time like utility? - Hewlett Packard

This is a discussion on time like utility? - Hewlett Packard ; Is there an easy utility to time, wallclock and processor, how long a process takes? I am thinking like the unix "time" utility. Measures down to 100 milliseconds I think. I've gotten some 9gig disks to work as data disks ...

+ Reply to Thread
Results 1 to 8 of 8

Thread: time like utility?

  1. time like utility?

    Is there an easy utility to time, wallclock and processor, how long a
    process takes? I am thinking like the unix "time" utility. Measures
    down to 100 milliseconds I think.

    I've gotten some 9gig disks to work as data disks (thought not yet as
    boot disks...) and the for file access is very startling to me, so
    much so that I need to test some big file writes and reads on 'em to
    be sure. They seem to be something like 9 to 10 times faster than poor
    old internal drive.

    Now admittedly, these drives *can* perform a whole lot faster than
    Telzy can drive them (they can and do handle 160mbs), even if they are
    considered slow and small by today's standards. But that much
    difference on a 33mhz processor? Even going through a Fast/Wide
    adapter??

    -Paul

    * To join/leave the list, search archives, change list settings, *
    * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


  2. Re: time like utility?

    Paul,

    The 937LX that you have (it was a 937 right?) has an NIO bus, it is rated at 32mbits/sec with a sustained rate of 20mbits per second.

    Somebody help me here, senior moment, what does NIO stand for?

    -Craig



    --- On Thu, 10/23/08, Paul Raulerson wrote:

    > From: Paul Raulerson
    > Subject: time like utility?
    > To: HP3000-L@RAVEN.UTC.EDU
    > Date: Thursday, October 23, 2008, 8:22 PM
    > Is there an easy utility to time, wallclock and processor,
    > how long a
    > process takes? I am thinking like the unix "time"
    > utility. Measures
    > down to 100 milliseconds I think.
    >
    > I've gotten some 9gig disks to work as data disks
    > (thought not yet as
    > boot disks...) and the for file access is very startling to
    > me, so
    > much so that I need to test some big file writes and reads
    > on 'em to
    > be sure. They seem to be something like 9 to 10 times
    > faster than poor
    > old internal drive.
    >
    > Now admittedly, these drives *can* perform a whole lot
    > faster than
    > Telzy can drive them (they can and do handle 160mbs), even
    > if they are
    > considered slow and small by today's standards. But
    > that much
    > difference on a 33mhz processor? Even going through a
    > Fast/Wide
    > adapter??
    >
    > -Paul
    >
    > * To join/leave the list, search archives, change list
    > settings, *
    > * etc., please visit
    > http://raven.utc.edu/archives/hp3000-l.html *


    * To join/leave the list, search archives, change list settings, *
    * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


  3. Re: time like utility?

    Nope -- it is a 917LX. I think that the processors are the same
    speed, but I have no idea about data bus differences.
    -Paul

    On Oct 23, 2008, at 8:39 PM, Craig Lalley wrote:

    > Paul,
    >
    > The 937LX that you have (it was a 937 right?) has an NIO bus, it is
    > rated at 32mbits/sec with a sustained rate of 20mbits per second.
    >
    > Somebody help me here, senior moment, what does NIO stand for?
    >
    > -Craig
    >
    >
    >
    > --- On Thu, 10/23/08, Paul Raulerson wrote:
    >
    >> From: Paul Raulerson
    >> Subject: time like utility?
    >> To: HP3000-L@RAVEN.UTC.EDU
    >> Date: Thursday, October 23, 2008, 8:22 PM
    >> Is there an easy utility to time, wallclock and processor,
    >> how long a
    >> process takes? I am thinking like the unix "time"
    >> utility. Measures
    >> down to 100 milliseconds I think.
    >>
    >> I've gotten some 9gig disks to work as data disks
    >> (thought not yet as
    >> boot disks...) and the for file access is very startling to
    >> me, so
    >> much so that I need to test some big file writes and reads
    >> on 'em to
    >> be sure. They seem to be something like 9 to 10 times
    >> faster than poor
    >> old internal drive.
    >>
    >> Now admittedly, these drives *can* perform a whole lot
    >> faster than
    >> Telzy can drive them (they can and do handle 160mbs), even
    >> if they are
    >> considered slow and small by today's standards. But
    >> that much
    >> difference on a 33mhz processor? Even going through a
    >> Fast/Wide
    >> adapter??
    >>
    >> -Paul
    >>
    >> * To join/leave the list, search archives, change list
    >> settings, *
    >> * etc., please visit
    >> http://raven.utc.edu/archives/hp3000-l.html *


    * To join/leave the list, search archives, change list settings, *
    * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


  4. Re: time like utility?

    A few years back I wrote a script that does that. I was using it to fine
    tune some scripts that I was developing, but it can be used to 'time' any
    command on the HP3000.

    I'll try to post it before Monday, as at the moment I don't have access to
    that server.

    Regards
    Paul Christidis

    HP-3000 Systems Discussion wrote on 10/23/2008
    06:22:16 PM:

    > Is there an easy utility to time, wallclock and processor, how long a
    > process takes? I am thinking like the unix "time" utility. Measures
    > down to 100 milliseconds I think.
    >
    > I've gotten some 9gig disks to work as data disks (thought not yet as
    > boot disks...) and the for file access is very startling to me, so
    > much so that I need to test some big file writes and reads on 'em to
    > be sure. They seem to be something like 9 to 10 times faster than poor
    > old internal drive.
    >
    > Now admittedly, these drives *can* perform a whole lot faster than
    > Telzy can drive them (they can and do handle 160mbs), even if they are
    > considered slow and small by today's standards. But that much
    > difference on a 33mhz processor? Even going through a Fast/Wide
    > adapter??
    >
    > -Paul
    >
    > * To join/leave the list, search archives, change list settings, *
    > * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


    * To join/leave the list, search archives, change list settings, *
    * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


  5. Re: time like utility?

    This might be close to what you want

    :PRINT TIMETEST.CMD.SHR
    anyparm cmd_ = 'MUST ENTER A COMMAND TO RUN TIME-TESTS ON'
    option nolist
    setvar connsecs_ hpconnsecs
    setvar cpumsecs_ hpcpumsecs
    setvar cpusecs_ hpcpusecs
    !cmd_
    setvar connsecs_ hpconnsecs - connsecs_
    setvar cpumsecs_ hpcpumsecs - cpumsecs_
    setvar cpusecs_ hpcpusecs - cpusecs_
    setvar connsecs_X rht(" !connsecs_",6)
    setvar cpusecs_X rht(" !cpusecs_",6)
    setvar cpumsecs_X rht(" !cpumsecs_",8)
    echo Took!connsecs_x seconds (wall time)&
    !cpusecs_x cpu seconds&
    !cpumsecs_x cpu milliseconds
    return
    #
    # TIMETEST - command-file to tell how long it takes to
    # run one specified program, mpe-command,
    # command-file, etc.
    # - Cannot be called on itself, works on other
    # tested command files.
    # - Not tested with any redirection.
    # - CPU times not accurate if run under Editor,
    # etc. Best run from : prompt.
    #
    # Examples:
    # XEQ TIMETEST.TESTCMD.SHR SHOWME (MPE command)
    # TIMETEST RUN YADAYADA.UNDER.PRIV;INFO="Y" (program)
    # timetest devwait.syscmd.shr 7 AVAIL 1 1 (command file)
    #
    # -------------------------------------------------
    #
    # Orig version 12/01/2003 Dave Powell, MMfab, Inc.



    ----- Original Message -----
    From: "Paul Raulerson"
    To:
    Sent: Thursday, October 23, 2008 18:22
    Subject: [HP3000-L] time like utility?


    > Is there an easy utility to time, wallclock and processor, how long a
    > process takes? I am thinking like the unix "time" utility. Measures
    > down to 100 milliseconds I think.
    >
    > I've gotten some 9gig disks to work as data disks (thought not yet as
    > boot disks...) and the for file access is very startling to me, so
    > much so that I need to test some big file writes and reads on 'em to
    > be sure. They seem to be something like 9 to 10 times faster than poor
    > old internal drive.
    >
    > Now admittedly, these drives *can* perform a whole lot faster than
    > Telzy can drive them (they can and do handle 160mbs), even if they are
    > considered slow and small by today's standards. But that much
    > difference on a 33mhz processor? Even going through a Fast/Wide
    > adapter??
    >
    > -Paul
    >
    > * To join/leave the list, search archives, change list settings, *
    > * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


    * To join/leave the list, search archives, change list settings, *
    * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


  6. Re: time like utility?

    Hi Paul,

    If you're looking for something like the unix "time" utility, why not just use the unix "time" utility?

    :hello user.account

    :sh.hpbin.sys
    ################################################## ######################

    MPE/iX Shell and Utilities (A.50.02)
    COPYRIGHT (c) Hewlett-Packard Company 1992, All Rights Reserved.

    ################################################## ######################


    shell/iX> time date
    Fri Oct 24 14:24:40 EDT 2008

    real 0m00.151s
    user 0m00.031s
    sys 0m00.030s

    [ or an example running an MPE program from the shell using the 'callci' command ]

    shell/iX> time callci "run discfree.pub.sys;info='d'"


    DISCFREE A.75.01 Copyright (C) Hewlett-Packard 1992. All rights reserved.
    FRI, OCT 24, 2008, 11:33 AM

    -------------------------------------------------------------------------------
    TOTALS (IN SECTORS):

    DEVICE SIZE : 782461856
    TRANS SPACE : 787568 PERM SPACE : 577544544
    MAX TRANS SPACE : 773570304 MAX PERM SPACE : 773570304

    FREE SPACE : 204129744
    AVAIL TO TRANS SPACE : 204129744 AVAIL TO PERM SPACE : 195439328


    END OF PROGRAM

    real 0m00.030s
    user 0m00.010s
    sys 0m00.011s


    I hope this helps!

    Take Care,
    Craig

    > -----Original Message-----
    > From: HP-3000 Systems Discussion
    > [mailto:HP3000-L@RAVEN.UTC.EDU] On Behalf Of Paul Raulerson
    > Sent: Thursday, October 23, 2008 6:22 PM
    > To: HP3000-L@RAVEN.UTC.EDU
    > Subject: [HP3000-L] time like utility?
    >
    > Is there an easy utility to time, wallclock and processor,
    > how long a process takes? I am thinking like the unix "time"
    > utility. Measures down to 100 milliseconds I think.
    > -Paul


    * To join/leave the list, search archives, change list settings, *
    * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


  7. Re: time like utility?

    Several people have suggested this now, and you as point it, it works just great. Whodathunkit?

    Thanks Craig & everyone else who suggested this.
    -Paul

    On Friday, October 24, 2008, at 11:37AM, "Fairchild, Craig D" wrote:
    >Hi Paul,
    >
    >If you're looking for something like the unix "time" utility, why not just use the unix "time" utility?
    >
    >:hello user.account
    >
    >:sh.hpbin.sys
    >################################################## ######################
    >
    > MPE/iX Shell and Utilities (A.50.02)
    > COPYRIGHT (c) Hewlett-Packard Company 1992, All Rights Reserved.
    >
    >################################################## ######################
    >
    >
    >shell/iX> time date
    >Fri Oct 24 14:24:40 EDT 2008
    >
    >real 0m00.151s
    >user 0m00.031s
    >sys 0m00.030s
    >
    >[ or an example running an MPE program from the shell using the 'callci' command ]
    >
    >shell/iX> time callci "run discfree.pub.sys;info='d'"
    >
    >
    >DISCFREE A.75.01 Copyright (C) Hewlett-Packard 1992. All rights reserved.
    > FRI, OCT 24, 2008, 11:33 AM
    >
    >-------------------------------------------------------------------------------
    >TOTALS (IN SECTORS):
    >
    > DEVICE SIZE : 782461856
    > TRANS SPACE : 787568 PERM SPACE : 577544544
    > MAX TRANS SPACE : 773570304 MAX PERM SPACE : 773570304
    >
    > FREE SPACE : 204129744
    > AVAIL TO TRANS SPACE : 204129744 AVAIL TO PERM SPACE : 195439328
    >
    >
    >END OF PROGRAM
    >
    >real 0m00.030s
    >user 0m00.010s
    >sys 0m00.011s
    >
    >
    >I hope this helps!
    >
    >Take Care,
    >Craig
    >
    >> -----Original Message-----
    >> From: HP-3000 Systems Discussion
    >> [mailto:HP3000-L@RAVEN.UTC.EDU] On Behalf Of Paul Raulerson
    >> Sent: Thursday, October 23, 2008 6:22 PM
    >> To: HP3000-L@RAVEN.UTC.EDU
    >> Subject: [HP3000-L] time like utility?
    >>
    >> Is there an easy utility to time, wallclock and processor,
    >> how long a process takes? I am thinking like the unix "time"
    >> utility. Measures down to 100 milliseconds I think.
    >> -Paul

    >
    >


    * To join/leave the list, search archives, change list settings, *
    * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


  8. Re: time like utility?

    Craig wrote: The 937LX that you have (it was a 937 right?) has an NIO bus, it is rated at 32mbits/sec with a sustained rate of 20mbits per second.
    "Native I/O" -- meaning the I/O cards (Host Bus Adapters or Device Adapters) communicating using the PA RISC IO ACD definitions. The buses you plugged I/O cards into had characteristics close to what Craig quotes. However later systems had NIO buses that ran up to 100Mhz and where 64/128 bits wide for the CPU & memory communication. The earliest MPE/XL systems used I/O cards that were based upon the MPE Classic I/O cards. So they were termed CIO for (Compatibility I/O or Channel I/O as I/O driver writers did "channel programming" on MPE V) and so required an extra Channel Adapter hardware to convert their CIO signals to NIO for the Memory and CPU -- there wasalso a CAM or Channel Adapter Manager software which managed the CA. So for most system before 9x7 (like 950 "family" and 922/32/48/58 & 49) you'd have a hardware & software stack like:

    CPU <> NIO bus <> CA & CAM <> CIO bus <> DA & DAM <> HBIB bus -> Device & Disk DM

    With 9x7 you had only NIO buses,

    CPU <> NIO bus <> DA & SCSI DAM <> SCSI bus <> Device & SCSI DM

    9x9 & 99x systems were "all" NIO though they had higher speed NIO for CPU/Memory and more layers for I/O expansion capabilities.
    What's a little interesting is that the last generation of HP e3000s we were moving back something like CIO but based upon PCI:

    CPU <> NIO ("runway") <> Ike <> "ropes" high speed serial <> Elroy <> PCI bus <> HBA & PCI-SCSI DAM <> SCSI bus <> SCSI Device & SCSI DM.

    These components are all visible in the I/O paths and you can see the namesfor the "manager" software for each layer in the "pmgr" field in Sysgen IO> LP. There are software managers for the Ike and Elroy but they exist only for some boot time initialization and don't get involved in moving bytes.So they data flow isn't really any more complex from an OS standpoint than it was for 9x7. . .

    Hope you all had a safe and fun Halloween -- spotty rain here but enough time to fill modest buckets for my two little ghosts.

    Jim


    --- On Thu, 10/23/08, Paul Raulerson wrote:

    > From: Paul Raulerson
    > Subject: time like utility?
    > To: HP3000-L@RAVEN.UTC.EDU
    > Date: Thursday, October 23, 2008, 8:22 PM
    > Is there an easy utility to time, wallclock and processor,
    > how long a
    > process takes? I am thinking like the unix "time"
    > utility. Measures
    > down to 100 milliseconds I think.
    >
    > I've gotten some 9gig disks to work as data disks
    > (thought not yet as
    > boot disks...) and the for file access is very startling to
    > me, so
    > much so that I need to test some big file writes and reads
    > on 'em to
    > be sure. They seem to be something like 9 to 10 times
    > faster than poor
    > old internal drive.
    >
    > Now admittedly, these drives *can* perform a whole lot
    > faster than
    > Telzy can drive them (they can and do handle 160mbs), even
    > if they are
    > considered slow and small by today's standards. But
    > that much
    > difference on a 33mhz processor? Even going through a
    > Fast/Wide
    > adapter??
    >
    > -Paul
    >
    > * To join/leave the list, search archives, change list
    > settings, *
    > * etc., please visit
    > http://raven.utc.edu/archives/hp3000-l.html *


    * To join/leave the list, search archives, change list settings, *
    * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *

    * To join/leave the list, search archives, change list settings, *
    * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


+ Reply to Thread