Other Directory Sites: SeekWonder | Directory Owners Forum
The following archive was created by hippie-mail 7.98617-22 on Sat Apr 10 1999 - 01:10:55
Enzo Michelangeli (em@who.net)
Sat, 1 Aug 1998 08:35:29 +0800
-----Original Message-----
From: Ben Laurie <ben@algroup.co.uk>
Date: Saturday, August 01, 1998 12:01 AM
>Just thought I'd explain the 50/60 ms thing seen in PC timings for those
>who don't know (note that the 40ms Enzo reported didn't actually occur
>in the data).
>
>50/60 ms is a consequence of the standard PC tick rate, which is the
>maximum divisor for the clock, yielding 18.2 ticks/second, or very close
>to 55 ms. Clearly something else is also rounding to 10 ms intervals,
>thus yielding a neat mixture of 50 and 60 ms intervals, in more-or-less
>equal proportions (in fact 50 ms should be just a teeny bit more
>common).
Yes, but additional resolution could be gained if the Java Virtual Machine
(System.currentTimeMillis() is a native method) read the registers of the
counter/timer present in every PC. This is what was done by many old DOS
applications; under protected mode, of course, you need some form of OS
support, either through a specific API, or by virtualization of the hardware
registers. Both Linux and WinNT seem to provide that support: the returned
values change every millisecond, not every 40 or 50. So probably the fault
lies in the limited support provided by the Win32 implementation in
Win95/98, not in the Win32 JVM itself.
Cheers --
Enzo
The following archive was created by hippie-mail 7.98617-22 on Sat Apr 10 1999 - 01:10:55