| Message ID | 20240701214657.3163066-2-quic_apinski@quicinc.com |
|---|---|
| State | New |
| Headers | show |
| Series | Some improvements to complex lowering (cabs related) | expand |
On Mon, Jul 1, 2024 at 11:47 PM Andrew Pinski <quic_apinski@quicinc.com> wrote: > > This is just a small optimization for the case where the real and imag > parts are the same when lowering complex addition/subtraction. We only > need to do the addition once when the real and imag parts are the same (on > both sides of the operator). This gets done later on by FRE/PRE/DOM but > having it done soon allows the cabs lowering to remove the sqrt and > just change it to a multiply by a constant. > > Bootstrapped and tested on x86_64-linux-gnu. OK. > gcc/ChangeLog: > > * tree-complex.cc (expand_complex_addition): If both > operands have the same real and imag parts, only > add the addition once. > > gcc/testsuite/ChangeLog: > > * gcc.dg/tree-ssa/complex-8.c: New test. > > Signed-off-by: Andrew Pinski <quic_apinski@quicinc.com> > --- > gcc/testsuite/gcc.dg/tree-ssa/complex-8.c | 12 ++++++++++++ > gcc/tree-complex.cc | 7 ++++++- > 2 files changed, 18 insertions(+), 1 deletion(-) > create mode 100644 gcc/testsuite/gcc.dg/tree-ssa/complex-8.c > > diff --git a/gcc/testsuite/gcc.dg/tree-ssa/complex-8.c b/gcc/testsuite/gcc.dg/tree-ssa/complex-8.c > new file mode 100644 > index 00000000000..a9636ff9e9a > --- /dev/null > +++ b/gcc/testsuite/gcc.dg/tree-ssa/complex-8.c > @@ -0,0 +1,12 @@ > +/* { dg-do compile } */ > +/* { dg-options "-O1 -fdump-tree-cplxlower1-raw" } */ > + > +_Complex double f(double a, double c) > +{ > + _Complex double d = __builtin_complex (a, a); > + d+=__builtin_complex(c, c); > + return d; > +} > + > +/* There should only be one plus as (a+c) is still (a+c) */ > +/* { dg-final { scan-tree-dump-times "plus_expr, " 1 "cplxlower1" } } */ > diff --git a/gcc/tree-complex.cc b/gcc/tree-complex.cc > index 8a879acffca..dfebec18ec3 100644 > --- a/gcc/tree-complex.cc > +++ b/gcc/tree-complex.cc > @@ -984,7 +984,12 @@ expand_complex_addition (gimple_stmt_iterator *gsi, tree inner_type, > case PAIR (VARYING, VARYING): > general: > rr = gimple_build (&stmts, loc, code, inner_type, ar, br); > - ri = gimple_build (&stmts, loc, code, inner_type, ai, bi); > + /* (a+ai) + (b+bi) -> (a+b)+(a+b)i > + small optimization to remove one new statement. */ > + if (operand_equal_p (ar, ai) && operand_equal_p (br, bi)) > + ri = rr; > + else > + ri = gimple_build (&stmts, loc, code, inner_type, ai, bi); > break; > > default: > -- > 2.43.0 >
diff --git a/gcc/testsuite/gcc.dg/tree-ssa/complex-8.c b/gcc/testsuite/gcc.dg/tree-ssa/complex-8.c new file mode 100644 index 00000000000..a9636ff9e9a --- /dev/null +++ b/gcc/testsuite/gcc.dg/tree-ssa/complex-8.c @@ -0,0 +1,12 @@ +/* { dg-do compile } */ +/* { dg-options "-O1 -fdump-tree-cplxlower1-raw" } */ + +_Complex double f(double a, double c) +{ + _Complex double d = __builtin_complex (a, a); + d+=__builtin_complex(c, c); + return d; +} + +/* There should only be one plus as (a+c) is still (a+c) */ +/* { dg-final { scan-tree-dump-times "plus_expr, " 1 "cplxlower1" } } */ diff --git a/gcc/tree-complex.cc b/gcc/tree-complex.cc index 8a879acffca..dfebec18ec3 100644 --- a/gcc/tree-complex.cc +++ b/gcc/tree-complex.cc @@ -984,7 +984,12 @@ expand_complex_addition (gimple_stmt_iterator *gsi, tree inner_type, case PAIR (VARYING, VARYING): general: rr = gimple_build (&stmts, loc, code, inner_type, ar, br); - ri = gimple_build (&stmts, loc, code, inner_type, ai, bi); + /* (a+ai) + (b+bi) -> (a+b)+(a+b)i + small optimization to remove one new statement. */ + if (operand_equal_p (ar, ai) && operand_equal_p (br, bi)) + ri = rr; + else + ri = gimple_build (&stmts, loc, code, inner_type, ai, bi); break; default:
This is just a small optimization for the case where the real and imag parts are the same when lowering complex addition/subtraction. We only need to do the addition once when the real and imag parts are the same (on both sides of the operator). This gets done later on by FRE/PRE/DOM but having it done soon allows the cabs lowering to remove the sqrt and just change it to a multiply by a constant. Bootstrapped and tested on x86_64-linux-gnu. gcc/ChangeLog: * tree-complex.cc (expand_complex_addition): If both operands have the same real and imag parts, only add the addition once. gcc/testsuite/ChangeLog: * gcc.dg/tree-ssa/complex-8.c: New test. Signed-off-by: Andrew Pinski <quic_apinski@quicinc.com> --- gcc/testsuite/gcc.dg/tree-ssa/complex-8.c | 12 ++++++++++++ gcc/tree-complex.cc | 7 ++++++- 2 files changed, 18 insertions(+), 1 deletion(-) create mode 100644 gcc/testsuite/gcc.dg/tree-ssa/complex-8.c