* [committed] [v2] More logical op simplifications in simplify-rtx.cc
@ 2024-05-25 18:42 Jeff Law
2024-06-11 22:38 ` Andrew Pinski
0 siblings, 1 reply; 2+ messages in thread
From: Jeff Law @ 2024-05-25 18:42 UTC (permalink / raw)
To: gcc-patches
[-- Attachment #1: Type: text/plain, Size: 1206 bytes --]
This is a revamp of what started as a target specific patch.
Basically xalan (corrected, I originally thought it was perlbench) has a
bitset implementation with a bit of an oddity. Specifically setBit will
clear the bit before it is set:
> if (bitToSet < 32)
> {
> fBits1 &= ~mask;
> fBits1 |= mask;
> }
> else
> {
> fBits2 &= ~mask;
> fBits2 |= mask;
> }
We can clean this up pretty easily in RTL with a small bit of code in
simplify-rtx. While xalan doesn't have other cases, we can synthesize
tests pretty easily and handle them as well.
It turns out we don't actually have to recognize this stuff at the bit
level, just standard logical identities are sufficient. For example
(X | Y) & ~Y -> X & ~Y
Andrew P. might poke at this at the gimple level. The type changes
kindof get in the way in gimple but he's much better at match.pd than I
am, so if he wants to chase it from the gimple side, I'll fully support
that.
Bootstrapped and regression tested on x86. Also run through my tester
on its embedded targets.
Pushing to the trunk.
jeff
[-- Attachment #2: P --]
[-- Type: text/plain, Size: 6622 bytes --]
commit 05daf617ea22e1d818295ed2d037456937e23530
Author: Jeff Law <jlaw@ventanamicro.com>
Date: Sat May 25 12:39:05 2024 -0600
[committed] [v2] More logical op simplifications in simplify-rtx.cc
This is a revamp of what started as a target specific patch.
Basically xalan (corrected, I originally thought it was perlbench) has a bitset
implementation with a bit of an oddity. Specifically setBit will clear the bit
before it is set:
> if (bitToSet < 32)
> {
> fBits1 &= ~mask;
> fBits1 |= mask;
> }
> else
> {
> fBits2 &= ~mask;
> fBits2 |= mask;
> }
We can clean this up pretty easily in RTL with a small bit of code in
simplify-rtx. While xalan doesn't have other cases, we can synthesize tests
pretty easily and handle them as well.
It turns out we don't actually have to recognize this stuff at the bit level,
just standard logical identities are sufficient. For example
(X | Y) & ~Y -> X & ~Y
Andrew P. might poke at this at the gimple level. The type changes kindof get
in the way in gimple but he's much better at match.pd than I am, so if he wants
to chase it from the gimple side, I'll fully support that.
Bootstrapped and regression tested on x86. Also run through my tester on its
embedded targets.
Pushing to the trunk.
gcc/
* simplify-rtx.cc (simplify_context::simplify_binary_operation_1): Handle
more logical simplifications.
gcc/testsuite/
* g++.target/riscv/redundant-bitmap-1.C: New test.
* g++.target/riscv/redundant-bitmap-2.C: New test.
* g++.target/riscv/redundant-bitmap-3.C: New test.
* g++.target/riscv/redundant-bitmap-4.C: New test.
diff --git a/gcc/simplify-rtx.cc b/gcc/simplify-rtx.cc
index 53f54d1d392..5caf1dfd957 100644
--- a/gcc/simplify-rtx.cc
+++ b/gcc/simplify-rtx.cc
@@ -3549,6 +3549,12 @@ simplify_context::simplify_binary_operation_1 (rtx_code code,
return tem;
}
+ /* Convert (ior (and (not A) B) A) into A | B. */
+ if (GET_CODE (op0) == AND
+ && GET_CODE (XEXP (op0, 0)) == NOT
+ && rtx_equal_p (XEXP (XEXP (op0, 0), 0), op1))
+ return simplify_gen_binary (IOR, mode, XEXP (op0, 1), op1);
+
tem = simplify_byte_swapping_operation (code, mode, op0, op1);
if (tem)
return tem;
@@ -3801,6 +3807,12 @@ simplify_context::simplify_binary_operation_1 (rtx_code code,
return tem;
}
+ /* Convert (xor (and (not A) B) A) into A | B. */
+ if (GET_CODE (op0) == AND
+ && GET_CODE (XEXP (op0, 0)) == NOT
+ && rtx_equal_p (XEXP (XEXP (op0, 0), 0), op1))
+ return simplify_gen_binary (IOR, mode, XEXP (op0, 1), op1);
+
tem = simplify_byte_swapping_operation (code, mode, op0, op1);
if (tem)
return tem;
@@ -4006,6 +4018,23 @@ simplify_context::simplify_binary_operation_1 (rtx_code code,
&& rtx_equal_p (op1, XEXP (XEXP (op0, 1), 0)))
return simplify_gen_binary (AND, mode, op1, XEXP (op0, 0));
+ /* (and (ior/xor (X Y) (not Y)) -> X & ~Y */
+ if ((GET_CODE (op0) == IOR || GET_CODE (op0) == XOR)
+ && GET_CODE (op1) == NOT
+ && rtx_equal_p (XEXP (op1, 0), XEXP (op0, 1)))
+ return simplify_gen_binary (AND, mode, XEXP (op0, 0),
+ simplify_gen_unary (NOT, mode,
+ XEXP (op1, 0),
+ mode));
+ /* (and (ior/xor (Y X) (not Y)) -> X & ~Y */
+ if ((GET_CODE (op0) == IOR || GET_CODE (op0) == XOR)
+ && GET_CODE (op1) == NOT
+ && rtx_equal_p (XEXP (op1, 0), XEXP (op0, 0)))
+ return simplify_gen_binary (AND, mode, XEXP (op0, 1),
+ simplify_gen_unary (NOT, mode,
+ XEXP (op1, 0),
+ mode));
+
/* Convert (and (ior A C) (ior B C)) into (ior (and A B) C). */
if (GET_CODE (op0) == GET_CODE (op1)
&& (GET_CODE (op0) == AND
diff --git a/gcc/testsuite/g++.target/riscv/redundant-bitmap-1.C b/gcc/testsuite/g++.target/riscv/redundant-bitmap-1.C
new file mode 100644
index 00000000000..37066f10eea
--- /dev/null
+++ b/gcc/testsuite/g++.target/riscv/redundant-bitmap-1.C
@@ -0,0 +1,14 @@
+/* { dg-do compile } */
+/* { dg-options "-O2 -march=rv64gc_zba_zbb_zbs -mabi=lp64" } */
+
+void setBit(char &a, int b) {
+ char c = 0x1UL << b;
+ a &= ~c;
+ a |= c;
+}
+
+/* { dg-final { scan-assembler-not "bclr\t" } } */
+/* { dg-final { scan-assembler-not "andn\t" } } */
+/* { dg-final { scan-assembler-not "or\t" } } */
+/* { dg-final { scan-assembler-times "bset\t" 1 } } */
+
diff --git a/gcc/testsuite/g++.target/riscv/redundant-bitmap-2.C b/gcc/testsuite/g++.target/riscv/redundant-bitmap-2.C
new file mode 100644
index 00000000000..86acaba298f
--- /dev/null
+++ b/gcc/testsuite/g++.target/riscv/redundant-bitmap-2.C
@@ -0,0 +1,14 @@
+/* { dg-do compile } */
+/* { dg-options "-O2 -march=rv64gc_zba_zbb_zbs -mabi=lp64" } */
+
+void setBit(char &a, int b) {
+ char c = 0x1UL << b;
+ a &= ~c;
+ a ^= c;
+}
+
+/* { dg-final { scan-assembler-not "bclr\t" } } */
+/* { dg-final { scan-assembler-not "xor\t" } } */
+/* { dg-final { scan-assembler-not "andn\t" } } */
+/* { dg-final { scan-assembler-times "bset\t" 1 } } */
+
diff --git a/gcc/testsuite/g++.target/riscv/redundant-bitmap-3.C b/gcc/testsuite/g++.target/riscv/redundant-bitmap-3.C
new file mode 100644
index 00000000000..16bd7c1785e
--- /dev/null
+++ b/gcc/testsuite/g++.target/riscv/redundant-bitmap-3.C
@@ -0,0 +1,14 @@
+/* { dg-do compile } */
+/* { dg-options "-O2 -march=rv64gc_zba_zbb_zbs -mabi=lp64" } */
+
+void setBit(char &a, int b) {
+ char c = 0x1UL << b;
+ a |= c;
+ a &= ~c;
+}
+
+/* { dg-final { scan-assembler-not "bset\t" } } */
+/* { dg-final { scan-assembler-not "or\t" } } */
+/* { dg-final { scan-assembler-not "andn\t" } } */
+/* { dg-final { scan-assembler-times "bclr\t" 1 } } */
+
diff --git a/gcc/testsuite/g++.target/riscv/redundant-bitmap-4.C b/gcc/testsuite/g++.target/riscv/redundant-bitmap-4.C
new file mode 100644
index 00000000000..f664ee01a01
--- /dev/null
+++ b/gcc/testsuite/g++.target/riscv/redundant-bitmap-4.C
@@ -0,0 +1,14 @@
+/* { dg-do compile } */
+/* { dg-options "-O2 -march=rv64gc_zba_zbb_zbs -mabi=lp64" } */
+
+void setBit(char &a, int b) {
+ char c = 0x1UL << b;
+ a ^= c;
+ a &= ~c;
+}
+
+/* { dg-final { scan-assembler-not "binv\t" } } */
+/* { dg-final { scan-assembler-not "xor\t" } } */
+/* { dg-final { scan-assembler-not "andn\t" } } */
+/* { dg-final { scan-assembler-times "bclr\t" 1 } } */
+
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [committed] [v2] More logical op simplifications in simplify-rtx.cc
2024-05-25 18:42 [committed] [v2] More logical op simplifications in simplify-rtx.cc Jeff Law
@ 2024-06-11 22:38 ` Andrew Pinski
0 siblings, 0 replies; 2+ messages in thread
From: Andrew Pinski @ 2024-06-11 22:38 UTC (permalink / raw)
To: Jeff Law; +Cc: gcc-patches
On Sat, May 25, 2024 at 11:42 AM Jeff Law <jlaw@ventanamicro.com> wrote:
>
> This is a revamp of what started as a target specific patch.
>
> Basically xalan (corrected, I originally thought it was perlbench) has a
> bitset implementation with a bit of an oddity. Specifically setBit will
> clear the bit before it is set:
>
> > if (bitToSet < 32)
> > {
> > fBits1 &= ~mask;
> > fBits1 |= mask;
> > }
> > else
> > {
> > fBits2 &= ~mask;
> > fBits2 |= mask;
> > }
>
> We can clean this up pretty easily in RTL with a small bit of code in
> simplify-rtx. While xalan doesn't have other cases, we can synthesize
> tests pretty easily and handle them as well.
>
>
> It turns out we don't actually have to recognize this stuff at the bit
> level, just standard logical identities are sufficient. For example
>
> (X | Y) & ~Y -> X & ~Y
>
>
>
> Andrew P. might poke at this at the gimple level. The type changes
> kindof get in the way in gimple but he's much better at match.pd than I
> am, so if he wants to chase it from the gimple side, I'll fully support
> that.
So we already have this pattern (without the type change) in gimple:
/* (~x | y) & x -> x & y */
/* (~x & y) | x -> x | y */
(simplify
(bitop:c (rbitop:c @2 @1) @0)
(with { bool wascmp; }
(if (bitwise_inverted_equal_p (@0, @2, wascmp)
&& (!wascmp || element_precision (type) == 1))
(bitop @0 @1))))
The problem is bitwise_inverted_equal_p does not see that:
c.0_4 = (signed char) _1;
_5 = ~c.0_4;
_16 = (charD.11) _5;
and
c_11 = (charD.11) _1;
are bitwise inversions of each other.
I filed https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115449 to keep
track of this.
Thanks,
Andrew Pinski
>
> Bootstrapped and regression tested on x86. Also run through my tester
> on its embedded targets.
>
> Pushing to the trunk.
>
> jeff
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2024-06-11 22:38 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-05-25 18:42 [committed] [v2] More logical op simplifications in simplify-rtx.cc Jeff Law
2024-06-11 22:38 ` Andrew Pinski
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).