Date: Wed, 19 Jan 2000 17:59:38 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Nate Eldredge <neldredge AT hmc DOT edu>
cc: djgpp AT delorie DOT com
Subject: Re: `void main' in Gnu packages---?!
In-Reply-To: <837lh789ti.fsf@mercury.st.hmc.edu>
Message-ID: <Pine.SUN.3.91.1000119175730.13343N-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp AT delorie DOT com
Errors-To: dj-admin AT delorie DOT com
X-Mailing-List: djgpp AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk


On 18 Jan 2000, Nate Eldredge wrote:

> Eli Zaretskii <eliz AT is DOT elta DOT co DOT il> writes:
> 
> > Shawn Hargreaves wrote:
> > > 
> > > Wouldn't this accounting add a lot of additional overhead to these calls,
> > > though? __dpmi_int() in particular is used frequently enough to be a
> > > major performance bottleneck in some situations.
> > 
> > All you need to do is to farpeekb at the BIOS data area where the BIOS tick
> > count is maintained.  I'm sure we can sustain a few simple instructions...
> 
> Didn't see the context of this post, but isn't that exactly what
> `rawclock' does?

Sure.  But the context was how to implement the missing features of 
`getrusage'.  This requires `rawclock' to be called before and after
__dpmi_int calls the DPMI host (to record the time consumed by the
``system'').

Shawn was concerned that this would steal too many CPU cycles.