Hello Murray!

In a message written on Wednesday May 17 1995, Murray Lesser insanely
babbles to Paul Sidorsky:

 ML>     The DOS system clock (the PC timer counter) "ticks"
(is updated)
 ML> about 18.2065 times per second (actually 1,573,040 ticks before
 ML> rollover, which is every 86,400 seconds, assuming the clock is
 ML> accurate).  The OS/2 system clock (the "real-time clock")
"ticks"
 ML> exactly 31.25 times per second (same assumption).  These "tick
rates"
 ML> indicate only the granularity of the system clock readings--an
 ML> indication of the significance of the value of a time interval measured
 ML> by two readings of the system clock (same assumption).

We were actually discussing this in OS2DOORS a while back.  The thing I'm
wondering is how you get, say, the 1/10000th of a second precision timings
you see in Turbo Profiler for DOS.  From what I know the computer has a
timing "port" that you can configure to time up to something like
1/100000th of a second.  I can only assume that this port doesn't use the
system clock, but if that's the case then what _does_ it use?  Or are times
not as accurate as they appear to be?

 ML>     Perhaps you misunderstood the meaning of the value given in your
 ML> OS/2 C compiler's time.h header for CLOCKS_PER_SECOND.  This value is an
 ML> artifact of the clock() function implementation in that particular
 ML> language processor, and has nothing to do with the actual system clock
 ML> tick rate.  A couple of examples from DOS C compilers I have owned:  The
 ML> MSC DOS C compiler v 5.1 used a value of 1000 for CLOCKS_PER_SECOND by
 ML> computing the system clock time from the PC timer counter value to the
 ML> nearest millisecond each time the program called for a value of clock().

I see.  I don't know very much about the hardware goings-on, so I usually
only guess at that kind of stuff.  One of these days I'll sit down and
learn all this fancy stuff, but until then I'll have to stumble along.

 ML> On the other hand, the Borland DOS C compiler (v 3.1) effectively
 ML> counted system clock ticks by using the readings directly from the PC
 ML> timer counter; CLOCKS_PER_SECOND for that compiler is given as 18.2.

It's called CLK_TCK in BC++ actually.  :-)

 ML>     In any case, the system clock tick interval has nothing to do with
 ML> how fast the display screen is rewritten.

Well, it was a logical deduction from the facts I had, but when I don't
have all the facts it's hard to get it right.  :-)  Thanks for correcting
me.

 -- PaulS

--- FMail/386 1.0g