Re: Randy downloaded KESYS

On Jun 28, 4:14 am, "Wolfgang Kern" <nowh...@xxxxxxxxxxx> wrote:
Robert Wessel replied:

... M$ file time base is 01.01.1980
Only for FAT. NTFS stores the number of 100ns periods since 1-1-1601.

I see, and I'm not fully through with my NT-reader yet.
But where is this 100nS timer ?
I wont believe that NT let the PIT fire IRQs at ~10MHz rate.


There isn't one. Usually NT, et al, run with no faster than a 1KHz
timer. The stored format is units of 100ns, and that's the format
returned by the basic APIs, although there are functions to convert
those to other formats, including UTC, system time (y/m/d/h/m/s/m) and
DOS.

The old 32 bit formats (both *nix 32 bit number of seconds and DOS
16+16) were clearly too short, both in terms of resolution and the
covered period, so the question was clearly how to use a 64 bit
number. Clearly 1ns would leave too short an overall period, but I
always wondered why 100ns instead of 1us, not that it really matters,
of course.

I would certainly not be surprised if some of those unused bits of
resolution eventually got used, especially if PCs ever develop better
clock hardware (something along the lines of S/370s time-of-day
clock).

.

Relevant Pages

  • Re: [patch 00/21] hrtimer - High-resolution timer subsystem
    ... Right now the primary function of the state is to tell whether the timer ... > Of course if you consider the possibility of including high resolution ... problem requires solving problems in the clock abstraction first, ...
    (Linux-Kernel)
  • Re: [patch 00/21] hrtimer - High-resolution timer subsystem
    ... >> This way the list head is only necessary for the high resolution case. ... > (e.g. wakeup + timer restart). ... Even if the clock ...
    (Linux-Kernel)
  • Re: [PATCH] ktimers subsystem 2.6.14-rc2-kt5
    ... >>The cleanup I pointed out for the posix timer interval timers is pretty ... >>the posix timer structure. ... The rounding to the resolution value is explicitly required by the ... This says a) round to the next resolution, and b) don't allow the resulting timer to expire early. ...
    (Linux-Kernel)
  • Re: [PATCH] ktimers subsystem 2.6.14-rc2-kt5
    ... > a clock with a better resolution. ... > resolution than the physical clock has. ... reaches this count the timer is expired. ... OTOH We could also do this at the timer interrupt: ...
    (Linux-Kernel)
  • Re: VB6 code runs different speeds on different PCs
    ... As you've discovered, the minimum Sleep period can be different on different machines (5 milliseconds on my own WinXP system, but longer than that on others I think). ... You can get whatever "game rate" you want in a closed loop of course, by checking a high resolution timing source. ... If I were you I would use a proper Timer with a higher resolution that the standard VB Timer and run your entire game in the Timer event. ... Admittedly, updating their positions more than once every ten milliseconds does appear to give smoother animation (even though the video frame rate simply cannot draw a new frame at a rate greater than every 10 milliseconds,or so, depending on your display settings) because it tends iron out the differences more. ...
    (microsoft.public.vb.general.discussion)