From: "Roger Sayle" <roger@nextmovesoftware.com>
To: "'GCC Patches'" <gcc-patches@gcc.gnu.org>
Subject: [x86 PATCH] Pre-reload splitter to transform and;cmp into not;test.
Date: Thu, 26 May 2022 19:41:29 +0100 [thread overview]
Message-ID: <006301d87130$3747a240$a5d6e6c0$@nextmovesoftware.com> (raw)
[-- Attachment #1: Type: text/plain, Size: 1674 bytes --]
A common idiom for testing if a specific set of bits is set in a value
is to use "(X & Y) == Y", which on x86 results in an AND followed by a
CMP. A slightly improved implementation is to instead use (~X & Y)==0,
that uses a NOT and a TEST (or ANDN where available); still two "fast"
instructions, but typically shorter especially if Y is an immediate
constant. Because the above transformation would require more gimple
statements in SSA, and may only be a win on targets with flags registers,
it isn't performed by the middle-end, instead leaving this choice to
the backend.
As an example, here's the change in code generation for pr91400-1.c
[which now requires a tweak to its dg-final clauses].
Before:
movl __cpu_model+12(%rip), %eax
andl $68, %eax // 3 bytes
cmpl $68, %eax // 3 bytes
sete %al
ret
After:
movl __cpu_model+12(%rip), %eax
notl %eax // 2 bytes
testb $68, %al // 2 bytes
sete %al
ret
This patch has been tested on x86_64-pc-linux-gnu with make bootstrap
and make -k check, both with and without --target_board=unix{-m32},
with no new failures. Ok for mainline?
2022-05-26 Roger Sayle <roger@nextmovesoftware.com>
gcc/ChangeLog
* config/i386/i386.md (*test<mode>_not): New define_insn_and_split
to split a combined "and;cmp" sequence into "not;test".
gcc/testsuite/ChangeLog
* gcc.target/i386/pr91400-1.c: Update for improved code generation.
* gcc.target/i386/pr91400-2.c: Likewise.
* gcc.target/i386/testnot-1.c: New test case.
* gcc.target/i386/testnot-2.c: Likewise.
Thanks in advance,
Roger
--
[-- Attachment #2: patchtn3.txt --]
[-- Type: text/plain, Size: 3665 bytes --]
diff --git a/gcc/config/i386/i386.md b/gcc/config/i386/i386.md
index b9b8f78..602dfa7 100644
--- a/gcc/config/i386/i386.md
+++ b/gcc/config/i386/i386.md
@@ -9716,6 +9716,27 @@
operands[2] = gen_rtx_AND (mode, val, immed_wide_int_const (mask, mode));
})
+;; Split and;cmp (as optimized by combine) into not;test
+;; Except when TARGET_BMI provides andn (*andn_<mode>_ccno).
+(define_insn_and_split "*test<mode>_not"
+ [(set (reg:CCZ FLAGS_REG)
+ (compare:CCZ
+ (and:SWI
+ (not:SWI (match_operand:SWI 0 "register_operand"))
+ (match_operand:SWI 1 "<nonmemory_szext_operand>"))
+ (const_int 0)))]
+ "ix86_pre_reload_split ()
+ && (!TARGET_BMI || !REG_P (operands[1]))"
+ "#"
+ "&& 1"
+ [(set (match_dup 2) (not:SWI (match_dup 0)))
+ (set (reg:CCZ FLAGS_REG)
+ (compare:CCZ (and:SWI (match_dup 2) (match_dup 1))
+ (const_int 0)))]
+{
+ operands[2] = gen_reg_rtx (<MODE>mode);
+})
+
;; Convert HImode/SImode test instructions with immediate to QImode ones.
;; i386 does not allow to encode test with 8bit sign extended immediate, so
;; this is relatively important trick.
diff --git a/gcc/testsuite/gcc.target/i386/pr91400-1.c b/gcc/testsuite/gcc.target/i386/pr91400-1.c
index 6124058..751dc6c 100644
--- a/gcc/testsuite/gcc.target/i386/pr91400-1.c
+++ b/gcc/testsuite/gcc.target/i386/pr91400-1.c
@@ -1,8 +1,8 @@
/* PR target/91400 */
/* { dg-do compile } */
/* { dg-options "-O2" } */
-/* { dg-final { scan-assembler-times "andl" 1 } } */
-/* { dg-final { scan-assembler-times "cmpl" 1 } } */
+/* { dg-final { scan-assembler-times "notl" 1 } } */
+/* { dg-final { scan-assembler-times "testb" 1 } } */
/* { dg-final { scan-assembler-times "sete" 1 } } */
/* { dg-final { scan-assembler-not "cmove" } } */
diff --git a/gcc/testsuite/gcc.target/i386/pr91400-2.c b/gcc/testsuite/gcc.target/i386/pr91400-2.c
index 1af5a2f..914acd7 100644
--- a/gcc/testsuite/gcc.target/i386/pr91400-2.c
+++ b/gcc/testsuite/gcc.target/i386/pr91400-2.c
@@ -1,8 +1,8 @@
/* PR target/91400 */
/* { dg-do compile } */
/* { dg-options "-O2" } */
-/* { dg-final { scan-assembler-times "andl" 1 } } */
-/* { dg-final { scan-assembler-times "cmpl" 1 } } */
+/* { dg-final { scan-assembler-times "notl" 1 } } */
+/* { dg-final { scan-assembler-times "testb" 1 } } */
/* { dg-final { scan-assembler-times "sete" 1 } } */
/* { dg-final { scan-assembler-not "cmove" } } */
diff --git a/gcc/testsuite/gcc.target/i386/testnot-1.c b/gcc/testsuite/gcc.target/i386/testnot-1.c
new file mode 100644
index 0000000..9ebcb5c
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/testnot-1.c
@@ -0,0 +1,24 @@
+/* { dg-do compile } */
+/* { dg-options "-O2" } */
+
+int foo(int x)
+{
+ return (x & 1234) == 1234;
+}
+
+int foos(short x)
+{
+ return (x & 1234) == 1234;
+}
+
+int fooc(char x)
+{
+ return (x & 123) == 123;
+}
+
+int fool(long long x)
+{
+ return (x & 1234) == 1234;
+}
+
+/* { dg-final { scan-assembler-not "cmp" } } */
diff --git a/gcc/testsuite/gcc.target/i386/testnot-2.c b/gcc/testsuite/gcc.target/i386/testnot-2.c
new file mode 100644
index 0000000..52fdaf3
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/testnot-2.c
@@ -0,0 +1,24 @@
+/* { dg-do compile { target { ! ia32 } } } */
+/* { dg-options "-O2" } */
+
+int foo(int x, int y)
+{
+ return (x & y) == y;
+}
+
+int foos(short x, short y)
+{
+ return (x & y) == y;
+}
+
+int fooc(char x, char y)
+{
+ return (x & y) == y;
+}
+
+int fool(long long x, long long y)
+{
+ return (x & y) == y;
+}
+
+/* { dg-final { scan-assembler-not "cmp" } } */
next reply other threads:[~2022-05-26 18:41 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-05-26 18:41 Roger Sayle [this message]
2022-05-26 18:59 ` [x86 PATCH] Pre-reload splitter to transform and; cmp into not; test Uros Bizjak
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='006301d87130$3747a240$a5d6e6c0$@nextmovesoftware.com' \
--to=roger@nextmovesoftware.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).