From: Kyrylo Tkachov <Kyrylo.Tkachov@arm.com>
To: "gcc-patches@gcc.gnu.org" <gcc-patches@gcc.gnu.org>
Subject: [PATCH][committed] aarch64: PR target/PR99195 Annotate more simple integer binary patterns with vcz subst rules
Date: Tue, 25 Apr 2023 13:52:17 +0000 [thread overview]
Message-ID: <AS8PR08MB69184D5200F02499A51AF21393649@AS8PR08MB6918.eurprd08.prod.outlook.com> (raw)
[-- Attachment #1: Type: text/plain, Size: 663 bytes --]
Hi all,
This patch adds more straightforward annotations to some more integer binary ops to
eliminate redundant fmovs around 64-bit SIMD results.
Bootstrapped and tested on aarch64-none-linux.
Pushing to trunk.
Thanks,
Kyrill
gcc/ChangeLog:
PR target/PR99195
* config/aarch64/aarch64-simd.md (orn<mode>3): Rename to...
(orn<mode>3<vczle><vczbe>): ... This.
(bic<mode>3): Rename to...
(bic<mode>3<vczle><vczbe>): ... This.
(<su><maxmin><mode>3): Rename to...
(<su><maxmin><mode>3<vczle><vczbe>): ... This.
gcc/testsuite/ChangeLog:
PR target/PR99195
* gcc.target/aarch64/simd/pr99195_1.c: Add tests for orn, bic, max and min.
[-- Attachment #2: binvcz.patch --]
[-- Type: application/octet-stream, Size: 3185 bytes --]
diff --git a/gcc/config/aarch64/aarch64-simd.md b/gcc/config/aarch64/aarch64-simd.md
index 0bf4346227ead8c8568c1221cb4f1ed546079efc..d2321113671053ec4c4ec3f7379d3287f9d922ee 100644
--- a/gcc/config/aarch64/aarch64-simd.md
+++ b/gcc/config/aarch64/aarch64-simd.md
@@ -398,7 +398,7 @@ (define_insn "aarch64_simd_mov_from_<mode>high"
(set_attr "length" "4")]
)
-(define_insn "orn<mode>3"
+(define_insn "orn<mode>3<vczle><vczbe>"
[(set (match_operand:VDQ_I 0 "register_operand" "=w")
(ior:VDQ_I (not:VDQ_I (match_operand:VDQ_I 1 "register_operand" "w"))
(match_operand:VDQ_I 2 "register_operand" "w")))]
@@ -407,7 +407,7 @@ (define_insn "orn<mode>3"
[(set_attr "type" "neon_logic<q>")]
)
-(define_insn "bic<mode>3"
+(define_insn "bic<mode>3<vczle><vczbe>"
[(set (match_operand:VDQ_I 0 "register_operand" "=w")
(and:VDQ_I (not:VDQ_I (match_operand:VDQ_I 1 "register_operand" "w"))
(match_operand:VDQ_I 2 "register_operand" "w")))]
@@ -1697,7 +1697,7 @@ (define_insn "aarch64_mls_n<mode>"
)
;; Max/Min operations.
-(define_insn "<su><maxmin><mode>3"
+(define_insn "<su><maxmin><mode>3<vczle><vczbe>"
[(set (match_operand:VDQ_BHSI 0 "register_operand" "=w")
(MAXMIN:VDQ_BHSI (match_operand:VDQ_BHSI 1 "register_operand" "w")
(match_operand:VDQ_BHSI 2 "register_operand" "w")))]
diff --git a/gcc/testsuite/gcc.target/aarch64/simd/pr99195_1.c b/gcc/testsuite/gcc.target/aarch64/simd/pr99195_1.c
index 3fe0e53bcd086a202ec1580a0b118121ef9334cc..5304e14b573aa7dea6b4eb2a3103882bb482a00d 100644
--- a/gcc/testsuite/gcc.target/aarch64/simd/pr99195_1.c
+++ b/gcc/testsuite/gcc.target/aarch64/simd/pr99195_1.c
@@ -37,18 +37,17 @@ OPFOUR (T, IS, OS, S, OP2, OP3, OP4, OP5)
FUNC (T, IS, OS, OP1, S) \
OPFIVE (T, IS, OS, S, OP2, OP3, OP4, OP5, OP6)
-#define OPSEVEN(T,IS,OS,S,OP1,OP2,OP3,OP4,OP5,OP6,OP7) \
-FUNC (T, IS, OS, OP1, S) \
-OPSIX (T, IS, OS, S, OP2, OP3, OP4, OP5, OP6, OP7)
-
+#define OPELEVEN(T,IS,OS,S,OP1,OP2,OP3,OP4,OP5,OP6,OP7,OP8,OP9,OP10,OP11) \
+OPFIVE (T, IS, OS, S, OP1, OP2, OP3, OP4, OP5) \
+OPSIX (T, IS, OS, S, OP6, OP7, OP8, OP9, OP10, OP11)
-OPSEVEN (int8, 8, 16, s8, padd, add, sub, mul, and, orr, eor)
-OPSEVEN (int16, 4, 8, s16, padd, add, sub, mul, and, orr, eor)
-OPSEVEN (int32, 2, 4, s32, padd, add, sub, mul, and, orr, eor)
+OPELEVEN (int8, 8, 16, s8, padd, add, sub, mul, and, orr, eor, orn, bic, max, min)
+OPELEVEN (int16, 4, 8, s16, padd, add, sub, mul, and, orr, eor, orn, bic, max, min)
+OPELEVEN (int32, 2, 4, s32, padd, add, sub, mul, and, orr, eor, orn, bic, max, min)
-OPSEVEN (uint8, 8, 16, u8, padd, add, sub, mul, and, orr, eor)
-OPSEVEN (uint16, 4, 8, u16, padd, add, sub, mul, and, orr, eor)
-OPSEVEN (uint32, 2, 4, u32, padd, add, sub, mul, and, orr, eor)
+OPELEVEN (uint8, 8, 16, u8, padd, add, sub, mul, and, orr, eor, orn, bic, max, min)
+OPELEVEN (uint16, 4, 8, u16, padd, add, sub, mul, and, orr, eor, orn, bic, max, min)
+OPELEVEN (uint32, 2, 4, u32, padd, add, sub, mul, and, orr, eor, orn, bic, max, min)
/* { dg-final { scan-assembler-not {\tfmov\t} } } */
/* { dg-final { scan-assembler-not {\tmov\t} } } */
reply other threads:[~2023-04-25 13:52 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=AS8PR08MB69184D5200F02499A51AF21393649@AS8PR08MB6918.eurprd08.prod.outlook.com \
--to=kyrylo.tkachov@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).