diff options
author | Richard Sandiford <richard.sandiford@arm.com> | 2020-04-23 15:45:43 +0100 |
---|---|---|
committer | Richard Sandiford <richard.sandiford@arm.com> | 2020-04-23 15:45:43 +0100 |
commit | 901f5289d9465d4c388ae288f850ad4f29e99a2c (patch) | |
tree | 328b47dedcb13f4a994dc4705f3779e6eefb5572 /gcc | |
parent | e73a32d6d47ec7c5fb5a5fe7eb896c0e1258ea68 (diff) | |
download | gcc-901f5289d9465d4c388ae288f850ad4f29e99a2c.zip gcc-901f5289d9465d4c388ae288f850ad4f29e99a2c.tar.gz gcc-901f5289d9465d4c388ae288f850ad4f29e99a2c.tar.bz2 |
vect: Fix comparisons between invariant booleans [PR94727]
This PR was caused by mismatched expectations between
vectorizable_comparison and SLP. We had a "<" comparison
between two booleans that were leaves of the SLP tree, so
vectorizable_comparison fell back on:
/* Invariant comparison. */
if (!vectype)
{
vectype = get_vectype_for_scalar_type (vinfo, TREE_TYPE (rhs1),
slp_node);
if (maybe_ne (TYPE_VECTOR_SUBPARTS (vectype), nunits))
return false;
}
rhs1 and rhs2 were *unsigned* boolean types, so we got back a vector
of unsigned integers. This in itself was OK, and meant that "<"
worked as expected without the need for the boolean fix-ups:
/* Boolean values may have another representation in vectors
and therefore we prefer bit operations over comparison for
them (which also works for scalar masks). We store opcodes
to use in bitop1 and bitop2. Statement is vectorized as
BITOP2 (rhs1 BITOP1 rhs2) or
rhs1 BITOP2 (BITOP1 rhs2)
depending on bitop1 and bitop2 arity. */
bool swap_p = false;
if (VECTOR_BOOLEAN_TYPE_P (vectype))
{
However, vectorizable_comparison then used vect_get_slp_defs to get
the actual operands. The request went to vect_get_constant_vectors,
which also has logic to calculate the vector type. The problem was
that this type was different from the one chosen above:
if (VECT_SCALAR_BOOLEAN_TYPE_P (TREE_TYPE (op))
&& vect_mask_constant_operand_p (stmt_vinfo))
vector_type = truth_type_for (stmt_vectype);
else
vector_type = get_vectype_for_scalar_type (vinfo, TREE_TYPE (op), op_node);
So the function gave back a vector of mask types, which here are vectors
of *signed* booleans. This meant that "<" gave:
true (-1) < false (0)
and so the boolean fixup above was needed after all.
Fixed by making vectorizable_comparison also pick a mask type in
this case.
2020-04-23 Richard Sandiford <richard.sandiford@arm.com>
gcc/
PR tree-optimization/94727
* tree-vect-stmts.c (vectorizable_comparison): Use mask_type when
comparing invariant scalar booleans.
gcc/testsuite/
PR tree-optimization/94727
* gcc.dg/vect/pr94727.c: New test.
Diffstat (limited to 'gcc')
-rw-r--r-- | gcc/ChangeLog | 6 | ||||
-rw-r--r-- | gcc/testsuite/ChangeLog | 5 | ||||
-rw-r--r-- | gcc/testsuite/gcc.dg/vect/pr94727.c | 24 | ||||
-rw-r--r-- | gcc/tree-vect-stmts.c | 7 |
4 files changed, 40 insertions, 2 deletions
diff --git a/gcc/ChangeLog b/gcc/ChangeLog index 71c480a..041894e 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,3 +1,9 @@ +2020-04-23 Richard Sandiford <richard.sandiford@arm.com> + + PR tree-optimization/94727 + * tree-vect-stmts.c (vectorizable_comparison): Use mask_type when + comparing invariant scalar booleans. + 2020-04-23 Matthew Malcomson <matthew.malcomson@arm.com> Jakub Jelinek <jakub@redhat.com> diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog index 7e676f0..2b28272 100644 --- a/gcc/testsuite/ChangeLog +++ b/gcc/testsuite/ChangeLog @@ -1,3 +1,8 @@ +2020-04-23 Richard Sandiford <richard.sandiford@arm.com> + + PR tree-optimization/94727 + * gcc.dg/vect/pr94727.c: New test. + 2020-04-23 Szabolcs Nagy <szabolcs.nagy@arm.com> PR target/94514 diff --git a/gcc/testsuite/gcc.dg/vect/pr94727.c b/gcc/testsuite/gcc.dg/vect/pr94727.c new file mode 100644 index 0000000..3840871 --- /dev/null +++ b/gcc/testsuite/gcc.dg/vect/pr94727.c @@ -0,0 +1,24 @@ +/* { dg-additional-options "-O3" } */ + +unsigned char a[16][32]; +long b[16][32]; +unsigned long c; +_Bool d; + +void __attribute__((noipa)) +foo (void) +{ + for (int j = 0; j < 8; j++) + for (int i = 0; i < 17; ++i) + b[j][i] = (a[j][i] < c) > d; +} + +int +main (void) +{ + c = 1; + foo (); + if (!b[0][0]) + __builtin_abort (); + return 0; +} diff --git a/gcc/tree-vect-stmts.c b/gcc/tree-vect-stmts.c index 7f3a9fb..88a1e2c 100644 --- a/gcc/tree-vect-stmts.c +++ b/gcc/tree-vect-stmts.c @@ -10566,8 +10566,11 @@ vectorizable_comparison (stmt_vec_info stmt_info, gimple_stmt_iterator *gsi, /* Invariant comparison. */ if (!vectype) { - vectype = get_vectype_for_scalar_type (vinfo, TREE_TYPE (rhs1), - slp_node); + if (VECT_SCALAR_BOOLEAN_TYPE_P (TREE_TYPE (rhs1))) + vectype = mask_type; + else + vectype = get_vectype_for_scalar_type (vinfo, TREE_TYPE (rhs1), + slp_node); if (!vectype || maybe_ne (TYPE_VECTOR_SUBPARTS (vectype), nunits)) return false; } |