This change detects smax(a, -a) and
smax(-a, a) at the RTL level for integer
operands and optimizes that to abs(a).
PR rtl-optimization/126726
gcc/ChangeLog:
* simplify-rtx.cc (simplify_context::simplify_binary_operation_1):
Detect smax(a, -a) for integers and optimize it to abs(a).
Signed-off-by: Kevin Stefanov <[email protected]>
---
Hi all,
This is my patch for Easy Issue of Week 52
to detect smax(a, -a) at the RTL level for
integers and optimize it to abs(a).
Bugzilla link:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126726
I bootstrapped and regression tested with
enable-languages=c,c++,fortran and using
"make check-gcc" instead of "make -k check"
because the latter was ending with an error
even on pristine sources, both on x86_64 and
on aarch64, and I was advised the former should
suffice for a patch like this.
gcc/simplify-rtx.cc | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/gcc/simplify-rtx.cc b/gcc/simplify-rtx.cc
index 285826a3fed..5616a6b889f 100644
--- a/gcc/simplify-rtx.cc
+++ b/gcc/simplify-rtx.cc
@@ -5035,6 +5035,13 @@ simplify_ashift:
return op1;
if (rtx_equal_p (trueop0, trueop1) && ! side_effects_p (op0))
return op0;
+ if (INTEGRAL_MODE_P (mode))
+ {
+ if (GET_CODE (op0) == NEG && rtx_equal_p (XEXP (op0, 0), op1))
+ return simplify_gen_unary (ABS, mode, op1, mode);
+ else if (GET_CODE (op1) == NEG && rtx_equal_p (XEXP (op1, 0), op0))
+ return simplify_gen_unary (ABS, mode, op0, mode);
+ }
tem = simplify_associative_operation (code, mode, op0, op1);
if (tem)
return tem;
--
2.55.0
