https://gcc.gnu.org/g:283eb27d5f674b6fb11d8640495993e2ec693608
commit r16-5561-g283eb27d5f674b6fb11d8640495993e2ec693608 Author: Robin Dapp <[email protected]> Date: Sun Nov 16 18:42:04 2025 +0100 forwprop: Allow nop conversions for vector constructor. I observed a vect-construct forwprop opportunity in x264 that we could handle when checking for a nop conversion instead of a useless conversion. IMHO a nop-conversion check is sufficient as we're only dealing with permutations in simplify_vector_constructor. This patch replaces uses of useless_type_conversion_p with tree_nop_conversion_p in simplify_vector_constructor. It was bootstrapped and regtested on x86 and power10, regtested on aarch64 and riscv64. There is a single scan-test failure on power (gcc.target/powerpc/builtins-1.c). The code actually looks better so I took the liberty of adjusting the test expectation. gcc/ChangeLog: * tree-ssa-forwprop.cc (simplify_vector_constructor): Allow nop conversions. gcc/testsuite/ChangeLog: * gcc.target/powerpc/builtins-1.c: Adjust test expectation. * gcc.target/riscv/rvv/autovec/pr118019-3.c: New test. Diff: --- gcc/testsuite/gcc.target/powerpc/builtins-1.c | 2 +- .../gcc.target/riscv/rvv/autovec/pr118019-3.c | 51 ++++++++++++++++++++++ gcc/tree-ssa-forwprop.cc | 16 +++---- 3 files changed, 60 insertions(+), 9 deletions(-) diff --git a/gcc/testsuite/gcc.target/powerpc/builtins-1.c b/gcc/testsuite/gcc.target/powerpc/builtins-1.c index 8410a5fd4319..4e7e5384675f 100644 --- a/gcc/testsuite/gcc.target/powerpc/builtins-1.c +++ b/gcc/testsuite/gcc.target/powerpc/builtins-1.c @@ -1035,4 +1035,4 @@ foo156 (vector unsigned short usa) /* { dg-final { scan-assembler-times {\mvmrglb\M} 3 } } */ /* { dg-final { scan-assembler-times {\mvmrgew\M} 4 } } */ /* { dg-final { scan-assembler-times {\mvsplth|xxsplth\M} 4 } } */ -/* { dg-final { scan-assembler-times {\mxxpermdi\M} 44 } } */ +/* { dg-final { scan-assembler-times {\mxxpermdi\M} 42 } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/pr118019-3.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/pr118019-3.c new file mode 100644 index 000000000000..6cb761f63c0f --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/pr118019-3.c @@ -0,0 +1,51 @@ +/* { dg-do compile } */ +/* { dg-options "-O3 -march=rv64gcv_zvl512b -mabi=lp64d -mno-vector-strict-align -mmax-vectorization -fdump-tree-optimized" } */ + +/* Ensure we don't use a vector constructor but a permute. */ + +typedef unsigned char uint8_t; +typedef unsigned short uint16_t; +typedef unsigned int uint32_t; + +#define HADAMARD4(d0, d1, d2, d3, s0, s1, s2, s3) {\ + int t0 = s0 + s1;\ + int t1 = s0 - s1;\ + int t2 = s2 + s3;\ + int t3 = s2 - s3;\ + d0 = t0 + t2;\ + d2 = t0 - t2;\ + d1 = t1 + t3;\ + d3 = t1 - t3;\ +} + +uint32_t +abs2 (uint32_t a) +{ + uint32_t s = ((a >> 15) & 0x10001) * 0xffff; + return (a + s) ^ s; +} + +int +x264_pixel_satd_8x4 (uint8_t *pix1, int i_pix1, uint8_t *pix2, int i_pix2) +{ + uint32_t tmp[4][4]; + uint32_t a0, a1, a2, a3; + int sum = 0; + for (int i = 0; i < 4; i++, pix1 += i_pix1, pix2 += i_pix2) + { + a0 = (pix1[0] - pix2[0]) + ((pix1[4] - pix2[4]) << 16); + a1 = (pix1[1] - pix2[1]) + ((pix1[5] - pix2[5]) << 16); + a2 = (pix1[2] - pix2[2]) + ((pix1[6] - pix2[6]) << 16); + a3 = (pix1[3] - pix2[3]) + ((pix1[7] - pix2[7]) << 16); + HADAMARD4 (tmp[i][0], tmp[i][1], tmp[i][2], tmp[i][3], a0, a1, a2, a3); + } +#pragma GCC unroll 4 + for (int i = 0; i < 4; i++) + { + HADAMARD4 (a0, a1, a2, a3, tmp[0][i], tmp[1][i], tmp[2][i], tmp[3][i]); + sum += abs2 (a0) + abs2 (a1) + abs2 (a2) + abs2 (a3); + } + return (((uint16_t) sum) + ((uint32_t) sum >> 16)) >> 1; +} + +/* { dg-final { scan-tree-dump-not "BIT_FIELD_REF" "optimized" } } */ diff --git a/gcc/tree-ssa-forwprop.cc b/gcc/tree-ssa-forwprop.cc index c9a57d67adec..00140ce950c3 100644 --- a/gcc/tree-ssa-forwprop.cc +++ b/gcc/tree-ssa-forwprop.cc @@ -3873,11 +3873,11 @@ simplify_vector_constructor (gimple_stmt_iterator *gsi) if (op1 && TREE_CODE ((ref = TREE_OPERAND (op1, 0))) == SSA_NAME && VECTOR_TYPE_P (TREE_TYPE (ref)) - && (useless_type_conversion_p (TREE_TYPE (op1), - TREE_TYPE (TREE_TYPE (ref))) + && (tree_nop_conversion_p (TREE_TYPE (op1), + TREE_TYPE (TREE_TYPE (ref))) || (VECTOR_TYPE_P (TREE_TYPE (op1)) - && useless_type_conversion_p (TREE_TYPE (TREE_TYPE (op1)), - TREE_TYPE (TREE_TYPE (ref))) + && tree_nop_conversion_p (TREE_TYPE (TREE_TYPE (op1)), + TREE_TYPE (TREE_TYPE (ref))) && TYPE_VECTOR_SUBPARTS (TREE_TYPE (op1)) .is_constant (&nsubelts))) && constant_multiple_p (bit_field_size (op1), nsubelts, @@ -4101,8 +4101,8 @@ simplify_vector_constructor (gimple_stmt_iterator *gsi) { gcc_assert (known_eq (TYPE_VECTOR_SUBPARTS (type), TYPE_VECTOR_SUBPARTS (src_type)) - && useless_type_conversion_p (TREE_TYPE (type), - TREE_TYPE (src_type))); + && tree_nop_conversion_p (TREE_TYPE (type), + TREE_TYPE (src_type))); tree rhs = build1 (VIEW_CONVERT_EXPR, type, orig[0]); orig[0] = make_ssa_name (type); gassign *assign = gimple_build_assign (orig[0], rhs); @@ -4238,8 +4238,8 @@ simplify_vector_constructor (gimple_stmt_iterator *gsi) { gcc_assert (known_eq (TYPE_VECTOR_SUBPARTS (type), TYPE_VECTOR_SUBPARTS (perm_type)) - && useless_type_conversion_p (TREE_TYPE (type), - TREE_TYPE (perm_type))); + && tree_nop_conversion_p (TREE_TYPE (type), + TREE_TYPE (perm_type))); res = gimple_build (&stmts, VIEW_CONVERT_EXPR, type, res); } /* Blend in the actual constant. */
