> > On Thu, 3 Sep 1998, Cameron Kaiser wrote: > > > Also, since TI$ counts jiffies, it *should* be accurate (except for the fact > > that it counts to 240000 and THEN rolls over to 000000, so you need to account > > for an extra second per diem) ... unless there's cause for the 60Hz IRQ rate > > to be fouled? > > If you load something from the floppy, there are a lot of SEIs in the way, > causing a slowdown of TI/TI$. The TOD clock is not influenced by SEI/CLI, > though. Well, that won't cause any problems -- the CIA timer (the interrupt timer) runs in continuous mode. So although there might be some little fluctuations, after 1000 interrupts the TI$ counters will have been incremented 1000 times. (Kind of like the way the AC line is only 60Hz or 50Hz on average). On the other hand, I think the tape routines change the IRQ vector, and of course anything that changes the timer value or skips the update will mess up the clock. Let's see now... the interrupt timer uses a latch value of 17045. This means an interrupt takes place every 17046 cycles * 14/14318181 sec/cycle = 0.0166672 seconds (on an NTSC machine). This is about 1/59.998 seconds, or 1/60 second + 5.34*10^-7 seconds. This means that every 1/5.34*10^-7 or every 1871657 jiffies, TI accumulates a little error; this corresponds to 1/60 second every 8.6 hours or so. So, TI seems to stay pretty accurate (I always had it in my head that it wasn't). -Steve - This message was sent through the cbm-hackers mailing list. To unsubscribe: echo unsubscribe | mail cbm-hackers-request@dot.tcm.hut.fi.
Archive generated by hypermail 2.1.1.