> Des Bromilow writes:
> >My thesis was based on writing a virtual instruments package under win =
> >3.x/95.
> >I could get the 55ms interupt operable, but anything faster than that is =
> >impossible under the normal PC architecture.
>
>         This pretty well goes along with what I have found in DOS.

I really don't get this.  From what you guys are saying a serial port
set over 300bps will have problems with lost data...

In around 1990 I wrote a piece of BBS software.  It had a new terminal
protocol and part of it was "music", set a frequency and duration and
it lumped it onto the play queue.  The timing was in 1/200ths of a
second.  All I did was change the system clock to 200Hz and ran the
"music queue" out of the timer interrupt.  (This 200Hz timing is no
coincidence as it perfectly divided into the old timer frequency.)

This was running on 286s at the time.  Even then I could easily bump
the timer frequency up to a few thousand Hz without any serious
consequences.

If you want to do timing on the PC, it's laughable that you can't get
better than 55ms.  Generate an interrupt, from the ISR, read the timer.
You can get pretty good accuracy out of this.  Don't count timer ticks.
That's just silly.  Seriously high accuracy and resolution?  Maybe not,
for that get dedicated hardware.  The National Inst. boards I use in
the lab can quite accurately count intervals with 1/20MHz resolution.

In the earlier post I wrote about sub 1ms latency I was referring to the
amount of time that it takes my software to _get_ the data from the
aquisition boards and _respond_ with output.  This is under _Windows_
with real time updates of the display!  You can get a ridiculous amount
of work done on a PC.  A 133MHz Pentium (recently retired) showed
25% CPU usage.  The 450MHz P3's are almost wasted showing ~5% CPU usage.

No where near enough research was done on the aforementioned paper.

Chris
(Under Windows the <1ms latency could actually be quite a bit better,
but I'm only sampling AD data at 1,000Hz and I'm measuring latency in
AD clocks.)

P.S. Sorry if I seem "excited" but I can't believe what I'm reading.