Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!usc!samsung!uunet!mcsun!hp4nl!philapd!ssp2!pb
From: pb@idca.tds.PHILIPS.nl (Peter Brouwer)
Newsgroups: comp.arch
Subject: Re: yet-another-benchmark
Message-ID: <539@ssp2.idca.tds.philips.nl>
Date: 13 Nov 89 08:22:06 GMT
References: <1989Nov12.160221.26921@ux1.cso.uiuc.edu>
Organization: Philips Telecommunication and Data Systems, The Netherlands
Lines: 64

In article <1989Nov12.160221.26921@ux1.cso.uiuc.edu> ejk@ux1.cso.uiuc.edu (Ed Kubaitis) writes:
>Attached is yet-another-benchmark that might cast some light on aspects
>of architecture.  As with all benchmarks, there is a very serious question 
>of relevance to one's own applications. However, unlike many others, it 
>is small enough to see in detail what is being measured.
>
>   hopalong() {
>      int wc=W/8, cx=W/2, cy=H/2, ix, iy; 
>      double x=0, y=0, xx, yy, t;
>
Is there a reason that no register declarations are used. I think this setup
also introduces an dependency on compilers.
I know lost of people think a compiler should do the register setup but
in my opinion I programmer should always do the thinking. He knows ( at least
should know ) which variables need register declarations. A compiler can
never detect the critical loops in a program and might give a wrong 
variable a register declaration. This is very likely for CPU's that cannot
handle a fair number of register declarations , like the 386 ( Our compiler
can handle 3, one more should be possible).

-- 
Peter Brouwer,                # Philips Telecommunications and Data Systems,
NET  : pb@idca.tds.philips.nl # Department SSP-P9000 Building V2,
UUCP : ....!mcvax!philapd!pb  # P.O.Box 245, 7300AE Apeldoorn, The Netherlands.
PHONE:ext [+31] [0]55 432523, # Never underestimate the power of human stupidity
Newsgroups: yet-another-benchmark
Subject: Re: yet-another-benchmark
Summary: 
Expires: 
References: <1989Nov12.160221.26921@ux1.cso.uiuc.edu>
Sender: 
Reply-To: 
Followup-To: 
Distribution: 
Organization: Philips Telecommunication and Data Systems, The Netherlands
Keywords: 

In article <1989Nov12.160221.26921@ux1.cso.uiuc.edu> ejk@ux1.cso.uiuc.edu (Ed Kubaitis) writes:
>Here is an updated list of results. Thanks for the responses.
>-------------------------------------------------------------------------------
>Attached is yet-another-benchmark that might cast some light on aspects
>of architecture.  As with all benchmarks, there is a very serious question 
>of relevance to one's own applications. However, unlike many others, it 
>is small enough to see in detail what is being measured.
>
>   hopalong() {
>      int wc=W/8, cx=W/2, cy=H/2, ix, iy; 
>      double x=0, y=0, xx, yy, t;
>
Is there a reason that no register declarations are used. I think this setup
also introduces an dependency on compilers.
I know lost of people think a compiler should do the register setup but
in my opinion I programmer should always do the thinking. He knows ( at least
should know ) which variables need register declarations. A compiler can
never detect the critical loops in a program and might give a wrong 
variable a register declaration. This is very likely for CPU's that cannot
handle a fair number of register declarations , like the 386 ( Our compiler
can handle 3, one more should be possible).

-- 
Peter Brouwer,                # Philips Telecommunications and Data Systems,
NET  : pb@idca.tds.philips.nl # Department SSP-P9000 Building V2,
UUCP : ....!mcvax!philapd!pb  # P.O.Box 245, 7300AE Apeldoorn, The Netherlands.
PHONE:ext [+31] [0]55 432523, # Never underestimate the power of human stupidity