Path: utzoo!mnetor!uunet!husc6!think!ames!pasteur!ucbvax!dewey.soe.berkeley.edu!robinson
From: robinson@dewey.soe.berkeley.edu (Michael Robinson)
Newsgroups: comp.arch
Subject: Re: conditional branches
Message-ID: <23064@ucbvax.BERKELEY.EDU>
Date: 20 Feb 88 03:33:02 GMT
References: <191@telesoft.UUCP> <1556@gumby.mips.COM> <375@imagine.PAWL.RPI.EDU> <1610@gumby.mips.COM> <4252@aw.sei.cmu.edu> <7445@apple.UUCP>
Sender: usenet@ucbvax.BERKELEY.EDU
Reply-To: robinson@dewey.soe.berkeley.edu.UUCP (Michael Robinson)
Organization: School of Education, UC-Berkeley
Lines: 33

In article <7445@apple.UUCP> bcase@apple.UUCP (Brian Case) writes:
>(Here we go again...)  There is a paper by Robert D. Russell, "The PDP-11:
>A Case Study of How _NOT_ To Design Condition Codes," that, although for
>a different architecture, says exactly the same thing.  Excerpts from this
>old, old paper:
>[...]
>This means that the side
>effect of setting the condition codes is wasted in 512 (92%) of the 555
>instructions whose primary function is not condition code manipulation.
>(Of the 43 that did set the condition codes for later use, 36 (84%) were MOV
>instructions.)

But:

>Of course, these are measurements of "static" instruction
>sequences, not a dynamic trace during execution.

Maybe I'm just stupid, but aren't the dynamic measurements far more
significant?

In my own assembly language programming, my experience has been that 
condition-code-setting instructions are used to advantage most often in
tight, processor-intensive loops (copies, sorts, searches, etc.)--the type
of instructions, therefore, that tend to get executed with much greater 
frequency than usual code.

Surely, someone out there must have done dynamic analysis of condition
code utilization.  I won't be convinced on this issue until I've seen the
results.

------------------------------------------------------------------------------
Michael Robinson                              USENET:  ucbvax!ernie!robinson
                                              ARPA: robinson@ernie.berkeley.edu