Re: Possible data allocation and instruction scheduling algo...

bryan.hayes@hayestechnologies.com (Bryan Hayes)
30 Jan 2002 20:39:58 -0500

          From comp.compilers

Related articles
Possible data allocation and instruction scheduling algo... ollanes@prodigy.net (Orlando Llanes) (2002-01-24)
Re: Possible data allocation and instruction scheduling algo... bryan.hayes@hayestechnologies.com (2002-01-30)
Re: Possible data allocation and instruction scheduling algo... bear@sonic.net (Ray Dillinger) (2002-02-06)
Re: Possible data allocation and instruction scheduling algo... ollanes@pobox.com (2002-02-16)
Re: Possible data allocation and instruction scheduling algo... ollanes@pobox.com (2002-02-16)
| List of all articles for this month |
From: bryan.hayes@hayestechnologies.com (Bryan Hayes)
Newsgroups: comp.compilers
Date: 30 Jan 2002 20:39:58 -0500
Organization: http://groups.google.com/
References: 02-01-118
Keywords: optimize
Posted-Date: 30 Jan 2002 20:39:58 EST

Regarding instruction scheduling: You are referring to the Pentium
processor. The PPro upwards operates very differently. See the
optimization manual at developer.intel.com for a description of the
internal operation of the P4 (plus a huge amount of optimization
advice).


> On a side note, when will compilers use some form of artificial
> intelligence to optimize the compiler's assembly output? :)


Well, this depends very much on your definition of "artificial
intelligence". If you define "ai" as some kind of magic which has not
been implemented or thought of yet, then the answer is "I don't know".
If you more realistically define it as the compiler doing something
similar to a good programmer or achieving similar results then the
answer is that good compilers already do optimize the "assembly
output" (instruction selection + scheduling etc.) to a significant
extent.
Anyhow, no compiler matches an experienced human assembly programmer
(or even comes close, at least in most cases).


Best regards
Bryan Hayes
http://www.hayestechnologies.com - Software Speed Optimization


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.