|Code selection trade-offs firstname.lastname@example.org (Tim Frink) (2008-05-12)|
|Re: Code selection trade-offs email@example.com (=?windows-1252?Q?Bj=F6rn_Franke?=) (2008-05-15)|
|Re: Code selection trade-offs firstname.lastname@example.org (IndianTechie) (2008-05-23)|
|Re: Code selection trade-offs email@example.com (Walter Banks) (2008-05-24)|
|From:||Walter Banks <firstname.lastname@example.org>|
|Date:||Sat, 24 May 2008 17:30:07 -0400|
|Organization:||NewsGuy - Unlimited Usenet $19.95|
|Posted-Date:||24 May 2008 17:43:08 EDT|
Tim Frink wrote:
> Do you know of any works where profiling information/ static program
> analysis is used as a heuristic to control code selection, i.e. the
> (profiling) execution counts of particular code structures are
> influencing the choice of instructions used to translate the source
> code into assembler? ...
There are some embedded applications where profiling can be
significant in code generation. The real implementation problem is the
evaluation of when to ignore the profile information. In cases where
in some code sections information from profiling shows that little can
be gained also identifies the area where performance may be further
sacrificed in exchange for performance in a critical section.
Byte Craft Limited
Tel. (519) 888-6911
Return to the
Search the comp.compilers archives again.