Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!rpi!bu.edu!wang!news
From: marcs@slc.com (Marc San Soucie)
Newsgroups: comp.arch
Subject: Re: Page size and linkers
Message-ID: <azzlya.7wk@wang.com>
Date: 28 Jan 91 19:00:55 GMT
References: <1899@cluster.cs.su.oz.au>
Sender: news@wang.com
Organization: Servio Corporation
Lines: 35

Bruce Janson writes:

> Marc San Soucie writes:

> > One of the real goodies, which I have yet to see in a UNIX system, was a
> > service provided by the Kernel (and making liberal use of a nifty 68020/30
> > hardware feature) whereby one could run one's program under the auspices of
> > the debugger, and cause it to emit a trace log of all jumps and calls.

> On the newer RISC chips the instruction sets and their encodings
> are fairly straightforward.
> It's not too much trouble to write a simulator that runs your
> program (albeit slowly), printing out the jump/call info., and
> thereby avoiding the need for os support.
>
> Cheers,
> Bruce Janson
> Basser Department of Computer Science
> University of Sydney
> Sydney, N.S.W., 2006
> AUSTRALIA

Right, but this kind of technique is even less portable than one which relies
on hardware support that could be generic if more hardware provided it. ALso,
the beauty of OS-based analysis support is that it can be performed on-site in
a customer situation. Since customer performance nightmares can seldom be
reproduced back in the manufacturer's lab, on-site analysis is extremely handy.

Also, the OS-based support, with hardware assist, is leagues faster - in our
case about 10% of original speed - than a software simulation would be.

    Marc San Soucie
    Servio Corporation
    Beaverton, Oregon
    marcs@slc.com