CACHE386.EXE into Config.sys or Startup.cmd ? - OS2

This is a discussion on CACHE386.EXE into Config.sys or Startup.cmd ? - OS2 ; Anyone got any justified reason why to put CACHE386.EXE into either Config.sys or Startup.cmd? I wonder, I ponder ... Thanks DD...

+ Reply to Thread
Results 1 to 19 of 19

Thread: CACHE386.EXE into Config.sys or Startup.cmd ?

  1. CACHE386.EXE into Config.sys or Startup.cmd ?

    Anyone got any justified reason why to put CACHE386.EXE into either
    Config.sys or Startup.cmd?

    I wonder, I ponder ...

    Thanks
    DD


  2. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    Donald wrote:
    > Anyone got any justified reason why to put CACHE386.EXE into either
    > Config.sys or Startup.cmd?
    >
    > I wonder, I ponder ...
    >
    > Thanks
    > DD
    >


    ISTR that without running it once, the write-back cache wasn't enabled.
    There was a simple way to test: remove from config.sys / startup.cmd,
    reboot. Then, open a prompt, and start it. The 1st time after reboot
    gives another message than subsequent times.

  3. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    I've had this line in my config.sys for years, it works great (<- the reason
    :
    CALL=C:\OS2\CMD.EXE /Q /C C:\IBM386FS\CACHE386.EXE >NUL

    regards,
    - dink

    On 8 Feb 2005 16:39:01 -0800, Donald wrote:

    :Anyone got any justified reason why to put CACHE386.EXE into either
    :Config.sys or Startup.cmd?
    :
    :I wonder, I ponder ...
    :
    :Thanks
    : DD
    :




  4. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    Still doesn't answere the question.
    I heard that the author of the original HPFS FS strongly suggested
    CACHE.EXE to be frun from Startup.cmd, rather than from the IBM
    recommended Config.sys.
    So I wonder if that's still true for CACHE386.EXE...


  5. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    Donald wrote:
    > Still doesn't answere the question.
    > I heard that the author of the original HPFS FS strongly suggested
    > CACHE.EXE to be frun from Startup.cmd, rather than from the IBM
    > recommended Config.sys.
    > So I wonder if that's still true for CACHE386.EXE...
    >

    Do you remember the why of him suggesting that? Knowing that would
    probably make it easier to know if the same applies cache386 (might
    still be unknown but better chance).
    Andy

  6. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    Donald wrote:
    > Still doesn't answere the question.
    > I heard that the author of the original HPFS FS strongly suggested
    > CACHE.EXE to be frun from Startup.cmd, rather than from the IBM
    > recommended Config.sys.
    > So I wonder if that's still true for CACHE386.EXE...


    Hard to say. I've used CACHE and CACHE386, on different computers, both
    of them ran either from startup.cmd or CONFIG.SYS, and saw no difference
    and no problems in any case. But I use CONFIG.SYS to start them both,
    since I stopped using statup.cmd a long time ago.
    For me, both CACHE and CACHE386 always work well when started from
    CONFIG.SYS (since 1994 ;-).

    Cheers,
    Martin

  7. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    He claimed that HPFS.IFS took a long time to propperly initialize and
    thus CACHE.EXE might catch it "half way through" and then fail to
    properrly communicate with the file system from there on.


  8. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    [A complimentary Cc of this posting was sent to
    Donald
    ], who wrote in article <1107960857.928784.56460@z14g2000cwz.googlegroups.c om>:
    > Still doesn't answere the question.
    > I heard that the author of the original HPFS FS strongly suggested
    > CACHE.EXE to be frun from Startup.cmd, rather than from the IBM
    > recommended Config.sys.


    There is no need to run CACHE.EXE if you run PMSHELL. The server
    thread is started by PMSHELL (incredible! But [IIRC] Scott claimed so...).

    The only need to run CACHE.EXE is for text-mode boot.

    Hope this helps,
    Ilya

  9. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    Donald wrote:
    > Still doesn't answere the question.
    > I heard that the author of the original HPFS FS strongly suggested
    > CACHE.EXE to be frun from Startup.cmd, rather than from the IBM
    > recommended Config.sys.
    > So I wonder if that's still true for CACHE386.EXE...
    >

    AFAIK CACHE.EXE and CACHE386.EXE are two completely different animals,
    one written mainly by IBM, one mainly by M$. So I think what counts for
    the one, might not count for the other.
    I've always ran CACHE386.EXE from CONFIG.SYS, on Model 80's up to
    NetFinity's. Never saw problems with that.

  10. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    Jeroen Besse wrote:
    > Donald wrote:
    >
    >> Still doesn't answere the question.
    >> I heard that the author of the original HPFS FS strongly suggested
    >> CACHE.EXE to be frun from Startup.cmd, rather than from the IBM
    >> recommended Config.sys.
    >> So I wonder if that's still true for CACHE386.EXE...
    >>

    > AFAIK CACHE.EXE and CACHE386.EXE are two completely different animals,
    > one written mainly by IBM, one mainly by M$.

    What's funny about that is that I have always understood that M$ wrote
    hpfs and IBM wrote hpfs386 but the reason that hpfs386 is so hard to
    come by is that M$ gets a very high license fee for each copy.

  11. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    Donald wrote:
    > He claimed that HPFS.IFS took a long time to propperly initialize and
    > thus CACHE.EXE might catch it "half way through" and then fail to
    > properrly communicate with the file system from there on.
    >

    I can't argue with that as I don't know but it seems odd. I would think
    that by the time that the cache386 line was gotten to in config.sys the
    driver would be fully initiallized (but then I would think it would be
    fully initiallized before it passed to the next command, even the
    regular hpfs). Come to think of it there is a line that says hpfs386 is
    loaded so I would expect it to be fully initialized at that point.

  12. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    Andy Willis wrote:
    > Jeroen Besse wrote:
    >
    >> Donald wrote:
    >>
    >>> Still doesn't answere the question.
    >>> I heard that the author of the original HPFS FS strongly suggested
    >>> CACHE.EXE to be frun from Startup.cmd, rather than from the IBM
    >>> recommended Config.sys.
    >>> So I wonder if that's still true for CACHE386.EXE...
    >>>

    >> AFAIK CACHE.EXE and CACHE386.EXE are two completely different animals,
    >> one written mainly by IBM, one mainly by M$.

    >
    > What's funny about that is that I have always understood that M$ wrote
    > hpfs and IBM wrote hpfs386 but the reason that hpfs386 is so hard to
    > come by is that M$ gets a very high license fee for each copy.


    IIRC M$ did HPFS layout (Gordon Letwin), and prepared assembler version
    of the HPFS.IFS. However IBM rewrote HPFS.IFS in C, that's problably why
    IBM doesn't need to pay license fee per copy. OTOH, HPFS386 is known for
    "hand-optimized assembler code" and "tight assembler loops" so I sense a
    M$ code inside, the reason for high license fee per copy.

    HPFS386 even existed for LanManager server version of OS/2 1.x, 16-bit
    286 version.

    Cheers,
    Martin

  13. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    On Wed, 09 Feb 2005 17:24:46 -0700, Andy Willis
    wrote:
    >> AFAIK CACHE.EXE and CACHE386.EXE are two completely different animals,
    >> one written mainly by IBM, one mainly by M$.

    > What's funny about that is that I have always understood that M$ wrote
    > hpfs and IBM wrote hpfs386 but the reason that hpfs386 is so hard to
    > come by is that M$ gets a very high license fee for each copy.


    Other way round... Microsoft originally designed and wrote HPFS386. Hence
    their continuing rights to royalties. HPFS presumably is exempt from the
    royalties, either because the royalties are specific to the implementation
    and not the design (and HPFS.IFS is hence not covered), or else the joint
    properties agreement between IBM and MS simply had it that way.

    AIUI, anyway.
    --
    Alex Taylor
    http://www.cs-club.org/~alex

  14. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    I really don't understand what you're saying here?
    I'm 100% positive that w/o me running CACHE386.EXE from either
    Config.sys or Startup.cmd
    it just won't start up PERIOD.
    It most certainly isn't started by itself by PMSHELL.
    I know that for a fact (I checked with "ps")!


  15. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    Well, in that case I think its time for that beast to go back into
    Config.sys.

    Thankz!

    DD


  16. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    [A complimentary Cc of this posting was sent to
    Donald
    ], who wrote in article <1108049188.579375.56860@g14g2000cwa.googlegroups.c om>:
    > I really don't understand what you're saying here?


    I'm afraid I was very clear...

    > I'm 100% positive that w/o me running CACHE386.EXE from either
    > Config.sys or Startup.cmd
    > it just won't start up PERIOD.


    Who cares? I did not discuss CACHE386.EXE. While I *was* discussing
    CACHE.EXE, I did not claim CACHE.EXE was started.

    > I know that for a fact (I checked with "ps")!


    PS has no way to distinguish which of the threads running on the
    system is a lazy-write thread... Per Scott's statement, it may be
    started either by CACHE.EXE, or by PMSHELL.EXE.

    Hope this helps,
    Ilya



  17. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    Alex Taylor wrote:
    > On Wed, 09 Feb 2005 17:24:46 -0700, Andy Willis
    > wrote:
    >
    >>>AFAIK CACHE.EXE and CACHE386.EXE are two completely different animals,
    >>>one written mainly by IBM, one mainly by M$.

    >>
    >>What's funny about that is that I have always understood that M$ wrote
    >>hpfs and IBM wrote hpfs386 but the reason that hpfs386 is so hard to
    >>come by is that M$ gets a very high license fee for each copy.

    >
    >
    > Other way round... Microsoft originally designed and wrote HPFS386. Hence
    > their continuing rights to royalties. HPFS presumably is exempt from the
    > royalties, either because the royalties are specific to the implementation
    > and not the design (and HPFS.IFS is hence not covered), or else the joint
    > properties agreement between IBM and MS simply had it that way.
    >
    > AIUI, anyway.

    HPFS386 was written by Gordon Letwin (one of the first 10 employees of
    Microsoft) in the late 80's. It is composed of hand-tuned, pure assembly
    code. Gordon Letwin was one of the lead designers and programmers of
    OS/2 on the Microsoft side in the beggining years of OS/2. He was the
    one that came up with the genius way to switch the 286 from protected
    mode to real mode by a "hot reset"!

    Until Bootable JFS arrives with eComStation v2.0, HPFS386 is still the
    most efficient and fastest filesystem for OS/2 and eComStation boot volumes!

    Cheers!

  18. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    >did not discuss CACHE386.EXE. While I *was* discussing
    >CACHE.EXE, I did not claim CACHE.EXE was started.

    That doesn't make any sense and most certainly didn't help in a thread
    titled "CACHE386.EXE"

    You wrote:
    >There is no need to run CACHE.EXE if you run PMSHELL. The server
    >thread is started by PMSHELL


    And I say that PS proves this not to be the case, as simple as that.

    DD


  19. Re: CACHE386.EXE into Config.sys or Startup.cmd ?

    [A complimentary Cc of this posting was sent to
    Donald
    ], who wrote in article <1108414414.527721.89330@c13g2000cwb.googlegroups.c om>:
    > >did not discuss CACHE386.EXE. While I *was* discussing
    > >CACHE.EXE, I did not claim CACHE.EXE was started.


    > That doesn't make any sense and most certainly didn't help in a thread
    > titled "CACHE386.EXE"


    Eh??? Who do you think first mentioned CACHE.EXE in this thread???

    > >There is no need to run CACHE.EXE if you run PMSHELL. The server
    > >thread is started by PMSHELL


    > And I say that PS proves this not to be the case, as simple as that.


    As I explained in my previous message, this means that you do not know
    what you are talking about.

    Best regards,
    Ilya

+ Reply to Thread