public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* [PATCH, rs6000] Remove DImode from SLOW_UNALIGNED_ACCESS
@ 2012-04-17 14:56 Pat Haugen
  2012-04-17 15:09 ` David Edelsohn
  0 siblings, 1 reply; 2+ messages in thread
From: Pat Haugen @ 2012-04-17 14:56 UTC (permalink / raw)
  To: gcc-patches; +Cc: David Edelsohn

[-- Attachment #1: Type: text/plain, Size: 304 bytes --]

DImode references do not suffer a major performance hit for < 4-byte aligned access like the float types.

Bootstrap/regtest on powerpc64-linux with no new regressions. Ok for trunk?

-Pat


2012-04-17  Pat Haugen <pthaugen@us.ibm.com>

	* config/rs6000/rs6000.h (SLOW_UNALIGNED_ACCESS): Remove DImode.


[-- Attachment #2: slow_unalign.diff --]
[-- Type: text/x-patch, Size: 669 bytes --]

Index: gcc/config/rs6000/rs6000.h
===================================================================
--- gcc/config/rs6000/rs6000.h	(revision 186389)
+++ gcc/config/rs6000/rs6000.h	(working copy)
@@ -771,8 +771,7 @@ extern unsigned rs6000_pointer_size;
 #define SLOW_UNALIGNED_ACCESS(MODE, ALIGN)				\
   (STRICT_ALIGNMENT							\
    || (((MODE) == SFmode || (MODE) == DFmode || (MODE) == TFmode	\
-	|| (MODE) == SDmode || (MODE) == DDmode || (MODE) == TDmode	\
-	|| (MODE) == DImode)						\
+	|| (MODE) == SDmode || (MODE) == DDmode || (MODE) == TDmode)	\
        && (ALIGN) < 32)							\
    || (VECTOR_MODE_P ((MODE)) && (((int)(ALIGN)) < VECTOR_ALIGN (MODE))))
 

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [PATCH, rs6000] Remove DImode from SLOW_UNALIGNED_ACCESS
  2012-04-17 14:56 [PATCH, rs6000] Remove DImode from SLOW_UNALIGNED_ACCESS Pat Haugen
@ 2012-04-17 15:09 ` David Edelsohn
  0 siblings, 0 replies; 2+ messages in thread
From: David Edelsohn @ 2012-04-17 15:09 UTC (permalink / raw)
  To: Pat Haugen; +Cc: gcc-patches

On Tue, Apr 17, 2012 at 10:55 AM, Pat Haugen
<pthaugen@linux.vnet.ibm.com> wrote:
> DImode references do not suffer a major performance hit for < 4-byte aligned
> access like the float types.
>
> Bootstrap/regtest on powerpc64-linux with no new regressions. Ok for trunk?
>
> 2012-04-17  Pat Haugen <pthaugen@us.ibm.com>
>
>        * config/rs6000/rs6000.h (SLOW_UNALIGNED_ACCESS): Remove DImode.

Okay.

I think this may have been introduced at the time the port was 32 bit
and DImode values sometimes were allocated to FPRs.

Thanks, David

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2012-04-17 15:09 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2012-04-17 14:56 [PATCH, rs6000] Remove DImode from SLOW_UNALIGNED_ACCESS Pat Haugen
2012-04-17 15:09 ` David Edelsohn

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).