[gcc r14-4968] PR target/110551: Fix reg allocation for widening multiplications on x86.
Roger Sayle
sayle@gcc.gnu.org
Fri Oct 27 09:05:00 GMT 2023
https://gcc.gnu.org/g:89e5d902fc55ad375f149f25a84c516ad360a606
commit r14-4968-g89e5d902fc55ad375f149f25a84c516ad360a606
Author: Roger Sayle <roger@nextmovesoftware.com>
Date: Fri Oct 27 10:03:53 2023 +0100
PR target/110551: Fix reg allocation for widening multiplications on x86.
This patch contains clean-ups of the widening multiplication patterns in
i386.md, and provides variants of the existing highpart multiplication
peephole2 transformations (that tidy up register allocation after
reload), and thereby fixes PR target/110551, which is a superfluous
move instruction.
For the new test case, compiled on x86_64 with -O2.
Before:
mulx64: movabsq $-7046029254386353131, %rcx
movq %rcx, %rax
mulq %rdi
xorq %rdx, %rax
ret
After:
mulx64: movabsq $-7046029254386353131, %rax
mulq %rdi
xorq %rdx, %rax
ret
The clean-ups are (i) that operand 1 is consistently made register_operand
and operand 2 becomes nonimmediate_operand, so that predicates match the
constraints, (ii) the representation of the BMI2 mulx instruction is
updated to use the new umul_highpart RTX, and (iii) because operands
0 and 1 have different modes in widening multiplications, "a" is a more
appropriate constraint than "0" (which avoids spills/reloads containing
SUBREGs). The new peephole2 transformations are based upon those at
around line 9951 of i386.md, that begins with the comment
;; Highpart multiplication peephole2s to tweak register allocation.
;; mov imm,%rdx; mov %rdi,%rax; imulq %rdx -> mov imm,%rax; imulq %rdi
2023-10-27 Roger Sayle <roger@nextmovesoftware.com>
gcc/ChangeLog
PR target/110551
* config/i386/i386.md (<u>mul<mode><dwi>3): Make operands 1 and
2 take "regiser_operand" and "nonimmediate_operand" respectively.
(<u>mulqihi3): Likewise.
(*bmi2_umul<mode><dwi>3_1): Operand 2 needs to be register_operand
matching the %d constraint. Use umul_highpart RTX to represent
the highpart multiplication.
(*umul<mode><dwi>3_1): Operand 2 should use regiser_operand
predicate, and "a" rather than "0" as operands 0 and 2 have
different modes.
(define_split): For mul to mulx conversion, use the new
umul_highpart RTX representation.
(*mul<mode><dwi>3_1): Operand 1 should be register_operand
and the constraint %a as operands 0 and 1 have different modes.
(*<u>mulqihi3_1): Operand 1 should be register_operand matching
the constraint %0.
(define_peephole2): Providing widening multiplication variants
of the peephole2s that tweak highpart multiplication register
allocation.
gcc/testsuite/ChangeLog
PR target/110551
* gcc.target/i386/pr110551.c: New test case.
Diff:
---
gcc/config/i386/i386.md | 75 ++++++++++++++++++++++++--------
gcc/testsuite/gcc.target/i386/pr110551.c | 12 +++++
2 files changed, 68 insertions(+), 19 deletions(-)
diff --git a/gcc/config/i386/i386.md b/gcc/config/i386/i386.md
index 732c7a0323e9..eb4121b3f1e6 100644
--- a/gcc/config/i386/i386.md
+++ b/gcc/config/i386/i386.md
@@ -9725,33 +9725,29 @@
[(parallel [(set (match_operand:<DWI> 0 "register_operand")
(mult:<DWI>
(any_extend:<DWI>
- (match_operand:DWIH 1 "nonimmediate_operand"))
+ (match_operand:DWIH 1 "register_operand"))
(any_extend:<DWI>
- (match_operand:DWIH 2 "register_operand"))))
+ (match_operand:DWIH 2 "nonimmediate_operand"))))
(clobber (reg:CC FLAGS_REG))])])
(define_expand "<u>mulqihi3"
[(parallel [(set (match_operand:HI 0 "register_operand")
(mult:HI
(any_extend:HI
- (match_operand:QI 1 "nonimmediate_operand"))
+ (match_operand:QI 1 "register_operand"))
(any_extend:HI
- (match_operand:QI 2 "register_operand"))))
+ (match_operand:QI 2 "nonimmediate_operand"))))
(clobber (reg:CC FLAGS_REG))])]
"TARGET_QIMODE_MATH")
(define_insn "*bmi2_umul<mode><dwi>3_1"
[(set (match_operand:DWIH 0 "register_operand" "=r")
(mult:DWIH
- (match_operand:DWIH 2 "nonimmediate_operand" "%d")
+ (match_operand:DWIH 2 "register_operand" "%d")
(match_operand:DWIH 3 "nonimmediate_operand" "rm")))
(set (match_operand:DWIH 1 "register_operand" "=r")
- (truncate:DWIH
- (lshiftrt:<DWI>
- (mult:<DWI> (zero_extend:<DWI> (match_dup 2))
- (zero_extend:<DWI> (match_dup 3)))
- (match_operand:QI 4 "const_int_operand"))))]
- "TARGET_BMI2 && INTVAL (operands[4]) == <MODE_SIZE> * BITS_PER_UNIT
+ (umul_highpart:DWIH (match_dup 2) (match_dup 3)))]
+ "TARGET_BMI2
&& !(MEM_P (operands[2]) && MEM_P (operands[3]))"
"mulx\t{%3, %0, %1|%1, %0, %3}"
[(set_attr "type" "imulx")
@@ -9762,7 +9758,7 @@
[(set (match_operand:<DWI> 0 "register_operand" "=r,A")
(mult:<DWI>
(zero_extend:<DWI>
- (match_operand:DWIH 1 "nonimmediate_operand" "%d,0"))
+ (match_operand:DWIH 1 "register_operand" "%d,a"))
(zero_extend:<DWI>
(match_operand:DWIH 2 "nonimmediate_operand" "rm,rm"))))
(clobber (reg:CC FLAGS_REG))]
@@ -9798,11 +9794,7 @@
[(parallel [(set (match_dup 3)
(mult:DWIH (match_dup 1) (match_dup 2)))
(set (match_dup 4)
- (truncate:DWIH
- (lshiftrt:<DWI>
- (mult:<DWI> (zero_extend:<DWI> (match_dup 1))
- (zero_extend:<DWI> (match_dup 2)))
- (match_dup 5))))])]
+ (umul_highpart:DWIH (match_dup 1) (match_dup 2)))])]
{
split_double_mode (<DWI>mode, &operands[0], 1, &operands[3], &operands[4]);
@@ -9813,7 +9805,7 @@
[(set (match_operand:<DWI> 0 "register_operand" "=A")
(mult:<DWI>
(sign_extend:<DWI>
- (match_operand:DWIH 1 "nonimmediate_operand" "%0"))
+ (match_operand:DWIH 1 "register_operand" "%a"))
(sign_extend:<DWI>
(match_operand:DWIH 2 "nonimmediate_operand" "rm"))))
(clobber (reg:CC FLAGS_REG))]
@@ -9833,7 +9825,7 @@
[(set (match_operand:HI 0 "register_operand" "=a")
(mult:HI
(any_extend:HI
- (match_operand:QI 1 "nonimmediate_operand" "%0"))
+ (match_operand:QI 1 "register_operand" "%0"))
(any_extend:HI
(match_operand:QI 2 "nonimmediate_operand" "qm"))))
(clobber (reg:CC FLAGS_REG))]
@@ -9850,6 +9842,51 @@
(set_attr "bdver1_decode" "direct")
(set_attr "mode" "QI")])
+;; Widening multiplication peephole2s to tweak register allocation.
+;; mov imm,%rdx; mov %rdi,%rax; mulq %rdx -> mov imm,%rax; mulq %rdi
+(define_peephole2
+ [(set (match_operand:DWIH 0 "general_reg_operand")
+ (match_operand:DWIH 1 "immediate_operand"))
+ (set (match_operand:DWIH 2 "general_reg_operand")
+ (match_operand:DWIH 3 "general_reg_operand"))
+ (parallel [(set (match_operand:<DWI> 4 "general_reg_operand")
+ (mult:<DWI> (zero_extend:<DWI> (match_dup 2))
+ (zero_extend:<DWI> (match_dup 0))))
+ (clobber (reg:CC FLAGS_REG))])]
+ "REGNO (operands[3]) != AX_REG
+ && REGNO (operands[0]) != REGNO (operands[2])
+ && REGNO (operands[0]) != REGNO (operands[3])
+ && (REGNO (operands[0]) == REGNO (operands[4])
+ || REGNO (operands[0]) == DX_REG
+ || peep2_reg_dead_p (3, operands[0]))"
+ [(set (match_dup 2) (match_dup 1))
+ (parallel [(set (match_dup 4)
+ (mult:<DWI> (zero_extend:<DWI> (match_dup 2))
+ (zero_extend:<DWI> (match_dup 3))))
+ (clobber (reg:CC FLAGS_REG))])])
+
+;; mov imm,%rax; mov %rdi,%rdx; mulx %rax -> mov imm,%rdx; mulx %rdi
+(define_peephole2
+ [(set (match_operand:DWIH 0 "general_reg_operand")
+ (match_operand:DWIH 1 "immediate_operand"))
+ (set (match_operand:DWIH 2 "general_reg_operand")
+ (match_operand:DWIH 3 "general_reg_operand"))
+ (parallel [(set (match_operand:DWIH 4 "general_reg_operand")
+ (mult:DWIH (match_dup 2) (match_dup 0)))
+ (set (match_operand:DWIH 5 "general_reg_operand")
+ (umul_highpart:DWIH (match_dup 2) (match_dup 0)))])]
+ "REGNO (operands[3]) != DX_REG
+ && REGNO (operands[0]) != REGNO (operands[2])
+ && REGNO (operands[0]) != REGNO (operands[3])
+ && (REGNO (operands[0]) == REGNO (operands[4])
+ || REGNO (operands[0]) == REGNO (operands[5])
+ || peep2_reg_dead_p (3, operands[0]))"
+ [(set (match_dup 2) (match_dup 1))
+ (parallel [(set (match_dup 4)
+ (mult:DWIH (match_dup 2) (match_dup 3)))
+ (set (match_dup 5)
+ (umul_highpart:DWIH (match_dup 2) (match_dup 3)))])])
+
;; Highpart multiplication patterns
(define_insn "<s>mul<mode>3_highpart"
[(set (match_operand:DWIH 0 "register_operand" "=d")
diff --git a/gcc/testsuite/gcc.target/i386/pr110551.c b/gcc/testsuite/gcc.target/i386/pr110551.c
new file mode 100644
index 000000000000..142b808c675b
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/pr110551.c
@@ -0,0 +1,12 @@
+/* { dg-do compile { target int128 } } */
+/* { dg-options "-O2" } */
+
+typedef unsigned long long uint64_t;
+
+uint64_t mulx64(uint64_t x)
+{
+ __uint128_t r = (__uint128_t)x * 0x9E3779B97F4A7C15ull;
+ return (uint64_t)r ^ (uint64_t)( r >> 64 );
+}
+
+/* { dg-final { scan-assembler-not "movq" } } */
More information about the Gcc-cvs
mailing list