From: Jonathan Wright <Jonathan.Wright@arm.com>
To: "gcc-patches@gcc.gnu.org" <gcc-patches@gcc.gnu.org>
Cc: Richard Sandiford <Richard.Sandiford@arm.com>,
Kyrylo Tkachov <Kyrylo.Tkachov@arm.com>
Subject: [PATCH] aarch64: Use type-qualified builtins for [R]ADDHN[2] Neon intrinsics
Date: Thu, 11 Nov 2021 10:25:40 +0000 [thread overview]
Message-ID: <DB9PR08MB695957AE0E64364748045B02EB949@DB9PR08MB6959.eurprd08.prod.outlook.com> (raw)
[-- Attachment #1: Type: text/plain, Size: 1415 bytes --]
Hi,
This patch declares unsigned type-qualified builtins and uses them to
implement (rounding) halving-narrowing-add Neon intrinsics. This
removes the need for many casts in arm_neon.h.
Bootstrapped and regression tested on aarch64-none-linux-gnu - no
issues.
Ok for master?
Thanks,
Jonathan
---
gcc/ChangeLog:
2021-11-09 Jonathan Wright <jonathan.wright@arm.com>
* config/aarch64/aarch64-simd-builtins.def: Declare unsigned
builtins for [r]addhn[2].
* config/aarch64/arm_neon.h (vaddhn_s16): Remove unnecessary
cast.
(vaddhn_s32): Likewise.
(vaddhn_s64): Likewise.
(vaddhn_u16): Use type-qualified builtin and remove casts.
(vaddhn_u32): Likewise.
(vaddhn_u64): Likewise.
(vraddhn_s16): Remove unnecessary cast.
(vraddhn_s32): Likewise.
(vraddhn_s64): Likewise.
(vraddhn_u16): Use type-qualified builtin and remove casts.
(vraddhn_u32): Likewise.
(vraddhn_u64): Likewise.
(vaddhn_high_s16): Remove unnecessary cast.
(vaddhn_high_s32): Likewise.
(vaddhn_high_s64): Likewise.
(vaddhn_high_u16): Use type-qualified builtin and remove
casts.
(vaddhn_high_u32): Likewise.
(vaddhn_high_u64): Likewise.
(vraddhn_high_s16): Remove unnecessary cast.
(vraddhn_high_s32): Likewise.
(vraddhn_high_s64): Likewise.
(vraddhn_high_u16): Use type-qualified builtin and remove
casts.
(vraddhn_high_u32): Likewise.
(vraddhn_high_u64): Likewise.
[-- Attachment #2: rb15037.patch --]
[-- Type: application/octet-stream, Size: 8914 bytes --]
diff --git a/gcc/config/aarch64/aarch64-simd-builtins.def b/gcc/config/aarch64/aarch64-simd-builtins.def
index af04b732227439dcaaa2f3751097050d988eb729..6372da80be33c40cb27e5811bfb4f4f672f28a35 100644
--- a/gcc/config/aarch64/aarch64-simd-builtins.def
+++ b/gcc/config/aarch64/aarch64-simd-builtins.def
@@ -220,13 +220,17 @@
/* Implemented by aarch64_<sur><addsub>hn<mode>. */
BUILTIN_VQN (BINOP, addhn, 0, NONE)
+ BUILTIN_VQN (BINOPU, addhn, 0, NONE)
BUILTIN_VQN (BINOP, subhn, 0, NONE)
BUILTIN_VQN (BINOP, raddhn, 0, NONE)
+ BUILTIN_VQN (BINOPU, raddhn, 0, NONE)
BUILTIN_VQN (BINOP, rsubhn, 0, NONE)
/* Implemented by aarch64_<sur><addsub>hn2<mode>. */
BUILTIN_VQN (TERNOP, addhn2, 0, NONE)
+ BUILTIN_VQN (TERNOPU, addhn2, 0, NONE)
BUILTIN_VQN (TERNOP, subhn2, 0, NONE)
BUILTIN_VQN (TERNOP, raddhn2, 0, NONE)
+ BUILTIN_VQN (TERNOPU, raddhn2, 0, NONE)
BUILTIN_VQN (TERNOP, rsubhn2, 0, NONE)
/* Implemented by aarch64_<us>xtl<mode>. */
diff --git a/gcc/config/aarch64/arm_neon.h b/gcc/config/aarch64/arm_neon.h
index b2781f680d142b848f622d2f4965b42985885502..cb481542ba0d6ffb7cc8ffe7c1a098930fc5e746 100644
--- a/gcc/config/aarch64/arm_neon.h
+++ b/gcc/config/aarch64/arm_neon.h
@@ -713,186 +713,168 @@ __extension__ extern __inline int8x8_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vaddhn_s16 (int16x8_t __a, int16x8_t __b)
{
- return (int8x8_t) __builtin_aarch64_addhnv8hi (__a, __b);
+ return __builtin_aarch64_addhnv8hi (__a, __b);
}
__extension__ extern __inline int16x4_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vaddhn_s32 (int32x4_t __a, int32x4_t __b)
{
- return (int16x4_t) __builtin_aarch64_addhnv4si (__a, __b);
+ return __builtin_aarch64_addhnv4si (__a, __b);
}
__extension__ extern __inline int32x2_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vaddhn_s64 (int64x2_t __a, int64x2_t __b)
{
- return (int32x2_t) __builtin_aarch64_addhnv2di (__a, __b);
+ return __builtin_aarch64_addhnv2di (__a, __b);
}
__extension__ extern __inline uint8x8_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vaddhn_u16 (uint16x8_t __a, uint16x8_t __b)
{
- return (uint8x8_t) __builtin_aarch64_addhnv8hi ((int16x8_t) __a,
- (int16x8_t) __b);
+ return __builtin_aarch64_addhnv8hi_uuu (__a, __b);
}
__extension__ extern __inline uint16x4_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vaddhn_u32 (uint32x4_t __a, uint32x4_t __b)
{
- return (uint16x4_t) __builtin_aarch64_addhnv4si ((int32x4_t) __a,
- (int32x4_t) __b);
+ return __builtin_aarch64_addhnv4si_uuu (__a, __b);
}
__extension__ extern __inline uint32x2_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vaddhn_u64 (uint64x2_t __a, uint64x2_t __b)
{
- return (uint32x2_t) __builtin_aarch64_addhnv2di ((int64x2_t) __a,
- (int64x2_t) __b);
+ return __builtin_aarch64_addhnv2di_uuu (__a, __b);
}
__extension__ extern __inline int8x8_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vraddhn_s16 (int16x8_t __a, int16x8_t __b)
{
- return (int8x8_t) __builtin_aarch64_raddhnv8hi (__a, __b);
+ return __builtin_aarch64_raddhnv8hi (__a, __b);
}
__extension__ extern __inline int16x4_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vraddhn_s32 (int32x4_t __a, int32x4_t __b)
{
- return (int16x4_t) __builtin_aarch64_raddhnv4si (__a, __b);
+ return __builtin_aarch64_raddhnv4si (__a, __b);
}
__extension__ extern __inline int32x2_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vraddhn_s64 (int64x2_t __a, int64x2_t __b)
{
- return (int32x2_t) __builtin_aarch64_raddhnv2di (__a, __b);
+ return __builtin_aarch64_raddhnv2di (__a, __b);
}
__extension__ extern __inline uint8x8_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vraddhn_u16 (uint16x8_t __a, uint16x8_t __b)
{
- return (uint8x8_t) __builtin_aarch64_raddhnv8hi ((int16x8_t) __a,
- (int16x8_t) __b);
+ return __builtin_aarch64_raddhnv8hi_uuu (__a, __b);
}
__extension__ extern __inline uint16x4_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vraddhn_u32 (uint32x4_t __a, uint32x4_t __b)
{
- return (uint16x4_t) __builtin_aarch64_raddhnv4si ((int32x4_t) __a,
- (int32x4_t) __b);
+ return __builtin_aarch64_raddhnv4si_uuu (__a, __b);
}
__extension__ extern __inline uint32x2_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vraddhn_u64 (uint64x2_t __a, uint64x2_t __b)
{
- return (uint32x2_t) __builtin_aarch64_raddhnv2di ((int64x2_t) __a,
- (int64x2_t) __b);
+ return __builtin_aarch64_raddhnv2di_uuu (__a, __b);
}
__extension__ extern __inline int8x16_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vaddhn_high_s16 (int8x8_t __a, int16x8_t __b, int16x8_t __c)
{
- return (int8x16_t) __builtin_aarch64_addhn2v8hi (__a, __b, __c);
+ return __builtin_aarch64_addhn2v8hi (__a, __b, __c);
}
__extension__ extern __inline int16x8_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vaddhn_high_s32 (int16x4_t __a, int32x4_t __b, int32x4_t __c)
{
- return (int16x8_t) __builtin_aarch64_addhn2v4si (__a, __b, __c);
+ return __builtin_aarch64_addhn2v4si (__a, __b, __c);
}
__extension__ extern __inline int32x4_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vaddhn_high_s64 (int32x2_t __a, int64x2_t __b, int64x2_t __c)
{
- return (int32x4_t) __builtin_aarch64_addhn2v2di (__a, __b, __c);
+ return __builtin_aarch64_addhn2v2di (__a, __b, __c);
}
__extension__ extern __inline uint8x16_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vaddhn_high_u16 (uint8x8_t __a, uint16x8_t __b, uint16x8_t __c)
{
- return (uint8x16_t) __builtin_aarch64_addhn2v8hi ((int8x8_t) __a,
- (int16x8_t) __b,
- (int16x8_t) __c);
+ return __builtin_aarch64_addhn2v8hi_uuuu (__a, __b, __c);
}
__extension__ extern __inline uint16x8_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vaddhn_high_u32 (uint16x4_t __a, uint32x4_t __b, uint32x4_t __c)
{
- return (uint16x8_t) __builtin_aarch64_addhn2v4si ((int16x4_t) __a,
- (int32x4_t) __b,
- (int32x4_t) __c);
+ return __builtin_aarch64_addhn2v4si_uuuu (__a, __b, __c);
}
__extension__ extern __inline uint32x4_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vaddhn_high_u64 (uint32x2_t __a, uint64x2_t __b, uint64x2_t __c)
{
- return (uint32x4_t) __builtin_aarch64_addhn2v2di ((int32x2_t) __a,
- (int64x2_t) __b,
- (int64x2_t) __c);
+ return __builtin_aarch64_addhn2v2di_uuuu (__a, __b, __c);
}
__extension__ extern __inline int8x16_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vraddhn_high_s16 (int8x8_t __a, int16x8_t __b, int16x8_t __c)
{
- return (int8x16_t) __builtin_aarch64_raddhn2v8hi (__a, __b, __c);
+ return __builtin_aarch64_raddhn2v8hi (__a, __b, __c);
}
__extension__ extern __inline int16x8_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vraddhn_high_s32 (int16x4_t __a, int32x4_t __b, int32x4_t __c)
{
- return (int16x8_t) __builtin_aarch64_raddhn2v4si (__a, __b, __c);
+ return __builtin_aarch64_raddhn2v4si (__a, __b, __c);
}
__extension__ extern __inline int32x4_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vraddhn_high_s64 (int32x2_t __a, int64x2_t __b, int64x2_t __c)
{
- return (int32x4_t) __builtin_aarch64_raddhn2v2di (__a, __b, __c);
+ return __builtin_aarch64_raddhn2v2di (__a, __b, __c);
}
__extension__ extern __inline uint8x16_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vraddhn_high_u16 (uint8x8_t __a, uint16x8_t __b, uint16x8_t __c)
{
- return (uint8x16_t) __builtin_aarch64_raddhn2v8hi ((int8x8_t) __a,
- (int16x8_t) __b,
- (int16x8_t) __c);
+ return __builtin_aarch64_raddhn2v8hi_uuuu (__a, __b, __c);
}
__extension__ extern __inline uint16x8_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vraddhn_high_u32 (uint16x4_t __a, uint32x4_t __b, uint32x4_t __c)
{
- return (uint16x8_t) __builtin_aarch64_raddhn2v4si ((int16x4_t) __a,
- (int32x4_t) __b,
- (int32x4_t) __c);
+ return __builtin_aarch64_raddhn2v4si_uuuu (__a, __b, __c);
}
__extension__ extern __inline uint32x4_t
__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
vraddhn_high_u64 (uint32x2_t __a, uint64x2_t __b, uint64x2_t __c)
{
- return (uint32x4_t) __builtin_aarch64_raddhn2v2di ((int32x2_t) __a,
- (int64x2_t) __b,
- (int64x2_t) __c);
+ return __builtin_aarch64_raddhn2v2di_uuuu (__a, __b, __c);
}
__extension__ extern __inline float32x2_t
next reply other threads:[~2021-11-11 10:25 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-11 10:25 Jonathan Wright [this message]
2021-11-11 10:31 ` Richard Sandiford
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=DB9PR08MB695957AE0E64364748045B02EB949@DB9PR08MB6959.eurprd08.prod.outlook.com \
--to=jonathan.wright@arm.com \
--cc=Kyrylo.Tkachov@arm.com \
--cc=Richard.Sandiford@arm.com \
--cc=gcc-patches@gcc.gnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).