Re: A86: Interrupts

To: assembly-86@lists.ticalc.org
Subject: Re: A86: Interrupts
From: Nuno Joao <nmasj@camoes.rnl.ist.utl.pt>
Date: Wed, 09 Dec 1998 18:38:01 +0100
Delivered-To: assembly-86-outgoing@towerguard.unix.edu.sollentuna.se
References: <3a032780.366e0d1e@aol.com>
Reply-To: assembly-86@lists.ticalc.org


ZeromusMog@aol.com wrote:
> > >Are you sure about this? As I recall, when I checked this, the interrupt on
> > >my TI-86 (no user-interrupts or anything installed) ate almost 1/3 of the
> > >CPU cycles. . .
> >
> > Yes, I'm pretty sure about this... I would estimate the interrupt to
> > use about 1/100 or so of the CPU cycles.

> Great. Now I'm confused. I've been told, in conversations about just how much
> 6 megahertz is, that with interrupts, you can't get "nearly as many t-states"
> as when the interrupts are disabled. Yet Jimmy Mardell, Mr. Programming Guru
> here, says that it doesn't take that many cycles... help! My brain's gonna
> explode!

  Well, I went to 86 Central and counted the cycles of the ROM
  interrupt code they have there (in section "Interrupts"). It's about
  660 cycles, executed about 200 times a second. This count excludes 3
  or four routines that are called 'once in a while', like the run
  indicator.

  660 cycles * 200 Hz = 132000 cyles/sec => 2,2% of CPU time

  This is only a reference value, as it was not mesured accuratly.

  NSJ aka Viriato
  l41324@alfa.ist.utl.pt
  nmasj@camoes.rnl.ist.utl.pt
  http://camoes.rnl.ist.utl.pt/~nmasj - DemoAdict/TaradoPorDemos

References:

Re: A86: Interrupts

From: ZeromusMog@aol.com