Re: Semantics of .align pseudo-op

zstern@adobe.com (Zalman Stern)
Sun, 1 Aug 1993 08:10:25 GMT

          From comp.compilers

Related articles
Semantics of .align pseudo-op Dave.Reese@East.Sun.COM (1993-07-30)
Re: Semantics of .align pseudo-op zstern@adobe.com (1993-08-01)
| List of all articles for this month |

Newsgroups: comp.compilers
From: zstern@adobe.com (Zalman Stern)
Keywords: assembler
Organization: Adobe Systems Incorporated
References: 93-07-086
Date: Sun, 1 Aug 1993 08:10:25 GMT

David Reese writes
[Discussion of whether .align in a code section should fill with nop's or
illegal instruction values. In particular regarding alignment for
superscalar performance optimization.]


How about introducing a new pseudo op specifically for this purpose.
Preferably one that encodes the knowledge of what is the best alignment into
the assembler instead of the source code. (It could optionally specify the
alignment explicitly, but if left blank, the assembler should do the right
thing for the processor type it is generating code for.) Of course I also
like the MIPS "semi-optimizing" assembler, so I might be biased.
--
Zalman Stern zalman@adobe.com (415) 962 3824
Adobe Systems, 1585 Charleston Rd., POB 7900, Mountain View, CA 94039-7900
--


Post a followup to this message

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