-discuss] Affinity with MPICH_ASYNC_PROGRESS
by Jed Brown 25 Feb '13

25 Feb '13

On Mon, Feb 25, 2013 at 7:55 AM, Pavan Balaji <balaji(a)mcs.anl.gov> wrote: > On 02/23/2013 06:49 PM US Central Time, Jeff Hammond wrote: > > My understanding, having looked at the code for this, is that no > > thread binding is set. The OS will hopefully do the right thing when > > the cores are undersubscribed, but I see no thread affinity code in > > the MPICH source for the comm threads. > > This is incorrect. Hydra allows you to bind a process to a cpuset. If > you are using an affinity thread, make sure your cpuset has at least 2 > cores in it. It sounds like you are binding a process. Suppose hydra binds my process to cpuset 0xf and my process pthread_creates a total of three threads. If I bind those threads to 0x1, 0x2, 0x4, we still have the MPICH_ASYNC_PROGRESS running with affinity 0xf, and as far as I know, there is no way to bind it to 0x8. It would be interesting to be able to specify this, both to place the progress threads close to NICs and to make sure that progress threads do not disrupt application threads (to reduce jitter and possibly improve performance).

1 0

Re: [mpich-discuss] mpiexec: ignore exit codes
by Silvan Brändli 25 Feb '13

25 Feb '13

Thanks a lot! It works. And I can only support your suggestion to introduce a command line option... Silvan On 22.02.2013 19:01, Yauheni Zelenko wrote: > Hi, Silvan! > > I had similar problem. > > As workaround you need to change next line in src/pm/hydra/pm/pmiserv/pmiserv_cb.c from > > if (!WIFEXITED(proxy->exit_status[i]) || WEXITSTATUS(proxy->exit_status[i])) { > > to > > if (!WIFEXITED(proxy->exit_status[i]) /* || WEXITSTATUS(proxy->exit_status[i]) */) { > > I already suggested to introduce Hydra command line option which will allow to ignore non-zero exit codes. > > Eugene. > ________________________________________ > From: discuss-bounces(a)mpich.org [discuss-bounces(a)mpich.org] On Behalf Of Silvan Brï¿½ndli [silvan.braendli(a)tuhh.de] > Sent: Friday, February 22, 2013 5:46 AM > To: discuss(a)mpich.org > Subject: [mpich-discuss] mpiexec: ignore exit codes > > Dear all, > > one of my spawned processes returns Exit Code 1 when finishing, causing > mpiexec to abort (see below). Despite that I would like to continue > running my programs. Is there a way to make mpiexec ignore the exit codes? > > Best regards > Silvan Brï¿½ndli > > =================================================================================== > = BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES > = EXIT CODE: 1 > = CLEANING UP REMAINING PROCESSES > = YOU CAN IGNORE THE BELOW CLEANUP MESSAGES > =================================================================================== > [proxy:0:0@leibniz] HYD_pmcd_pmip_control_cmd_cb > (./pm/pmiserv/pmip_cb.c:883): assert (!closed) failed > [proxy:0:0@leibniz] HYDT_dmxu_poll_wait_for_event > (./tools/demux/demux_poll.c:77): callback returned error status > [proxy:0:0@leibniz] main (./pm/pmiserv/pmip.c:210): demux engine error > waiting for event > [proxy:2:0@leibniz] HYD_pmcd_pmip_control_cmd_cb > (./pm/pmiserv/pmip_cb.c:883): assert (!closed) failed > [proxy:2:0@leibniz] HYDT_dmxu_poll_wait_for_event > (./tools/demux/demux_poll.c:77): callback returned error status > [proxy:2:0@leibniz] main (./pm/pmiserv/pmip.c:210): demux engine error > waiting for event > [mpiexec@leibniz] HYDT_bscu_wait_for_completion > (./tools/bootstrap/utils/bscu_wait.c:76): one of the processes > terminated badly; aborting > [mpiexec@leibniz] HYDT_bsci_wait_for_completion > (./tools/bootstrap/src/bsci_wait.c:23): launcher returned error waiting > for completion > [mpiexec@leibniz] HYD_pmci_wait_for_completion > (./pm/pmiserv/pmiserv_pmci.c:216): launcher returned error waiting for > completion > [mpiexec@leibniz] main (./ui/mpich/mpiexec.c:325): process manager error > waiting for completion > > -- > Dipl.-Ing. Silvan Brï¿½ndli > Numerische Strukturanalyse mit Anwendungen in der Schiffstechnik (M-10) > > Technische Universitï¿½t Hamburg-Harburg > Schwarzenbergstraï¿½e 95c > 21073 Hamburg > > Tel. : +49 (0)40 42878 - 6187 > Fax. : +49 (0)40 42878 - 6090 > e-mail: silvan.braendli(a)tuhh.de > www : http://www.tuhh.de/skf > > 5th GACM Colloquium on Computational Mechanics > http://www.tu-harburg.de/gacm2013 > _______________________________________________ > 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 > -- Dipl.-Ing. Silvan Brï¿½ndli Numerische Strukturanalyse mit Anwendungen in der Schiffstechnik (M-10) Technische Universitï¿½t Hamburg-Harburg Schwarzenbergstraï¿½e 95c 21073 Hamburg Tel. : +49 (0)40 42878 - 6187 Fax. : +49 (0)40 42878 - 6090 e-mail: silvan.braendli(a)tuhh.de www : http://www.tuhh.de/skf 5th GACM Colloquium on Computational Mechanics http://www.tu-harburg.de/gacm2013

1 0

Re: [mpich-discuss] mpich 3.0.2 last release
by Dave Goodell 25 Feb '13

25 Feb '13

You should contact the Titan2d group about this issue. Their software is assuming MPICH2 and does not seem to be compatible with the newer MPICH. Furthermore, MPICH currently is not working/supported on Cygwin, so you might have trouble there as well. -Dave On Feb 24, 2013, at 7:47 PM CST, Angelo Paone wrote: > The matter of fact is that: > I am installing a Titan2d software (from Buffalo University, Geophysical Mass Flow Group) in cygwin, I make a directory titan-3.0.0 (3.0.0 is the last version), under this directory I run $ sh install-titan.sh > > This script will automatically check for both MPI and HDF5 installation. If MPI is not found in its standard location, then the user will be prompted to enter its absolute path, or quit installation. If this script is unable to locate HDF5, then the user will be prompted to enter its absolute path or continue without installing HDF5. > > Through the process of installetion I got this error: > > $ sh install-titan.sh /usr checking for a BSD-compatible install... /usr/bin/install -c > > checking whether build environment is sane... yes checking for gawk... gawk > > checking whether make sets $(MAKE)... yes > > checking for C++ compiler default output file name... a.exe > > checking whether the C++ compiler works... yes > > checking whether we are cross compiling... no > > checking for suffix of executables... .exe > > checking for suffix of object files... o > > checking whether we are using the GNU C++ compiler... yes > > checking whether /usr/local/bin/mpiCC accepts -g... yes > > checking for style of include used by make... GNU > > checking dependency style of /usr/local/bin/mpiCC... gcc3 > > checking for gcc... /usr/local/bin/mpicc checking whether we are using the GNU C compiler... yes > > checking whether /usr/local/bin/mpicc accepts -g... yes > > checking for /usr/local/bin/mpicc option to accept ANSI C... none needed > > checking dependency style of /usr/local/bin/mpicc... gcc3 > > checking for g77... g77 > > checking whether we are using the GNU Fortran 77 compiler... yes > > checking whether g77 accepts -g... yes > > checking how to run the C preprocessor... /usr/local/bin/mpicc -E > > checking for a BSD-compatible install... /usr/bin/install -c > > checking whether ln -s works... yes > > checking whether make sets $(MAKE)... (cached) yes > > checking for ranlib... ranlib > > checking how to get verbose linking output from g77... -v > > checking for Fortran libraries of g77... -L/usr/lib/gcc/i686-pc-cygwin/3.4.4 -L/usr/lib/gcc/i686-pc-cygwin/3.4.4/../../.. -lfrtbegin -lg2c -lcygwin -luser32 -lkernel32 -ladvapi32 -lshell32 > > checking for special C compiler options needed for large files... no > > checking for _FILE_OFFSET_BITS value needed for large files... no > > checking for _LARGE_FILES value needed for large files... no > > checking build system type... i686-pc-cygwin > > checking host system type... i686-pc-cygwin checking for pthread_create in -lpthread... yes > > checking for pkg-config... /usr/local/bin/pkg-config > > checking pkg-config is at least version 0.9.0... yes > > checking for MPICH2... configure: error: Package requirements (mpich2-ch3) were not met: > > Variable 'exec_prefix' not defined in '/usr/local/lib/pkgconfig/mpich2-ch3.pc' > > Consider adjusting the PKG_CONFIG_PATH environment variable if you installed software in a non-standard prefix. > > Alternatively, you may set the environment variables MPICH2_CFLAGS and MPICH2_LIBS to avoid the need to call pkg-config. See the pkg-config man page for more details. > > make: * No targets specified and no makefile found. Stop. make: Nothing to be done for `install'. > > I have installed either the openmpi from the cygwin setup, I choose full package installation so a series of mpi packages are automatically downloaded and installed, as well as I downloaded the last release of mpich-3.0.2 and an older release mpich2-1.2.1p1 and I installed in cygwin with configure, make, make install. However, titan2d installation error tell me that the Package requirements (mpich2-ch3) were not met. I know from the titan2d user guide that "If there are multiple copies of MPI on your machine, this script will find and use the first one in your path". I do not get why my instalaltion comes with this error. Any help will be greatly appreciated > > Angelo > > > > > > On Sun, Feb 24, 2013 at 5:14 PM, Angelo Paone <angelo.paone1(a)gmail.com> wrote: > When I install mpich 3.0.2, I am not obtaining the file mpich2-ch3.pc? > > Angelo > > > On Sun, Feb 24, 2013 at 1:40 PM, Dave Goodell <goodell(a)mcs.anl.gov> wrote: > On Feb 23, 2013, at 9:14 PM CST, Angelo Paone wrote: > > > I wonder if the last mpich release 3.0.2 has the following feature in it: pkgconfig(mpich2-ch3). > > I'm not sure I understand what you are asking. > > MPICH 3.0.2 should have a pkgconfig file ("mpich.pc", generated by running configure). The 3.0.2 release is a more recent version of the MPICH2 software, just the "2" has been dropped from the name. CH3 remains the default "device". > > Are you not obtaining an mpich.pc file when you install your copy of MPICH? > > -Dave > > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss > > > > -- > > > > > > > > Angelo Paone > *********************************************** > Research Professor > Post-Doctoral Research Fellow > Marie Curie Fellow > PhD, MSc, BSc > Pusan Natioanl University > Department of Earth Science > 30, Jangjeon-dong, Geumjeong-gu > Busan, Republic of Korea > Tel. +82-51-510-2987 > > > > -- > > > > > > > > Angelo Paone > *********************************************** > Research Professor > Post-Doctoral Research Fellow > Marie Curie Fellow > PhD, MSc, BSc > Pusan Natioanl University > Department of Earth Science > 30, Jangjeon-dong, Geumjeong-gu > Busan, Republic of Korea > Tel. +82-51-510-2987 > _______________________________________________ > 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] Affinity with MPICH_ASYNC_PROGRESS
by Pavan Balaji 25 Feb '13

25 Feb '13

On 02/23/2013 06:49 PM US Central Time, Jeff Hammond wrote: > My understanding, having looked at the code for this, is that no > thread binding is set. The OS will hopefully do the right thing when > the cores are undersubscribed, but I see no thread affinity code in > the MPICH source for the comm threads. This is incorrect. Hydra allows you to bind a process to a cpuset. If you are using an affinity thread, make sure your cpuset has at least 2 cores in it. -- Pavan -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0

Re: [mpich-discuss] mpich 3.0.2 last release
by Angelo Paone 24 Feb '13

24 Feb '13

The matter of fact is that: I am installing a Titan2d software (from Buffalo University, Geophysical Mass Flow Group) in cygwin, I make a directory titan-3.0.0 (3.0.0 is the last version), under this directory I run $ sh install-titan.sh This script will automatically check for both MPI and HDF5 installation. If MPI is not found in its standard location, then the user will be prompted to enter its absolute path, or quit installation. If this script is unable to locate HDF5, then the user will be prompted to enter its absolute path or continue without installing HDF5. Through the process of installetion I got this error: $ sh install-titan.sh /usr checking for a BSD-compatible install... /usr/bin/install -c checking whether build environment is sane... yes checking for gawk... gawk checking whether make sets $(MAKE)... yes checking for C++ compiler default output file name... a.exe checking whether the C++ compiler works... yes checking whether we are cross compiling... no checking for suffix of executables... .exe checking for suffix of object files... o checking whether we are using the GNU C++ compiler... yes checking whether /usr/local/bin/mpiCC accepts -g... yes checking for style of include used by make... GNU checking dependency style of /usr/local/bin/mpiCC... gcc3 checking for gcc... /usr/local/bin/mpicc checking whether we are using the GNU C compiler... yes checking whether /usr/local/bin/mpicc accepts -g... yes checking for /usr/local/bin/mpicc option to accept ANSI C... none needed checking dependency style of /usr/local/bin/mpicc... gcc3 checking for g77... g77 checking whether we are using the GNU Fortran 77 compiler... yes checking whether g77 accepts -g... yes checking how to run the C preprocessor... /usr/local/bin/mpicc -E checking for a BSD-compatible install... /usr/bin/install -c checking whether ln -s works... yes checking whether make sets $(MAKE)... (cached) yes checking for ranlib... ranlib checking how to get verbose linking output from g77... -v checking for Fortran libraries of g77... -L/usr/lib/gcc/i686-pc-cygwin/3.4.4 -L/usr/lib/gcc/i686-pc-cygwin/3.4.4/../../.. -lfrtbegin -lg2c -lcygwin -luser32 -lkernel32 -ladvapi32 -lshell32 checking for special C compiler options needed for large files... no checking for _FILE_OFFSET_BITS value needed for large files... no checking for _LARGE_FILES value needed for large files... no checking build system type... i686-pc-cygwin checking host system type... i686-pc-cygwin checking for pthread_create in -lpthread... yes checking for pkg-config... /usr/local/bin/pkg-config checking pkg-config is at least version 0.9.0... yes *checking for MPICH2... configure: error: Package requirements (mpich2-ch3) were not met:* *Variable 'exec_prefix' not defined in '/usr/local/lib/pkgconfig/mpich2-ch3.pc'* Consider adjusting the PKG_CONFIG_PATH environment variable if you installed software in a non-standard prefix. Alternatively, you may set the environment variables MPICH2_CFLAGS and MPICH2_LIBS to avoid the need to call pkg-config. See the pkg-config man page for more details. make: *** No targets specified and no makefile found. Stop. make: Nothing to be done for `install'. I have installed either the openmpi from the cygwin setup, I choose full package installation so a series of mpi packages are automatically downloaded and installed, as well as I downloaded the last release of mpich-3.0.2 and an older release mpich2-1.2.1p1 and I installed in cygwin with configure, make, make install. However, titan2d installation error tell me that the Package requirements (mpich2-ch3) were not met. I know from the titan2d user guide that "If there are multiple copies of MPI on your machine, this script will find and use the first one in your path". I do not get why my instalaltion comes with this error. Any help will be greatly appreciated Angelo On Sun, Feb 24, 2013 at 5:14 PM, Angelo Paone <angelo.paone1(a)gmail.com>wrote: > When I install mpich 3.0.2, I am not obtaining the file mpich2-ch3.pc? > > Angelo > > > On Sun, Feb 24, 2013 at 1:40 PM, Dave Goodell <goodell(a)mcs.anl.gov> wrote: > >> On Feb 23, 2013, at 9:14 PM CST, Angelo Paone wrote: >> >> > I wonder if the last mpich release 3.0.2 has the following feature in >> it: pkgconfig(mpich2-ch3). >> >> I'm not sure I understand what you are asking. >> >> MPICH 3.0.2 should have a pkgconfig file ("mpich.pc", generated by >> running configure). The 3.0.2 release is a more recent version of the >> MPICH2 software, just the "2" has been dropped from the name. CH3 remains >> the default "device". >> >> Are you not obtaining an mpich.pc file when you install your copy of >> MPICH? >> >> -Dave >> >> _______________________________________________ >> discuss mailing list discuss(a)mpich.org >> To manage subscription options or unsubscribe: >> https://lists.mpich.org/mailman/listinfo/discuss >> > > > > -- > > > > > > > > Angelo Paone > *********************************************** > Research Professor > Post-Doctoral Research Fellow > Marie Curie Fellow > PhD, MSc, BSc > Pusan Natioanl University > Department of Earth Science > 30, Jangjeon-dong, Geumjeong-gu > Busan, Republic of Korea > Tel. +82-51-510-2987 > -- Angelo Paone *********************************************** Research Professor Post-Doctoral Research Fellow Marie Curie Fellow PhD, MSc, BSc Pusan Natioanl University Department of Earth Science 30, Jangjeon-dong, Geumjeong-gu Busan, Republic of Korea Tel. +82-51-510-2987

1 0

Re: [mpich-discuss] mpich 3.0.2 last release
by Angelo Paone 24 Feb '13

24 Feb '13

When I install mpich 3.0.2, I am not obtaining the file mpich2-ch3.pc? Angelo On Sun, Feb 24, 2013 at 1:40 PM, Dave Goodell <goodell(a)mcs.anl.gov> wrote: > On Feb 23, 2013, at 9:14 PM CST, Angelo Paone wrote: > > > I wonder if the last mpich release 3.0.2 has the following feature in > it: pkgconfig(mpich2-ch3). > > I'm not sure I understand what you are asking. > > MPICH 3.0.2 should have a pkgconfig file ("mpich.pc", generated by running > configure). The 3.0.2 release is a more recent version of the MPICH2 > software, just the "2" has been dropped from the name. CH3 remains the > default "device". > > Are you not obtaining an mpich.pc file when you install your copy of MPICH? > > -Dave > > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss > -- Angelo Paone *********************************************** Research Professor Post-Doctoral Research Fellow Marie Curie Fellow PhD, MSc, BSc Pusan Natioanl University Department of Earth Science 30, Jangjeon-dong, Geumjeong-gu Busan, Republic of Korea Tel. +82-51-510-2987

1 0

Re: [mpich-discuss] mpich 3.0.2 last release
by Dave Goodell 24 Feb '13

24 Feb '13

On Feb 23, 2013, at 9:14 PM CST, Angelo Paone wrote: > I wonder if the last mpich release 3.0.2 has the following feature in it: pkgconfig(mpich2-ch3). I'm not sure I understand what you are asking. MPICH 3.0.2 should have a pkgconfig file ("mpich.pc", generated by running configure). The 3.0.2 release is a more recent version of the MPICH2 software, just the "2" has been dropped from the name. CH3 remains the default "device". Are you not obtaining an mpich.pc file when you install your copy of MPICH? -Dave

1 0

Re: [mpich-discuss] Maximum number of communicators
by Dave Goodell 24 Feb '13

24 Feb '13

Test code indeed would be helpful. You may get better warnings about context ID and/or communicator leaks if you configure a stock MPICH with "--enable-g=all". -Dave On Feb 23, 2013, at 7:05 PM CST, Jeff Hammond wrote: > It's 2048 in all MPICH2-derived implementations that I know about, > including the ones used on BGP and BGQ. > > Since I know what your issue is already, I think the problem has > something to do with context id leaking. You should post your test > code. > > Jeff > > On Sat, Feb 23, 2013 at 6:24 PM, Nichols A. Romero > <naromero(a)alcf.anl.gov> wrote: >> Hi, >> >> What is the maximum number of communicators in the MPICH that is used on BG/P vs. that used on BG/Q? >> >> Sent from my iPhone >> _______________________________________________ >> discuss mailing list discuss(a)mpich.org >> To manage subscription options or unsubscribe: >> https://lists.mpich.org/mailman/listinfo/discuss > > > > -- > Jeff Hammond > Argonne Leadership Computing Facility > University of Chicago Computation Institute > jhammond(a)alcf.anl.gov / (630) 252-5381 > http://www.linkedin.com/in/jeffhammond > https://wiki.alcf.anl.gov/parts/index.php/User:Jhammond > _______________________________________________ > 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] MPICH2 -Problem
by Andrey V. Malyshev 24 Feb '13

24 Feb '13

Hello, hamideh! You wrote to <discuss(a)mpich.org> on Sun, 24 Feb 2013 18:30:43 +0330: hj> I see file 'config.log' from the same directory where the 'configure' hj> was started. I attached its to this email. In the content of this file, hj> pthread libraries such as "pthread.h,pthread_yield" have make. but hj> there are related error... hj> Do you know that what do we do? 1) /home/inet/src/applications/sctpapp/SCTPlikeAPI/include: No such file or directory Is this directory exists? If yes, maybe the path at all is incorrect? Try exactly 'ls -la /home/inet/src/applications/sctpapp/SCTPlikeAPI/include' 2) /usr/bin/ld: cannot find -lsctplike Have you correctly installed your SCTPlikeAPI? Add its 'lib' directory to ldconfig paths (/etc/ld.so.conf.d). This directory have to contain the file (or softlink) libsctplike.so. Then run from root ldconfig -v, and try to configure again. hj> On Sun, Feb 24, 2013 at 12:22 PM, Andrey V. Malyshev hj> <amal(a)icm.krasn.ru>wrote: ??>> Hello, hamideh! ??>> You wrote to <discuss(a)mpich.org> on Sun, 24 Feb 2013 10:12:01 +0330: ??>> hj>>> I think configure doesn't done correctly! ??>> ??>> That's right. ??>> hj>>> configure command is : hj>>> ../configure --with-device=ch3:sctp hj>>> --with-sctp=/home/inet/src/applications/sctpapp/SCTPlikeAPI hj>>> --prefix=/home/mpich2-hginit ï¿½disable-mpe hj>>> What I am doing? :( ??>> ??>> See the file 'config.log' from the same directory where the ??>> 'configure' was started, or where the 'configure' file is ??>> located. Pthread library was not found, you should understand, why, ??>> and maybe then it becomes clear how to fix. ??>> hj>>> On Sun, Feb 24, 2013 at 10:07 AM, Andrey V. Malyshev hj>>> <amal(a)icm.krasn.ru>wrote: ??>> ??>>>> Hello, hamideh! ??>>>> You wrote to <discuss(a)mpich.org> on Sun, 24 Feb 2013 08:23:37 +0330: ??>>>> hj>>>>> checking for pthread.h... yes ??>>>> checking for pthread_yield... no ??>>>> checking for library containing pthread_key_create... no ??>>>> configure: error: unable to find pthreads library ??>>>> ??>>>> See config.log, what error occurs during this check? ??>>>> ??>>>> -- ??>>>> With best regards, Andrey V. Malyshev. ??>>>> E-mail: amal(a)krasn.ru ??>>>> _______________________________________________ ??>>>> discuss mailing list discuss(a)mpich.org ??>>>> To manage subscription options or unsubscribe: ??>>>> https://lists.mpich.org/mailman/listinfo/discuss ??>>>> hj>>> -- hj>>> *Best Regards* hj>>> * ??>> hj>>> Hamideh Jabalameli* ??>> hj>>> M.Sc Candidate, hj>>> School of Electrical and Computer Engineering, hj>>> College of Engineering, hj>>> Electronici Shahid Beheshti Tehran University hj>>> h.jabal(a)Mail.sbu.ac.ir hj>>> hamideh.jabalameli(a)gmail.com ??>> ??>> -- ??>> With best regards, Andrey V. Malyshev. ??>> E-mail: amal(a)krasn.ru ??>> ??>> _______________________________________________ ??>> discuss mailing list discuss(a)mpich.org ??>> To manage subscription options or unsubscribe: ??>> https://lists.mpich.org/mailman/listinfo/discuss ??>> hj> -- hj> *Best Regards* hj> * hj> Hamideh Jabalameli* hj> M.Sc Candidate, hj> School of Electrical and Computer Engineering, hj> College of Engineering, hj> Electronici Shahid Beheshti Tehran University hj> h.jabal(a)Mail.sbu.ac.ir hj> hamideh.jabalameli(a)gmail.com -- With best regards, Andrey V. Malyshev. E-mail: amal(a)krasn.ru

1 0

Re: [mpich-discuss] Affinity with MPICH_ASYNC_PROGRESS
by Jeff Hammond 24 Feb '13

24 Feb '13

I'll respond to the non-MPICH aspects of this thread offline... >> I think the only way for async progress to work well is to have >> fine-grain locking since of MPI, as is done in PAMID. Any >> implementation that resorts to fat-locking is probably better off >> without async progress unless the application is doing something >> really silly (like never calling MPI on a rank that is the target of >> MPI RMA). > > There's more than 100k cycles between times that I enter the MPI stack. > There are only two threads ever that contend for locks (my funneled thread > and MPICH's async-progress thread). I'm not convinced that you need super > fine-grained locks to make progress during that time period. FWIW, my > experience has been that standard Nemesis does a _much_ better job of making > asynchronous progress than Cray's implementation. If we could look at the > code for Cray's implementation, we might be able to get a better idea of > why. I would be interesting to analyze async progress in detail using MPICH and MVAPICH, since those are OSS. We are never going to get the CrayPICH source (I've asked, NCSA asked) so it's not prudent to debate what we would do with it if we did. We could reimplement the uGNI netmod if we had nothing better to do. Jeff -- Jeff Hammond Argonne Leadership Computing Facility University of Chicago Computation Institute jhammond(a)alcf.anl.gov / (630) 252-5381 http://www.linkedin.com/in/jeffhammond https://wiki.alcf.anl.gov/parts/index.php/User:Jhammond

1 0