September 2015 - Darshan-users - mailman3.cels.anl.gov

Darshan repository moved to gitlab
by Phil Carns 25 Sep '15

25 Sep '15

Hi everyone, The Darshan source code, issue tracking, and milestone management has been moved to the Darshan gitlab page at ANL. You can find instructions there for how to clone the repository: https://xgitlab.cels.anl.gov/darshan/darshan If you have an existing clone of the Darshan repository from its previous location, then you can switch it to point to the new repository with the following command: git remote set-url origin https://xgitlab.cels.anl.gov/darshan/darshan.git If you are interesting in contributed to Darshan, the gitlab page brings several new features to the project including the ability for external users to sign up for accounts, simplified forking and pull requests, and the ability to track issues and milestones. thanks, -Darshan development team

1 0

Re: [Darshan-users] Different builds for each MPI/compiler
by Phil Carns 10 Sep '15

10 Sep '15

On 09/09/2015 10:55 AM, Chih-Song Kuo wrote: > > *Von:*[email protected] > [mailto:[email protected]] *Im Auftrag von *Phil > Carns > *Gesendet:* Mittwoch, 9. September 2015 15:52 > *An:* darshan-users(a)lists.mcs.anl.gov > *Betreff:* Re: [Darshan-users] Different builds for each MPI/compiler > > On 09/09/2015 06:55 AM, Chih-Song Kuo wrote: > > Hello Darshan user community, > > I just started using Darshan. Despite it runs pretty well, I > wonder if I have to create different builds for each combination > of MPI and compiler. I had this concern because you wrote some > issues about libfmpich.so. I as a sales consultant of an > HPC-system vendor run benchmarks with many combinations of MPI and > compilers to advise our customer the best choice. > > Your insight will be greatly appreciated. > > Chih-Song > > > Hi Chih-Song, > > I have two pieces of information that might be helpful: > > a) The underlying compiler generally doesn't matter as much as the MPI > library itself. We usually compile Darshan itself with GNU compilers > for maximum compatibility; the resulting Darshan library should be > compatible with most any compiler/linker. > > b) As for the MPI library, the MPICH ABI compatibility initiative > (https://www.mpich.org/abi/) should help in theory by making sure that > quite a few of the MPI implementations have a compatible binary > interface. Our team hasn't tested this capability first hand, so I > can't guarantee it will work, but I think the prospects are good. > > thanks, > -Phil > > Hi Phil, > > Thanks. It looks like I might be worrying too much. According my > experience with IPMï¿½s I/O profiling module, different MPIs seem to > have some difference in the appearance of the ï¿½constï¿½ modifier which > caused some trouble in the compilation. > This would mainly be an issue for mixing MPI 2 and MPI 3 implementations (const correctness is part of the latter spec). I imagine there would be more significant roadblocks to that level of interoperability though. It's probably best to at least stick to one major version or the other for a given Darshan build. FWIW Darshan does detect the const issue at compile time (for the Darshan library itself), so you can build a version of Darshan for use with either API generation. > Also, on > https://www.nersc.gov/users/software/debugging-and-profiling/ipm/, > there seems to be some module explicitly built for Intel compilers. I > thought Darshan uses a similar profiling technique, so I was asking. > I don't believe this is the case. Darshan's various instrumentation methods (link-time wrappers, PMPI, or preloading) are all compiler-independent to my knowledge. IPM captures more in-depth information so it may rely on additional methods. > I read in some Darshan research papers that the tool has been enabled > by default on several NERSC clusters. Do the users there have a wide > choice of MPI implementations? Do you have any clue which MPIs are > chosen by the user and also get successfully profiled by Darshan? > I don't have that information myself, sorry. I know that Darshan works with each of the compilers that are available (and users do switch those for various reasons), but I'm not sure about the MPI implementations. thanks, -Phil > Cheers, > > Chh-Song >

3 2

Re: [Darshan-users] Interpretting Darshan rport
by Phil Carns 09 Sep '15

09 Sep '15

Thanks for pointing that out Frederico - that is indeed exactly where the terminology came from! -Phil On 09/09/2015 05:48 PM, Padua, Federico wrote: > Hi Richard, > Hi Shane, > > if I may add something just to give some credits, one of the original definitions and distinctions between consecutive and sequential access appears in the early paper in 1996 titled: > File-Access Characteristics of Parallel Scientific Workloads, > also known as Charisma project. > Probably some work before that had defined it but I donï¿½t know it, thatï¿½s the ï¿½oldestï¿½ Iï¿½m aware of... > > The definition is also mentioned here (same project, same authors): > http://www.cs.dartmouth.edu/~dfk/papers/ap-workload-tr.pdf > > From what it appears from the Darshan plots, they use this convention as Shean well explained. > > Best, > -Federico > > >> On 09 Sep 2015, at 13:44, Shane Snyder <ssnyder(a)mcs.anl.gov> wrote: >> >> Hi Richard, >> >> Sequential describes I/O accesses that are just at any offset greater than the immediately preceding I/O operation. For instance, say I write 100 bytes at offset 0, then my next access is a read at offset 500 -- that would be a sequential access. >> >> Consecutive is more strict in that it requires that the current operation be at the offset immediately following the end of the preceding access's end. An example would be if I write 100 bytes at offset 0, then read at offset 100, instead. So, all consecutive accesses are also sequential but not vice versa. >> >> Note that there is no distinction between the types of the 2 operations. If the current operation is a write, we don't care if the previous operation was a read or a write, just the ending offset of it. >> >> I should probably mention that we have fixed this graph to be more clear in the new development branch of darshan we are preparing to release, as it has been a source of confusion for a number of people. >> >> --Shane >> >> On 09/09/2015 03:23 PM, Hedges, Richard M. wrote: >>> In the default report, the fourth plot (labelled ?I/O Patternï¿½) distinguishes between ï¿½Consecutiveï¿½ and ï¿½Sequentialï¿½ . Could someone explain the distinction for me? >>> >>> Thanks, >>> - Richard >>> >>> ==================================================== >>> >>> Richard Hedges >>> Customer Support and Test - File Systems Project >>> Development Environment Group - Livermore Computing >>> Lawrence Livermore National Laboratory >>> 7000 East Avenue, MS L-557 >>> Livermore, CA 94551 >>> >>> v: (925) 423-2699 >>> f: (925) 423-6961 >>> E: richard-hedges(a)llnl.gov >>> >>> >>> >>> >>> _______________________________________________ >>> Darshan-users mailing list >>> >>> Darshan-users(a)lists.mcs.anl.gov >>> https://lists.mcs.anl.gov/mailman/listinfo/darshan-users >> _______________________________________________ >> Darshan-users mailing list >> Darshan-users(a)lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/darshan-users

1 0

Interpretting Darshan rport
by Hedges, Richard M. 09 Sep '15

09 Sep '15

In the default report, the fourth plot (labelled ?I/O Pattern”) distinguishes between ”Consecutive” and “Sequential” . Could someone explain the distinction for me? Thanks, - Richard ==================================================== Richard Hedges Customer Support and Test - File Systems Project Development Environment Group - Livermore Computing Lawrence Livermore National Laboratory 7000 East Avenue, MS L-557 Livermore, CA 94551 v: (925) 423-2699 f: (925) 423-6961 E: richard-hedges(a)llnl.gov

3 2

Re: [Darshan-users] Different builds for each MPI/compiler
by Phil Carns 09 Sep '15

09 Sep '15

On 09/09/2015 06:55 AM, Chih-Song Kuo wrote: > > Hello Darshan user community, > > I just started using Darshan. Despite it runs pretty well, I wonder if > I have to create different builds for each combination of MPI and > compiler. I had this concern because you wrote some issues about > libfmpich.so. I as a sales consultant of an HPC-system vendor run > benchmarks with many combinations of MPI and compilers to advise our > customer the best choice. > > Your insight will be greatly appreciated. > > Chih-Song > Hi Chih-Song, I have two pieces of information that might be helpful: a) The underlying compiler generally doesn't matter as much as the MPI library itself. We usually compile Darshan itself with GNU compilers for maximum compatibility; the resulting Darshan library should be compatible with most any compiler/linker. b) As for the MPI library, the MPICH ABI compatibility initiative (https://www.mpich.org/abi/) should help in theory by making sure that quite a few of the MPI implementations have a compatible binary interface. Our team hasn't tested this capability first hand, so I can't guarantee it will work, but I think the prospects are good. thanks, -Phil

2 1

Different builds for each MPI/compiler
by Chih-Song Kuo 09 Sep '15

09 Sep '15

Hello Darshan user community, I just started using Darshan. Despite it runs pretty well, I wonder if I have to create different builds for each combination of MPI and compiler. I had this concern because you wrote some issues about libfmpich.so. I as a sales consultant of an HPC-system vendor run benchmarks with many combinations of MPI and compilers to advise our customer the best choice. Your insight will be greatly appreciated. Chih-Song

1 0

Re: [Darshan-users] No PDF file generation with darshan-summary-per-file.sh or darshan-job-summary.pl utility
by Phil Carns 08 Sep '15

08 Sep '15

Hi Preeti, Sorry you ran into trouble with this- I've put some suggestions inline with your questions below. On 09/06/2015 11:07 PM, Suman, Preeti wrote: > > Hi, > > I have recently started working with Darshan 2.3.1 on my system. I > have installed all the prerequisites: Perl, pdflatex, epstopdf, and > gnuplot to enable darshan-job-summary.pl utility. I have few questions > specified below regarding the use of darshan-summary-per-file.sh and > darshan-job-summary.pl. > > System specification: > > Kernel Version: Linux 4.0.5-1.el7.elrepo.x86_64 > > Operating System: CentOS Linux release 7.1.1503 > > CPU: Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz (Haswell-EP) > > Architecture:x86_64 > > Compiler: GCC 5.1.0 > > MPI: Intel MPI > > #### BEGIN : output of darshan-summary-per-file.sh > > $ darshan-summary-per-file.sh > carns_mpi-io-test-ppc64-2.3.1_id426989_3-13-61875-4723177368895607730_1.darshan.gz > ../output > > Status: Generating summary for file 1 of 1: mpi-io-test.dat > > ======================================================= > > Slowest unique file time: 0 > > Slowest shared file time: 0.190554 > > Total bytes read and written by app (may be incorrect): 33554432 > > Total absolute I/O time: 0.190554 > > LaTeX generation (phase1) failed [256], aborting summary creation. > > error log: > > or enter new name. (Default extension: sty) > > Enter file name: > > ! Emergency stop. > > <read *> > > l.8 \usepackage > > {multirow}^^M > > ! ==> Fatal error occurred, no output PDF file produced! > > Transcript written on summary.log. > > ======================================================= > > darshan-summary-per-file.sh done. Results can be found in > /nfs/fx/proj/coralhpctools/users/psuman/myBuilds/Darshan/gcc/0904_logs/out1/*.pdf. > > #### END: darshan-summary-per-file.sh > > I am using the few benchmarks and darshan-test/example-output files to > test the installation. There is no PDF file generation while running > the darshan-summary-per-file.sh or darshan-job-summary.pl utility. > Here are few questions: > > a) Is the error message of darshan-summary-per-file.sh utility > mentioned above linked to Darshan or PDFLatex? > This is a pdflatex error. > b) Is there any other dependency needed to enable > darshan-summary-per-file.sh and other utilities? > Apparently so :) We might be able to add safety checks to prevent this in the future once we figure out what your problem is. > c) If this is not a dependency error, What could be the reason for the > error? > In my experience pdflatex errors are notoriously misleading, but it looks like the problem might be a missing multirow.sty file. Can you check to see if this is present on your system? On Ubuntu 15.04 it can be found in /usr/share/texlive/texmf-dist/tex/latex/multirow/multirow.sty and is a part of the texlive-latex-extra package. thanks, -Phil > Will greatly appreciate your response. > > Thanks, > > Preeti >

2 1

No PDF file generation with darshan-summary-per-file.sh or darshan-job-summary.pl utility
by Suman, Preeti 06 Sep '15

06 Sep '15

Hi, I have recently started working with Darshan 2.3.1 on my system. I have installed all the prerequisites: Perl, pdflatex, epstopdf, and gnuplot to enable darshan-job-summary.pl utility. I have few questions specified below regarding the use of darshan-summary-per-file.sh and darshan-job-summary.pl. System specification: Kernel Version: Linux 4.0.5-1.el7.elrepo.x86_64 Operating System: CentOS Linux release 7.1.1503 CPU: Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz (Haswell-EP) Architecture:x86_64 Compiler: GCC 5.1.0 MPI: Intel MPI #### BEGIN : output of darshan-summary-per-file.sh $ darshan-summary-per-file.sh carns_mpi-io-test-ppc64-2.3.1_id426989_3-13-61875-4723177368895607730_1.darshan.gz ../output Status: Generating summary for file 1 of 1: mpi-io-test.dat ======================================================= Slowest unique file time: 0 Slowest shared file time: 0.190554 Total bytes read and written by app (may be incorrect): 33554432 Total absolute I/O time: 0.190554 LaTeX generation (phase1) failed [256], aborting summary creation. error log: or enter new name. (Default extension: sty) Enter file name: ! Emergency stop. <read *> l.8 \usepackage {multirow}^^M ! ==> Fatal error occurred, no output PDF file produced! Transcript written on summary.log. ======================================================= darshan-summary-per-file.sh done. Results can be found in /nfs/fx/proj/coralhpctools/users/psuman/myBuilds/Darshan/gcc/0904_logs/out1/*.pdf. #### END: darshan-summary-per-file.sh I am using the few benchmarks and darshan-test/example-output files to test the installation. There is no PDF file generation while running the darshan-summary-per-file.sh or darshan-job-summary.pl utility. Here are few questions: a) Is the error message of darshan-summary-per-file.sh utility mentioned above linked to Darshan or PDFLatex? b) Is there any other dependency needed to enable darshan-summary-per-file.sh and other utilities? c) If this is not a dependency error, What could be the reason for the error? Will greatly appreciate your response. Thanks, Preeti

1 0