This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch] Enable Auto-vectorization for MIPS64
- From: "Chao-ying Fu" <fu at mips dot com>
- To: "Richard Sandiford" <rsandifo at redhat dot com>
- Cc: <gcc-patches at gcc dot gnu dot org>, "Jim Wilson" <wilson at specifixinc dot com>, "Stephens, Nigel" <nigel at mercury dot mips dot com>, "Thekkath, Radhika" <radhika at mercury dot mips dot com>, "Uhler, Mike" <uhler at mercury dot mips dot com>
- Date: Fri, 1 Oct 2004 14:02:18 -0700
- Subject: Re: [patch] Enable Auto-vectorization for MIPS64
- Organization: MIPS Technologies, Inc.
- References: <3CB54817FDF733459B230DD27C690CEC041C58@Exchange.MIPS.COM> <87sm8z2dhm.fsf@redhat.com>
- Reply-to: "Chao-ying Fu" <fu at mips dot com>
Ok. First, here is for "mips.h". Thanks!
2004-10-01 Chao-ying Fu <fu@mips.com>
* config/mips/mips.h (UNITS_PER_SIMD_WORD): Define.
Index: config/mips/mips.h
===================================================================
RCS file: /cvsroot/gcc/gcc/gcc/config/mips/mips.h,v
retrieving revision 1.371
diff -c -3 -p -r1.371 mips.h
*** config/mips/mips.h 25 Sep 2004 07:42:43 -0000 1.371
--- config/mips/mips.h 1 Oct 2004 20:58:07 -0000
*************** extern const struct mips_cpu_info *mips_
*** 1263,1268 ****
--- 1263,1270 ----
/* The number of bytes in a double. */
#define UNITS_PER_DOUBLE (TYPE_PRECISION (double_type_node) /
BITS_PER_UNIT)
+ #define UNITS_PER_SIMD_WORD (TARGET_PAIRED_SINGLE_FLOAT ? 8 : 0)
+
/* Set the sizes of the core types. */
#define SHORT_TYPE_SIZE 16
#define INT_TYPE_SIZE (TARGET_INT64 ? 64 : 32)
----- Original Message -----
From: "Richard Sandiford" <rsandifo@redhat.com>
To: "Fu, Chao-Ying" <fu@mips.com>
Cc: <gcc-patches@gcc.gnu.org>; "Jim Wilson" <wilson@specifixinc.com>;
"Stephens, Nigel" <nigel@mercury.mips.com>; "Thekkath, Radhika"
<radhika@mercury.mips.com>; "Uhler, Mike" <uhler@mercury.mips.com>
Sent: Thursday, September 30, 2004 10:50 PM
Subject: Re: [patch] Enable Auto-vectorization for MIPS64
> "Fu, Chao-Ying" <fu@mips.com> writes:
> > Here is the patch to enable auto-vectorization for MIPS64 that
> > has paired-single floating point instructions (V2SFmode).
> > We enable only floating point tests under gcc.dg/vect/*
> > for mipsisa64*-*-*.
>
> Thanks for doing this.
>
> Your patch was a missing a changelog, but aside from that,
> the MIPS part is fine. I can't approve the tree-vectorizer.c
> change of course.
>
> Richard
>