public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc(refs/vendors/riscv/heads/gcc-13-with-riscv-opts)] RISC-V: Add `movMODEcc' implementation for generic targets
@ 2023-11-22 5:11 Jeff Law
0 siblings, 0 replies; only message in thread
From: Jeff Law @ 2023-11-22 5:11 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:1bea5fb99ce0afa76fe49c6c961ddf4fecee8172
commit 1bea5fb99ce0afa76fe49c6c961ddf4fecee8172
Author: Maciej W. Rozycki <macro@embecosm.com>
Date: Wed Nov 22 01:18:28 2023 +0000
RISC-V: Add `movMODEcc' implementation for generic targets
Provide RTL expansion of conditional-move operations for generic targets
using a suitable sequence of base integer machine instructions according
to cost evaluation by if-conversion. Add `-mmovcc' command line option
to enable this transformation, off by default.
For the generic sequences small immediates as per the `arith_operand'
predicate are cost-equivalent to registers as we can use them as input,
alternative to a register, to the respective AND[I] machine operations,
however we need to reject immediates fulfilling `lui_operand', because
they would require reloading into a register, making the operation more
costly. Therefore add `movcc_operand' predicate and use it accordingly.
There is a need to adjust zbs-bext-02.c, which can also serve as emitted
code example, because with certain compilation options an AND operation
can now legitimately appear in output despite BEXT having been produced
as expected, such as with `-march=rv64gc -O2':
foo:
mv a3,a0
li a5,0
mv a0,a1
li a2,64
li a1,1
.L3:
sll a4,a1,a5
and a4,a4,a3
addiw a5,a5,1
beq a4,zero,.L2
addiw a0,a0,1
.L2:
bne a5,a2,.L3
ret
vs `-march=rv64gc_zbs -O2':
foo:
mv a4,a0
li a5,0
mv a0,a1
li a3,64
.L3:
bext a2,a4,a5
beq a2,zero,.L2
addiw a0,a0,1
.L2:
addiw a5,a5,1
bne a5,a3,.L3
ret
and then with `-march=rv64gc -mmovcc -mbranch-cost=7':
foo:
mv a6,a0
li a4,0
mv a0,a1
li a7,1
li a1,64
.L3:
sll a5,a7,a4
and a5,a5,a6
snez a5,a5
neg a5,a5
not a2,a5
addiw a3,a0,1
and a5,a5,a3
and a0,a2,a0
addiw a4,a4,1
or a0,a5,a0
bne a4,a1,.L3
ret
vs `-march=rv64gc_zbs -mmovcc -mbranch-cost=7':
foo:
mv a6,a0
li a4,0
mv a0,a1
li a1,64
.L3:
bext a5,a6,a4
neg a5,a5
not a2,a5
addiw a3,a0,1
and a5,a5,a3
and a0,a2,a0
addiw a4,a4,1
or a0,a5,a0
bne a4,a1,.L3
ret
However BEXT is supposed to replace an SLL operation so adjust the test
case to reject SLL rather than AND, letting the test case pass even with
`/-mmovcc/-mbranch-cost=7' specified as DejaGNU test flags (and in the
absence of target-specific conditional-move operations enabled either by
default or with other test flags).
gcc/
* config/riscv/predicates.md (movcc_operand): New predicate.
* config/riscv/riscv.cc (riscv_expand_conditional_move): Handle
generic targets.
* config/riscv/riscv.md (mov<mode>cc): Likewise.
* config/riscv/riscv.opt (mmovcc): New option.
* doc/invoke.texi (Option Summary): Document it.
gcc/testsuite/
* gcc.target/riscv/zbs-bext-02.c: Adjust to reject SLL rather
than AND.
(cherry picked from commit dc95b338c04281224f5c9c058acec170e4008b5c)
Diff:
---
gcc/config/riscv/predicates.md | 6 ++++
gcc/config/riscv/riscv.cc | 41 +++++++++++++++++++++++-----
gcc/config/riscv/riscv.md | 7 +++--
gcc/config/riscv/riscv.opt | 4 +++
gcc/doc/invoke.texi | 9 ++++++
gcc/testsuite/gcc.target/riscv/zbs-bext-02.c | 2 +-
6 files changed, 58 insertions(+), 11 deletions(-)
diff --git a/gcc/config/riscv/predicates.md b/gcc/config/riscv/predicates.md
index 90567a817a7..f943f31f4db 100644
--- a/gcc/config/riscv/predicates.md
+++ b/gcc/config/riscv/predicates.md
@@ -41,6 +41,12 @@
(ior (match_operand 0 "arith_operand")
(match_operand 0 "lui_operand")))
+(define_predicate "movcc_operand"
+ (if_then_else (match_test "TARGET_SFB_ALU || TARGET_XTHEADCONDMOV
+ || TARGET_ZICOND_LIKE")
+ (match_operand 0 "sfb_alu_operand")
+ (match_operand 0 "arith_operand")))
+
(define_predicate "const_csr_operand"
(and (match_code "const_int")
(match_test "IN_RANGE (INTVAL (op), 0, 31)")))
diff --git a/gcc/config/riscv/riscv.cc b/gcc/config/riscv/riscv.cc
index 5198da6d936..df08e9c831a 100644
--- a/gcc/config/riscv/riscv.cc
+++ b/gcc/config/riscv/riscv.cc
@@ -4128,7 +4128,9 @@ riscv_expand_conditional_move (rtx dest, rtx op, rtx cons, rtx alt)
rtx op0 = XEXP (op, 0);
rtx op1 = XEXP (op, 1);
- if ((TARGET_ZICOND_LIKE && GET_MODE_CLASS (mode) == MODE_INT)
+ if (((TARGET_ZICOND_LIKE
+ || (arith_operand (cons, mode) && arith_operand (alt, mode)))
+ && (GET_MODE_CLASS (mode) == MODE_INT))
|| TARGET_SFB_ALU || TARGET_XTHEADCONDMOV)
{
machine_mode mode0 = GET_MODE (op0);
@@ -4142,6 +4144,15 @@ riscv_expand_conditional_move (rtx dest, rtx op, rtx cons, rtx alt)
|| (mode1 != word_mode && mode1 != VOIDmode))
return false;
+ /* In the fallback generic case use MODE rather than WORD_MODE for
+ the output of the SCC instruction, to match the mode of the NEG
+ operation below. The output of SCC is 0 or 1 boolean, so it is
+ valid for input in any scalar integer mode. */
+ rtx tmp = gen_reg_rtx ((TARGET_ZICOND_LIKE
+ || TARGET_SFB_ALU || TARGET_XTHEADCONDMOV)
+ ? word_mode : mode);
+ bool invert = false;
+
/* Canonicalize the comparison. It must be an equality comparison
of integer operands, or with SFB it can be any comparison of
integer operands. If it isn't, then emit an SCC instruction
@@ -4150,7 +4161,6 @@ riscv_expand_conditional_move (rtx dest, rtx op, rtx cons, rtx alt)
|| !INTEGRAL_MODE_P (mode0))
{
bool *invert_ptr = nullptr;
- bool invert = false;
/* If riscv_expand_int_scc inverts the condition, then it will
flip the value of INVERT. We need to know where so that
@@ -4158,11 +4168,9 @@ riscv_expand_conditional_move (rtx dest, rtx op, rtx cons, rtx alt)
if (code == LE || code == LEU || code == GE || code == GEU)
invert_ptr = &invert;
- /* Emit an scc like instruction into a temporary
- so that we can use an EQ/NE comparison. */
- rtx tmp = gen_reg_rtx (word_mode);
-
- /* We can support both FP and integer conditional moves. */
+ /* Emit an SCC-like instruction into a temporary so that we can
+ use an EQ/NE comparison. We can support both FP and integer
+ conditional moves. */
if (INTEGRAL_MODE_P (mode0))
riscv_expand_int_scc (tmp, code, op0, op1, invert_ptr);
else if (FLOAT_MODE_P (mode0)
@@ -4178,6 +4186,8 @@ riscv_expand_conditional_move (rtx dest, rtx op, rtx cons, rtx alt)
op0 = XEXP (op, 0);
op1 = XEXP (op, 1);
}
+ else if (!TARGET_ZICOND_LIKE && !TARGET_SFB_ALU && !TARGET_XTHEADCONDMOV)
+ riscv_expand_int_scc (tmp, code, op0, op1, &invert);
if (TARGET_SFB_ALU || TARGET_XTHEADCONDMOV)
{
@@ -4195,6 +4205,23 @@ riscv_expand_conditional_move (rtx dest, rtx op, rtx cons, rtx alt)
cons, alt)));
return true;
}
+ else if (!TARGET_ZICOND_LIKE)
+ {
+ if (invert)
+ std::swap (cons, alt);
+
+ rtx reg1 = gen_reg_rtx (mode);
+ rtx reg2 = gen_reg_rtx (mode);
+ rtx reg3 = gen_reg_rtx (mode);
+ rtx reg4 = gen_reg_rtx (mode);
+
+ riscv_emit_unary (NEG, reg1, tmp);
+ riscv_emit_binary (AND, reg2, reg1, cons);
+ riscv_emit_unary (NOT, reg3, reg1);
+ riscv_emit_binary (AND, reg4, reg3, alt);
+ riscv_emit_binary (IOR, dest, reg2, reg4);
+ return true;
+ }
/* 0, reg or 0, imm */
else if (cons == CONST0_RTX (mode)
&& (REG_P (alt)
diff --git a/gcc/config/riscv/riscv.md b/gcc/config/riscv/riscv.md
index e6d90c2265f..37860cfe4d6 100644
--- a/gcc/config/riscv/riscv.md
+++ b/gcc/config/riscv/riscv.md
@@ -2658,9 +2658,10 @@
(define_expand "mov<mode>cc"
[(set (match_operand:GPR 0 "register_operand")
(if_then_else:GPR (match_operand 1 "comparison_operator")
- (match_operand:GPR 2 "sfb_alu_operand")
- (match_operand:GPR 3 "sfb_alu_operand")))]
- "TARGET_SFB_ALU || TARGET_XTHEADCONDMOV || TARGET_ZICOND_LIKE"
+ (match_operand:GPR 2 "movcc_operand")
+ (match_operand:GPR 3 "movcc_operand")))]
+ "TARGET_SFB_ALU || TARGET_XTHEADCONDMOV || TARGET_ZICOND_LIKE
+ || TARGET_MOVCC"
{
if (riscv_expand_conditional_move (operands[0], operands[1],
operands[2], operands[3]))
diff --git a/gcc/config/riscv/riscv.opt b/gcc/config/riscv/riscv.opt
index 555288e682a..11526f9cc2a 100644
--- a/gcc/config/riscv/riscv.opt
+++ b/gcc/config/riscv/riscv.opt
@@ -460,6 +460,10 @@ misa-spec=
Target RejectNegative Joined Enum(isa_spec_class) Var(riscv_isa_spec) Init(TARGET_DEFAULT_ISA_SPEC)
Set the version of RISC-V ISA spec.
+mmovcc
+Target Var(TARGET_MOVCC)
+Enable conditional moves unconditionally.
+
minline-atomics
Target Var(TARGET_INLINE_SUBWORD_ATOMIC) Init(1)
Always inline subword atomic operations.
diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi
index bb30dfc9479..2667a7d7acc 100644
--- a/gcc/doc/invoke.texi
+++ b/gcc/doc/invoke.texi
@@ -1228,6 +1228,7 @@ See RS/6000 and PowerPC Options.
-mstack-protector-guard=@var{guard} -mstack-protector-guard-reg=@var{reg}
-mstack-protector-guard-offset=@var{offset}
-mcsr-check -mno-csr-check
+-mmovcc -mno-movcc
-minline-atomics -mno-inline-atomics
-minline-strlen -mno-inline-strlen
-minline-strcmp -mno-inline-strcmp
@@ -29032,6 +29033,14 @@ Do or don't use smaller but slower prologue and epilogue code that uses
library function calls. The default is to use fast inline prologues and
epilogues.
+@opindex mmovcc
+@item -mmovcc
+@itemx -mno-movcc
+Do or don't produce branchless conditional-move code sequences even with
+targets that do not have specific instructions for conditional operations.
+If enabled, sequences of ALU operations are produced using base integer
+ISA instructions where profitable.
+
@opindex minline-atomics
@item -minline-atomics
@itemx -mno-inline-atomics
diff --git a/gcc/testsuite/gcc.target/riscv/zbs-bext-02.c b/gcc/testsuite/gcc.target/riscv/zbs-bext-02.c
index 8c5d8c7a41a..81d826d9dd4 100644
--- a/gcc/testsuite/gcc.target/riscv/zbs-bext-02.c
+++ b/gcc/testsuite/gcc.target/riscv/zbs-bext-02.c
@@ -15,4 +15,4 @@ foo(const long long B, int a)
/* { dg-final { scan-assembler-times "bext\t" 1 } } */
/* { dg-final { scan-assembler-not {\mbset} } } */
-/* { dg-final { scan-assembler-not {\mand} } } */
+/* { dg-final { scan-assembler-not {\msll} } } */
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2023-11-22 5:11 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-11-22 5:11 [gcc(refs/vendors/riscv/heads/gcc-13-with-riscv-opts)] RISC-V: Add `movMODEcc' implementation for generic targets Jeff Law
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).