Re: [off-list] Re: PATCH: Rest of Intel Atom optimization

["H.J. Lu" <hjl dot tools at gmail dot com>]

Hi,

Here is the updated patch with update in comments:

Index: config/i386/i386.c
===================================================================
--- config/i386/i386.c	(revision 146353)
+++ config/i386/i386.c	(working copy)
@@ -13099,7 +13099,7 @@ distance_non_agu_define (unsigned int re

 done:
   /* get_attr_type may modify recog data.  We want to make sure
-     that recog data is valid for instruction, INSN, on which
+     that recog data is valid for instruction INSN, on which
      distance_non_agu_define is called.  INSN is unchanged here.  */
   extract_insn_cached (insn);
   return distance;


Tested on Linux/ia32 and Linux/Intel64, using --with-arch=atom --with-cpu=atom:

http://gcc.gnu.org/ml/gcc-testresults/2009-04/msg02043.html
http://gcc.gnu.org/ml/gcc-testresults/2009-04/msg02039.html

All failures are expected.

>
>> 2009-04-17 ?Joey Ye ?<joey.ye@intel.com>
>> ? ? ? ? ? ?Xuepeng Guo <xuepeng.guo@intel.com>
>> ? ? ? ? ? ?H.J. Lu ?<hongjiu.lu@intel.com>
>>
>> ? ? ? ?* config/i386/atom.md: Add bypasses with ix86_dep_by_shift_count.
>>
>> ? ? ? ?* config/i386/i386.c (LEA_SEARCH_THRESHOLD): New macro.
>> ? ? ? ?(IX86_LEA_PRIORITY): Likewise.
>> ? ? ? ?(distance_non_agu_define): New function.
>> ? ? ? ?(distance_agu_use): Likewise.
>> ? ? ? ?(ix86_lea_for_add_ok): Likewise.
>> ? ? ? ?(ix86_dep_by_shift_count): Likewise.
>>
>> ? ? ? ?* config/i386/i386.md: Call ix86_lea_for_add_ok to decide we
>> ? ? ? ?should split for LEA.
>>
>> ? ? ? ?* config/i386/i386-protos.h (ix86_lea_for_add_ok): Declare new
>> ? ? ? ?function.
>> ? ? ? ?(ix86_dep_by_shift_count): Likewise.
>>
>


-- 
H.J.

Attachment: gcc-atom-4.patch
Description: Text document