Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!lll-tis!ames!amdcad!tim
From: tim@amdcad.AMD.COM (Tim Olson)
Newsgroups: comp.lang.c
Subject: Re: Bad optimizations (was Re: comma operator)
Message-ID: <20060@amdcad.AMD.COM>
Date: 21 Jan 88 10:10:37 GMT
References: <3819@sigi.Colorado.EDU> <5080013@hpfcdc.HP.COM> <7120@brl-smoke.ARPA> <269@vsi1.UUCP>
Organization: Advanced Micro Devices, Inc., Sunnyvale, Ca.
Lines: 32
Keywords: optimization swap assembly C sun compiler
Summary: stop doubting....

In article <269@vsi1.UUCP>, steve@vsi1.UUCP (Steve Maurer) writes:
|  This isn't really Sun's fault, I might add.   Most compilers would do
|  just about the same, or perhaps worse.  Of course, an extrodinarily
|  optimized compiler (one in which i = 5; i = 6; would delete the i = 5;
|  statement altogether), might approximate the original intent of the
|  code, but I doubt it.
|  
|  					Steve Maurer

Well, you can stop doubting.  Here is the .s file generated by
the MetaWare Am29000 compiler (minus procedure prologue and epilogue):


	call	lr0,_foo_marker


As you can see, there are no swaps whatsoever.  This is because they
weren't needed.  I then modified the source to pass 'a' and 'b' to the
function 'foo_marker', so that the values were required.  The resulting
code is [with foo_marker(a,b)]:


	const	lr3,3			<-- pass '3' as second parameter
	call	lr0,_foo_marker
	const	lr2,5			<-- pass '5' as the first
					    (delayed branch)
	
( Ain't RISC machines + optimizing compilers great! ;-)

	-- Tim Olson
	Advanced Micro Devices
	(tim@amdcad.amd.com)