50G "speed bumps" again
arturo
using the 50 as a stopwatch for short time intervals.
( see 50g speed bumps ).
the problem is basically that there is about a 1 second delay
from when a key is pressed before it does anything in about
1 out of 30 to 40 times the key is pressed. most of you
probably don't notice it and wouldn't unless you were timing
short intervals. MEM DROP did not help and it seems to have
nothing to do with what the key was assigned to do- also
nothing to do with keytime setting or interval of time between
key presses.
someone suggested the 50 was "multitasking" but the fact that
these bumps can come as often as 4 to 5 times in 20 key presses
indicates otherwise. there does not seem to be any way to force
this delay either.
if you set yor 50 to "click" when a key is pressed you won't get
the click or activity annunciator for at least a second when this
occurs.
for these reasons and the fact that this never happened on my 48's
(never owned a 49), this bump appears to me to be a "bug"- that is
to say a mistake. i also have noted that key strokes made in this
1 second pause sometimes register and sometimes don't. this
is a problem on both of my new 50s. if this were some multitasking
function it should have been made so that it could be forced like
MEM DROP does for "garbage collection".
i'm no longer looking for solutions and will live with it. the 50s are
great overall and i'm not trashing them. my point is to describe the
problem in hopes that those who create future roms or models (HP)
can fix a problem that did not exist previously.
thanks, Arturo
manjo
> (never owned a 49), this bump appears to me to be a "bug"- that is
> to say a mistake. i also have noted that key strokes made in this
> 1 second pause sometimes register and sometimes don't. this
> is a problem on both of my new 50s. if this were some multitasking
> function it should have been made so that it could be forced like
> MEM DROP does for "garbage collection".
>
> i'm no longer looking for solutions and will live with it. the 50s are
> great overall and i'm not trashing them. my point is to describe the
> problem in hopes that those who create future roms or models (HP)
> can fix a problem that did not exist previously.
>
> thanks, Arturo
Now when you descirbed the "bump" that way...
this bump is noticable in latest ROMs which fixed the missed keys issue,
i can only speculate based on indications and my own investigations of my
own.
If they used some way to synchronize the arm's keyboard state with emulated
Saturn's keyboard scanning.
(Note: some way of syncronisation is needed because it is possible that
ARM's keyboard status changes
beween two emulator key scans thus keypress would not be detected by
emulator)
Depending on implementation it may be possible that it 'locks' or 'delays'
the keyboard scanning for some time,
this time may be related to KEYTIME value. You could try to set KEYTIME to
zero if it affects the 'bump' this would
promote my speculations to educated guesses or even better :-)
manjo
arturo
> Depending on implementation it may be possible that it 'locks' or 'delays'
> the keyboard scanning for some time,
> this time may be related to KEYTIME value. You could try to set KEYTIME to
> zero if it affects the 'bump' this would
> promote my speculations to educated guesses or even better :-)
i've tried keytime 0 to 2000 with no effect.
it's not that the delay needs to dissapear but if it could be forced,
a la MEM DROP, that would be a solution. like i said i can certainly
live with it and just want to point it out in case it can ever be
controlled.
Arturo
John H Meyers
> the problem is basically that there is about a 1 second delay
> from when a key is pressed before it does anything in about
> 1 out of 30 to 40 times the key is pressed.
> if you set your 50 to "click" when a key is pressed you won't get
> the click or activity annunciator for at least a second when this
> occurs.
> for these reasons and the fact that this never happened on my 48's
> (never owned a 49), this bump appears to me to be a "bug"- that is
> to say a mistake. i also have noted that key strokes made in this
> 1 second pause sometimes register and sometimes don't.
If the delayed keystroke registers when the *next* key is pressed,
then it's the "busy bug," as may previously have been mentioned;
I've seen no fix ever reported for that bug, which is highly likely
to reside in manufacturer Kinpo's ARM OS (the emulator and
hardware layer), rather than in the HP OS (which runs
on the emulator that hides within).
Otherwise it's another bug, and if you're the first who noted it,
you can name it yourself, like a comet :)
http://seds.lpl.arizona.edu/sl9/sl9.html
[r->] [OFF]
werty
You are niave ! HP had a scam , they knew people would not upgrade
to a new box , if the 48GX worked .
So they had to promise a little extra speed or functions .
But it was all lies .
The 48GX will work much harder than the 49 or 50 .
There are only broken promises in 49 , 50 .
Its marketing .. They will never give more , when they can
control the demand ( laws , copyrights ) .
They and the government have a TOTALLY captive market .
But theres still enough competition in the world to evolve a
pocket PC that has everything ,
HP will be too Little , Too Late . There is free s/w coming ,
on the ARM cpus , that will be trully open source , you will be able
to hack in seconds from softkeys ( as the 48GX has on top )
Unlike Linux "open source" ....ha ha he he ha ha ha ...
It will kill WXP , Linux , C/C++ , GNU , and Bill Gates will
be out of the s/w business , in spite of govt efforts to stop it !!
Wanna be part of it ? Just ask for concepts and i'll do my best .
The first concepts are , it has absolute rules , no exceptions .
You can only program this OpSys with a hardware key . If key
is not "in" , you cant install any s/w , you can only send kenel
requests . You can simulate programming with scripts , later compile
in 3 seconds , using a h/w key to open and boot OpSys .
Everything is integrated into the kernel , thus it "knows" everything
and can multi-task w/ reentrancy .
Kernel NEVER executes threads as belonging to others .
All threads are "part" of the kernel . making coding easier , for
you cant write code that does not need to know the kernel .
You cant write a priority ONE driver , you must integrate it into
the kernel such it knows NOT what priority it has !
This is the failure of Linux . It runs threads that has authority .
CRASH . The kernel must always have authority to figure what
needs what at all times . One can not serve 2 masters !!
The kernel is infinately knowledgable about the h/w ,
so if you try to code a device driver , the kernel debugger
will 'show" you
real time , the DD will run out of time at "T-23 " for the core
functions
of the kernel need that time slot !!!
You cant force my OpSys to crash , not even with low level code !
And this is the fastest method to teach the latest software methods
.....
by doin .., not by reading a manual ...
it will be FREE .. , all you must do is get off ur lazy arse , study
it ..
#####################################################################
Joe Horn
> the problem is basically that there is about a 1 second delay
> from when a key is pressed before it does anything in about
> 1 out of 30 to 40 times the key is pressed.
You're experiencing the Busy Bug, which happens whenever a key is
pressed at exactly the same moment that the busy annunciator turns off.
The next system interrupt causes the delayed key to get executed.
Your observation of a 1-second delay reveals that you have the "Ticking
Clock Mode" turned on (flag -40 is set). Ticking Clock Mode causes a
system interrupt every second (even when HEADER < 2 which makes the
clock invisible), and THAT interrupt is what's causing your delayed
keystroke to get executed. If you turn the ticking clock off (-40 CF),
the delayed key will remain in the key buffer until the next keystroke.
Leaving the clock on is probably less annoying.
-Joe-
John H Meyers
The original song could use a one-second interrupt as well:
http://pages.sbcglobal.net/casaverde/Jody/2_Worried.html
arturo
> You're experiencing the Busy Bug, which happens whenever a key is
> pressed at exactly the same moment that the busy annunciator turns off.
thing is i will get this delay many seconds after the busy annunciator
has
dissapeared. i am not making a big deal out of this but suspect you
will
be hearing more about it unless both my 5o's (sequential serial no.s)
have
their own little problem.
ps- with -40 flag clear same delay occurred with same frequency.
happy holidays, Arturo
Joe Horn
> thing is i will get this delay many seconds after the busy annunciator
> has dissapeared.
Does this ever happen when the "ticking clock" is turned off? If so,
are you taking into account the fact that the busy annunciator turns on
(often very briefly) with EVERY keystroke?
-Joe-
Veli-Pekka Nousiainen
news:1166857685.2...@42g2000cwt.googlegroups.com...
BTW: is there a difference with User Keys on or off?
How about Vectored Enter?
ALGebraic?
arturo
> Does this ever happen when the "ticking clock" is turned off? If so,
> are you taking into account the fact that the busy annunciator turns on
> (often very briefly) with EVERY keystroke?
yes it does Joe- if by "ticking clock" off you mean either -40 flag
clear
or key "ticks" turned off.
most won't even know it is happening except those who
turn "key click" on in which case you will not hear the click
for 1 second or so after a key is pressed or others like myself
using the 50 to time short intervals (if any).
if you are interrested in checking this out more you could try to
reproduce it.
<< CLEAR TIME 1 1 + DROP TIME SWAP HMS- 1000000 X MEM DROP >>
either assign this to a key or use soft key. the program times how long
it takes the calc to add 1 + 1 and drop the sum in hundredths of a
second.
either way press the key about once a second (plenty of time for the
busy annunciator to turn off) until you get the delay (watch the
annunciator)
the strange thing is in the 1 in 40 or so times i get the delay on key
stroke
it also takes like 8 times longer for the calc to process 1 + 1 DROP??
in fact if no one can reproduce this then it is a particular problem
with the
2 new 50's i have.
thanks, Arturo
Joe Horn
> press the key about once a second (plenty of time for the
> busy annunciator to turn off) until you get the delay (watch the
> annunciator) ...
> in fact if no one can reproduce this then it is a particular problem
> with the 2 new 50's i have.
Ow! Ow! This is hurting my one remaining brain cell! >:-O (Just call
me Amoeba Man.)
I ran your program many hundreds of times, and I can only reproduce the
delay when flag -40 is set. It NEVER delayed when flag -40 was clear.
Are you REALLY REALLY SURE that you see the delay even after EXPLICITLY
clearing flag -40? It's not enough to set HEADER to 1 or 0; you must
actually clear flag -40. Is something in your machine messing with
flag -40, perhaps? Something like 'STARTUP' or 'EXITED' or
something???
Do you have any freaky libraries installed which cause timer
interrupts? Can you reproduce the delay after backing up your machine
to port 2 or 3 and then performing a Memory Clear and then warmstarting
with the backspace key held down (which prevents libraries from
attaching)? If it happens in a totally empty machine with default
settings, then EITHER it's not the Busy Bug... OR the Busy Bug is worse
than I thought.
'Tis a puzzlement!
-Joe-
Veli-Pekka Nousiainen
"arturo" <surv...@bellsouth.net> wrote in message
news:1166888591.3...@48g2000cwx.googlegroups.com...
I don't what's up to
but I get 1/40 a HALF time - that is: it's takes only half time to proceed
I have USER off, keyclick off, clock tivking, LASTARG off, Last Stack Off
Multiplier 1E7 gives me 6.7139-7.0801, half times are 3.4180-4.3945
Veli-Pekka Nousiainen
With the clock off I got timins between 6.7138-7.2022
What are we chasing?
Veli-Pekka Nousiainen
news:Cjjjh.8051$EE7....@reader1.news.saunalahti.fi...
X
> I don't what's up to
> but I get 1/40 a HALF time - that is: it's takes only half time to proceed
> I have USER off, keyclick off, clock tivking, LASTARG off, Last Stack Off
> Multiplier 1E7 gives me 6.7139-7.0801, half times are 3.4180-4.3945
After about 1/thousand tries I finally got it: about 10* timing!
50.6591 instead of 5.9814-6.5918 (new code)
Now someone please - explain about the "half-times"
arturo
> Ow! Ow! This is hurting my one remaining brain cell! >:-O (Just call
> me Amoeba Man.)
>
> I ran your program many hundreds of times, and I can only reproduce the
> delay when flag -40 is set. It NEVER delayed when flag -40 was clear.
> Are you REALLY REALLY SURE that you see the delay even after EXPLICITLY
> clearing flag -40? It's not enough to set HEADER to 1 or 0; you must
> actually clear flag -40. Is something in your machine messing with
> flag -40, perhaps? Something like 'STARTUP' or 'EXITED' or
> something???
Joe- my bad, yes there was "something???"
when i think of you folks trying hundreds of times with this to help
out a stranger
i am both very gratefull and smacked by guilt. the guilt is because Joe
gave me the
solution with flag -40 and all those trial runs you made weren't
needed. what i didn't
realize was that, even tho my own reset program had been revised to
clear -40, the
program i use to time intervals had its own STOF (store flag) in it
which set -40 (DUH).
the -40 CF did work for the test program i posted above- that is it
eliminated the
speed bump, but then i would try my "real" timer program, which would
set -40,
they were back. for some reason my "real" timer program with HALTs,
KILLs and
0 WAITs seemed to aggrivate the problem as i was hitting this 'bump"
sometimes as often
as 4 times in 20 key hits (and then disapear for a hundred hits).
Joe, John, Manjo, Veli-Pekka Nousiainen and others (Eric at hpcalc)-
thank you. my
program is used to win at roulette- a very tuff task on modern wheels.
every bit of an
edge is needed so that "bump" wasn't just a minor inconvenience.
i could NEVER have figured to CF -40 on my own.
after the holidays i am going to post a short description of the
algorithm to use for the
balls motion at roulette. it is the only thing i have to offer in
return for the help i have
received. many smarter people than myself have overlooked the simple
solution to
model a ball that decelerates at different rates thruout a spin.
my Christmas came a day early- thanks, Arturo
Veli-Pekka Nousiainen
X
> set -40,
X
> my Christmas came a day early- thanks, Arturo
In Finland it's 24.12 Xmas eve perhaps 16:00 for smaller kids
--
Veli-Pekka
PS: Merry Christmas!
(we don't have to say Happy Holidays here :-D )
BUT
also Happy Hanukka!