[daip] AIPSmark large
Eric Greisen
egreisen at nrao.edu
Thu Mar 27 10:36:44 EDT 2008
Glenn Morrison wrote:
> Took 177sec (realtime) to run the large test.
>
> Would I get better results if I used more than one disk?
>
>
>
> 2 0 26-MAR-2008 14:30:01 AIPS RUN Y2KEXEC
#'
> 2 5 26-MAR-2008 14:33:02 AIPS 'PRINTING MESSAGES SUGGESTIVE OF ERROR: PRIO > 5'
Actually - the real time between these two messages is 181 sec =>
aipsmark 221.
Yes - more than one disk might help but I do all tests using only one
disk. One thing that is important is to kill the message server before
starting and to direct all printed messages to a file with OUTPRINT.
The lower cpu to real ratios are probably a reflection of the accounting
algorithms more than something really different between this machine and
the usual Linux box.
Eric
> Sequential list secs secs
>
> UVSRT 2 1967 26-MAR-2008 14:30:01 2 0.47 246 TST
> UVDIF 2 1967 26-MAR-2008 14:30:03 0 0.23 30 TST
> IMAGR 2 1967 26-MAR-2008 14:30:03 4 1.70 205 TST
> COMB 2 1967 26-MAR-2008 14:30:07 0 0.11 50 TST
> COMB 2 1967 26-MAR-2008 14:30:08 0 0.10 50 TST
> APCLN 2 1967 26-MAR-2008 14:30:08 15 10.37 1869 TST
> COMB 2 1967 26-MAR-2008 14:30:23 1 0.11 50 TST
> SUBIM 2 1967 26-MAR-2008 14:30:24 1 0.08 34 TST
> APCLN 2 1967 26-MAR-2008 14:30:25 2 0.90 179 TST
> COMB 2 1967 26-MAR-2008 14:30:27 0 0.11 50 TST
> CCMRG 2 1967 26-MAR-2008 14:30:27 1 0.10 13 TST
> CALIB 2 1967 26-MAR-2008 14:30:28 12 6.72 980 TST
> UVDIF 2 1967 26-MAR-2008 14:30:40 0 0.23 30 TST
> IMAGR 2 1967 26-MAR-2008 14:30:40 105 71.65 7986 TST
> COMB 2 1967 26-MAR-2008 14:32:26 0 0.11 50 TST
> IMAGR 2 1967 26-MAR-2008 14:32:26 4 1.73 205 TST
> SUBIM 2 1967 26-MAR-2008 14:32:30 0 0.08 36 TST
> VTESS 2 1967 26-MAR-2008 14:32:31 29 20.33 4020 TST
> COMB 2 1967 26-MAR-2008 14:33:01 0 0.11 50 TST
> COMB 2 1967 26-MAR-2008 14:33:01 1 0.11 50 TST
>
> Version: ALL
> # tasks 20
> Tot real 3.0
> Tot cpu 1.923
> Tot ioc 16183.
> Av cp/re 0.59250
> # abort 0
> # users 1
> # tasks 9
>
More information about the Daip
mailing list