This patch gets the scalar mode of a reduction operation from the gimple stmt rather than the vectype. This makes it more suitable for use in scalar costs.
Tested on aarch64-linux-gnu & applied. Richard gcc/ * config/aarch64/aarch64.c (aarch64_sve_in_loop_reduction_latency): Remove vectype parameter and get floatness from the type of the stmt lhs instead. (arch64_in_loop_reduction_latency): Likewise. (aarch64_detect_vector_stmt_subtype): Update caller. (aarch64_vector_costs::count_ops): Likewise. --- gcc/config/aarch64/aarch64.c | 33 +++++++++++++-------------------- 1 file changed, 13 insertions(+), 20 deletions(-) diff --git a/gcc/config/aarch64/aarch64.c b/gcc/config/aarch64/aarch64.c index c8a3cb38473..416362beefd 100644 --- a/gcc/config/aarch64/aarch64.c +++ b/gcc/config/aarch64/aarch64.c @@ -15220,14 +15220,13 @@ aarch64_sve_only_stmt_p (stmt_vec_info stmt_info, tree vectype) return false; } -/* We are considering implementing STMT_INFO using SVE vector type VECTYPE. - If STMT_INFO is an in-loop reduction that SVE supports directly, return - its latency in cycles, otherwise return zero. SVE_COSTS specifies the - latencies of the relevant instructions. */ +/* We are considering implementing STMT_INFO using SVE. If STMT_INFO is an + in-loop reduction that SVE supports directly, return its latency in cycles, + otherwise return zero. SVE_COSTS specifies the latencies of the relevant + instructions. */ static unsigned int aarch64_sve_in_loop_reduction_latency (vec_info *vinfo, stmt_vec_info stmt_info, - tree vectype, const sve_vec_cost *sve_costs) { switch (vect_reduc_type (vinfo, stmt_info)) @@ -15236,7 +15235,7 @@ aarch64_sve_in_loop_reduction_latency (vec_info *vinfo, return sve_costs->clast_cost; case FOLD_LEFT_REDUCTION: - switch (GET_MODE_INNER (TYPE_MODE (vectype))) + switch (TYPE_MODE (TREE_TYPE (gimple_get_lhs (stmt_info->stmt)))) { case E_HFmode: case E_BFmode: @@ -15268,14 +15267,10 @@ aarch64_sve_in_loop_reduction_latency (vec_info *vinfo, Advanced SIMD implementation. - If VEC_FLAGS & VEC_ANY_SVE, return the loop carry latency of the - SVE implementation. - - VECTYPE is the type of vector that the vectorizer is considering using - for STMT_INFO, which might be different from the type of vector described - by VEC_FLAGS. */ + SVE implementation. */ static unsigned int aarch64_in_loop_reduction_latency (vec_info *vinfo, stmt_vec_info stmt_info, - tree vectype, unsigned int vec_flags) + unsigned int vec_flags) { const cpu_vector_cost *vec_costs = aarch64_tune_params.vec_costs; const sve_vec_cost *sve_costs = nullptr; @@ -15287,16 +15282,16 @@ aarch64_in_loop_reduction_latency (vec_info *vinfo, stmt_vec_info stmt_info, if (sve_costs) { unsigned int latency - = aarch64_sve_in_loop_reduction_latency (vinfo, stmt_info, vectype, - sve_costs); + = aarch64_sve_in_loop_reduction_latency (vinfo, stmt_info, sve_costs); if (latency) return latency; } /* Handle scalar costs. */ + bool is_float = FLOAT_TYPE_P (TREE_TYPE (gimple_get_lhs (stmt_info->stmt))); if (vec_flags == 0) { - if (FLOAT_TYPE_P (vectype)) + if (is_float) return vec_costs->scalar_fp_stmt_cost; return vec_costs->scalar_int_stmt_cost; } @@ -15305,7 +15300,7 @@ aarch64_in_loop_reduction_latency (vec_info *vinfo, stmt_vec_info stmt_info, with a vector reduction outside the loop. */ const simd_vec_cost *simd_costs = aarch64_simd_vec_costs_for_flags (vec_flags); - if (FLOAT_TYPE_P (vectype)) + if (is_float) return simd_costs->fp_stmt_cost; return simd_costs->int_stmt_cost; } @@ -15382,8 +15377,7 @@ aarch64_detect_vector_stmt_subtype (vec_info *vinfo, vect_cost_for_stmt kind, && sve_costs) { unsigned int latency - = aarch64_sve_in_loop_reduction_latency (vinfo, stmt_info, vectype, - sve_costs); + = aarch64_sve_in_loop_reduction_latency (vinfo, stmt_info, sve_costs); if (latency) return latency; } @@ -15570,8 +15564,7 @@ aarch64_vector_costs::count_ops (unsigned int count, vect_cost_for_stmt kind, && vect_is_reduction (stmt_info)) { unsigned int base - = aarch64_in_loop_reduction_latency (m_vinfo, stmt_info, vectype, - vec_flags); + = aarch64_in_loop_reduction_latency (m_vinfo, stmt_info, vec_flags); if (vect_reduc_type (m_vinfo, stmt_info) == FOLD_LEFT_REDUCTION) { if (aarch64_sve_mode_p (TYPE_MODE (vectype))) -- 2.25.1