[Darshan-users] Darshan & EPCC benchio different behaviour
Piero LANUCARA
p.lanucara at cineca.it
Wed Feb 5 04:56:20 CST 2020
p.s
to be more "verbose" I add to the discussion:
Darshan output for the "serial" run (serial.pdf)
Darshan output for the MPI-IO run (mpiio.pdf)
benchio output for "serial" run (serial.out)
benchio output for "MPI-IO" run (mpi-io.out)
thanks
Piero
Il 04/02/2020 19:44, Piero LANUCARA ha scritto:
> Dear all
>
> I'm using Darshan to measure EPCC benchio benchmark
> (https://github.com/EPCCed/benchio) behaviour on a given x86 Tier1
> machine.
>
> running two benchio tests (MPI-IO and serial) a different behaviour
> appear
>
> while Darhsan pdf log file is able to recover the estimated time and
> bandwidth in the MPI-IO case, the "serial" run is completely
> underestimated by Darshan (the time and bandwidth are less/greater
> than benchio output).
>
> Suggestions are welcomed
>
> thanks
>
> Piero
>
> _______________________________________________
> Darshan-users mailing list
> Darshan-users at lists.mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/darshan-users
-------------- next part --------------
A non-text attachment was scrubbed...
Name: serial.pdf
Type: application/pdf
Size: 67618 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/darshan-users/attachments/20200205/a52b3657/attachment-0002.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mpiio.pdf
Type: application/pdf
Size: 79709 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/darshan-users/attachments/20200205/a52b3657/attachment-0003.pdf>
-------------- next part --------------
Simple Parallel IO benchmark
----------------------------
Running on 36 process(es)
Process grid is ( 3 , 3 , 4 )
Array size is ( 256 , 256 , 256 )
Global size is ( 768 , 768 , 1024 )
Total amount of data = 4608.00000000000 MiB
Clock resolution is 1.00000000000000 , usecs
------
Serial
------
Writing to benchio_files/serial.dat
time = 4.84900808334351 , rate = 950.297446570284 MiB/s
time = 4.83415985107422 , rate = 953.216306857548 MiB/s
time = 4.82698416709900 , rate = 954.633336361116 MiB/s
time = 4.82474207878113 , rate = 955.076960541716 MiB/s
time = 4.82010507583618 , rate = 955.995756835366 MiB/s
time = 4.82056093215942 , rate = 955.905353100847 MiB/s
time = 4.81360197067261 , rate = 957.287292982415 MiB/s
time = 4.81427383422852 , rate = 957.153697248804 MiB/s
time = 4.81633496284485 , rate = 956.744087682433 MiB/s
time = 4.82088708877563 , rate = 955.840681423281 MiB/s
mintime = 4.81360197067261 , maxrate = 957.287292982415 MiB/s
avgtime = 4.82406580448151 , avgrate = 955.215091960381 MiB/s
Deleting: benchio_files/serial.dat
--------
Finished
--------
real 0m51.136s
user 0m0.028s
sys 0m0.019s
-------------- next part --------------
Simple Parallel IO benchmark
----------------------------
Running on 36 process(es)
Process grid is ( 3 , 3 , 4 )
Array size is ( 256 , 256 , 256 )
Global size is ( 768 , 768 , 1024 )
Total amount of data = 4608.00000000000 MiB
Clock resolution is 1.00000000000000 , usecs
------
MPI-IO
------
Writing to benchio_files/mpiio.dat
time = 3.85350108146667 , rate = 1195.79569398905 MiB/s
time = 3.90402007102966 , rate = 1180.32180064706 MiB/s
time = 3.88898396492004 , rate = 1184.88531749314 MiB/s
time = 3.89588212966919 , rate = 1182.78732431550 MiB/s
time = 3.88649392127991 , rate = 1185.64446345062 MiB/s
time = 3.90831685066223 , rate = 1179.02416208123 MiB/s
time = 3.90417695045471 , rate = 1180.27437241627 MiB/s
time = 3.89346003532410 , rate = 1183.52312806427 MiB/s
time = 3.89405012130737 , rate = 1183.34378255330 MiB/s
time = 3.89044594764709 , rate = 1184.44005186266 MiB/s
mintime = 3.85350108146667 , maxrate = 1195.79569398905 MiB/s
avgtime = 3.89193310737610 , avgrate = 1184.00400968731 MiB/s
Deleting: benchio_files/mpiio.dat
--------
Finished
--------
real 0m41.803s
user 0m0.033s
sys 0m0.021s
More information about the Darshan-users
mailing list