e4288271a66fba0fe5cd.newsdawg.com
s2.setup.misc:667
From: "Dariusz Piatkowski"
Hi Doug!
On Wed, 11 Jan 2012 17:36:49 UTC, "Doug Bissett"
wrote:
> On Wed, 11 Jan 2012 04:33:22 UTC, "Dariusz Piatkowski"
> wrote:
>
> > Hi Trevor!
> >
> > On Tue, 10 Jan 2012 21:36:30 UTC, "Trevor Hemsley"
> > wrote:
> >
> > > On Tue, 10 Jan 2012 05:01:30 UTC in comp.os.os2.setup.misc, "Dariusz
Piatkowski"
> > > wrote:
> > >
> > > > This is with "PSD=OS2APIC.PSD /P=6 /APIC" in CONFIG.SYS and the BIOS
set to
> > > > support MPS 1.4 version (which was my previous setting, working OK
with the 4
> > > > core CPU).
> > >
> > > /APIC never got much beyond the beta stage so you might try it without.
> > >
> > > Also, I have got past many issues like you describe by using SET
> > > RESTARTOBJECTS=NO in config.sys
> >
> >
> > Yup, as Lars suggested, I tried dropping the /APIC option, however that
did not
> > appear to improve teh situation.
> >
> > Now, regarding the follow-up suggestion of yours, I attempted the
following
> > (based on some previous Intenet searches):
> >
> > SET RUNWORKPLACE=G:\OS2\CMD.EXE
> > SET RESTARTOBJECTS=NO
> > SET SHELLEXCEPTIONHANDLER=OFF
> > SET SHAPIEXCEPTIONHANDLER=OFF
> >
> > ...the result was a hard-lock however...what I haven't tried was leaving
the full
> > CONFIG.SYS setup as-is in my 4-core configuration and attemptign the SET
> > RESTARTOBJECTS=NO on it's own.
> >
> > I will try that tomorrow and will follow-up with my results.
> >
> > So far, strangely enough (to me anyways) it doesn't appear like I'm going
to
> > have much luck moving to the 6-core configuration.
> >
> > I wish that the eCS ACPI solution could be tried out on some type of a
trial
> > basis...because if it works I would be happy to pay the price and move
along to
> > more happy times...LOL...yes, but I'm just getting 'greedy' here trying to
get
> > the full 6 cores running...
>
> This is actually a very old problem, known as a "startup conflict". It
> can usually be caused/fixed, by simply changing the order of starting
> things in CONFIG.SYS, so that they don't step on each other.
> Unfortunately, it is not possible to say "do this" and have positive
> results (just as often, the results are negative). You need to play
> with things, to work around the problem.
>
> FWIW, one of the main causes of that problem (but not the only one),
> is the same thing that causes looping sounds at startup. The fix for
> that one is to use the mmfix.exe program, which comes with eCS 2.1.
> See "A Note About Repeating System Sounds" at:
>
> > http://svn.netlabs.org/uniaud/
>
> for more information about that.
>
> The latest indication, that I saw, happened after I installed the new
> ACPI package to my T43. It worked perfectly (well as perfectly as ACPI
> has ever worked) with eCS 2.0, but in eCS 2.1, boot stopped at not
> being able to load LANMSGEX.EXE (the message indicated that it
> couldn't read the file). The work around turned out to be to use
> BASEDEV=BOOTDLY.SYS /D:2
> (comes with eCS 2.1) immediately after
> PSD=ACPI.PSD
>
> Previous to that, I have seen many other things that have caused the
> problem, and it still happens intermittently, without changing
> anything. Reboot always works, in that case.
>
> Your problem is probably caused because OS2ACPI.PSD takes a little
> longer to get 6 processors going, or, 6 processors can get other
> things started sooner than was possible with only 4 processors.
>
> You must also remember that all RUN, and CALL statements, in
> CONFIG.SYS, are started at the same time that STARTUP.CMD is executed.
> The programs can step on each other. The work around for that is to
> take things out of STARTUP.CMD, and start things by making an icon in
> the STARTUP folder (you can do the same with RUN and CALL statements
> that are in CONFIG.SYS, but that can cause more trouble than it
> fixes). That delays the program startup slightly, but STARTUP.CMD is
> not stepped on). Of course, then you can have the problem where all
> programs in the STARTUP folder get started at the same time. The fix
> for that is to use REXX AUTOSTART to start the STARTUP folder
> contents, one at a time. Startup may take a few extra seconds, but you
> don't have conflicts (once you get it adjusted properly). I find that
> it is essential to use REXX AUTOSTART if there are more than about 4
> items in the STARTUP folder:
>
> > http://www.7cities.net/~mckinnis/rxautost/index.html
>
> The startup sound is also involved, so turning that off/on can also
> change the timing.
>
> Note that startup conflicts can, and do, cause problems at startup,
> while running, and at shutdown. It all depends on exactly what the
> conflict is.
>
> Hope something here helps...
Alright...so re-boot, after re-boot, after re-boot...I had tried all the
combinations I could think off, including basically stripping down the
CONFIG.SYS to what seemed like just the bare essentials...basically the OS,
-sound -USB -utilities, what remained was the Network stuff.
I had removed all my entries from the regular STARTUP folder and moved them
all
to the XWorkplace startup, which is pre-defined to start up 2 seconds AFTER
the
Desktop populates.
I also dwindled down the HPFS386 cache to 64Meg, tried a run with cache turned
OFF, various VIRTUALADDRESSLIMIT combinations (2048 is what my 4 core system
seems to like) and the OS2APCI with, and w/o the /APIC parameter.
At this point in time I am basically giving up on this...maybe eventually the
news will reach me that eCS ACPI is OK, and then I will give it a try.
I'm curious though if anyone out there is running a 6 core combination I would
appreciate hearing from you.
Thanks again for all your help guys...keep the ideas coming...I am always
happy
to give it another whirl...but for now, I'm ready to get back to normal
computing that doesn't keep me up until the wee 1-1:30am morning hours...LOL!
--- Internet Rex 2.31
* Origin: NewsGuy - Unlimited Usenet $19.95 (1:261/20.999)
|
