[committed] aarch64: Fold aarch64_sve_op_count into aarch64_vec_op_count
Richard Sandiford
richard.sandiford@arm.com
Fri Nov 12 17:36:34 GMT 2021
Later patches make aarch64 use the new vector hooks. We then
only need to track one set of ops for each aarch64_vector_costs
structure. This in turn means that it's more convenient to merge
aarch64_sve_op_count and aarch64_vec_op_count.
The patch also adds issue info and vec flags to aarch64_vec_op_count,
so that the structure is more self-descriptive. This simplifies some
things later.
Tested on aarch64-linux-gnu & applied.
Richard
gcc/
* config/aarch64/aarch64.c (aarch64_sve_op_count): Fold into...
(aarch64_vec_op_count): ...this. Add a constructor.
(aarch64_vec_op_count::vec_flags): New function.
(aarch64_vec_op_count::base_issue_info): Likewise.
(aarch64_vec_op_count::simd_issue_info): Likewise.
(aarch64_vec_op_count::sve_issue_info): Likewise.
(aarch64_vec_op_count::m_issue_info): New member variable.
(aarch64_vec_op_count::m_vec_flags): Likewise.
(aarch64_vector_costs): Add a constructor.
(aarch64_vector_costs::m_sve_ops): Change type to aarch64_vec_op_count.
(aarch64_vector_costs::aarch64_vector_costs): New function.
Initialize m_scalar_ops, m_advsimd_ops and m_sve_ops.
(aarch64_vector_costs::count_ops): Remove vec_flags and
issue_info parameters, using the new aarch64_vec_op_count
functions instead.
(aarch64_vector_costs::add_stmt_cost): Update call accordingly.
(aarch64_sve_op_count::dump): Fold into...
(aarch64_vec_op_count::dump): ..here.
---
gcc/config/aarch64/aarch64.c | 153 ++++++++++++++++++++++-------------
1 file changed, 96 insertions(+), 57 deletions(-)
diff --git a/gcc/config/aarch64/aarch64.c b/gcc/config/aarch64/aarch64.c
index 850288d0e01..c8a3cb38473 100644
--- a/gcc/config/aarch64/aarch64.c
+++ b/gcc/config/aarch64/aarch64.c
@@ -14707,8 +14707,16 @@ aarch64_first_cycle_multipass_dfa_lookahead_guard (rtx_insn *insn,
/* Information about how the CPU would issue the scalar, Advanced SIMD
or SVE version of a vector loop, using the scheme defined by the
aarch64_base_vec_issue_info hierarchy of structures. */
-struct aarch64_vec_op_count
+class aarch64_vec_op_count
{
+public:
+ aarch64_vec_op_count (const aarch64_vec_issue_info *, unsigned int);
+
+ unsigned int vec_flags () const { return m_vec_flags; }
+ const aarch64_base_vec_issue_info *base_issue_info () const;
+ const aarch64_simd_vec_issue_info *simd_issue_info () const;
+ const aarch64_sve_vec_issue_info *sve_issue_info () const;
+
void dump () const;
/* The number of individual "general" operations. See the comments
@@ -14724,23 +14732,71 @@ struct aarch64_vec_op_count
operations, which in the vector code become associated with
reductions. */
unsigned int reduction_latency = 0;
-};
-
-/* Extends aarch64_vec_op_count with SVE-specific information. */
-struct aarch64_sve_op_count : aarch64_vec_op_count
-{
- void dump () const;
/* The number of individual predicate operations. See the comments
in aarch64_sve_vec_issue_info for details. */
unsigned int pred_ops = 0;
+
+private:
+ /* The issue information for the core. */
+ const aarch64_vec_issue_info *m_issue_info;
+
+ /* - If M_VEC_FLAGS is zero then this structure describes scalar code
+ - If M_VEC_FLAGS & VEC_ADVSIMD is nonzero then this structure describes
+ Advanced SIMD code.
+ - If M_VEC_FLAGS & VEC_ANY_SVE is nonzero then this structure describes
+ SVE code. */
+ unsigned int m_vec_flags;
};
+aarch64_vec_op_count::
+aarch64_vec_op_count (const aarch64_vec_issue_info *issue_info,
+ unsigned int vec_flags)
+ : m_issue_info (issue_info),
+ m_vec_flags (vec_flags)
+{
+}
+
+/* Return the base issue information (i.e. the parts that make sense
+ for both scalar and vector code). Return null if we have no issue
+ information. */
+const aarch64_base_vec_issue_info *
+aarch64_vec_op_count::base_issue_info () const
+{
+ if (auto *ret = simd_issue_info ())
+ return ret;
+ if (m_issue_info)
+ return m_issue_info->scalar;
+ return nullptr;
+}
+
+/* If the structure describes vector code and we have associated issue
+ information, return that issue information, otherwise return null. */
+const aarch64_simd_vec_issue_info *
+aarch64_vec_op_count::simd_issue_info () const
+{
+ if (auto *ret = sve_issue_info ())
+ return ret;
+ if (m_issue_info && m_vec_flags)
+ return m_issue_info->advsimd;
+ return nullptr;
+}
+
+/* If the structure describes SVE code and we have associated issue
+ information, return that issue information, otherwise return null. */
+const aarch64_sve_vec_issue_info *
+aarch64_vec_op_count::sve_issue_info () const
+{
+ if (m_issue_info && (m_vec_flags & VEC_ANY_SVE))
+ return m_issue_info->sve;
+ return nullptr;
+}
+
/* Information about vector code that we're in the process of costing. */
class aarch64_vector_costs : public vector_costs
{
public:
- using vector_costs::vector_costs;
+ aarch64_vector_costs (vec_info *, bool);
unsigned int add_stmt_cost (int count, vect_cost_for_stmt kind,
stmt_vec_info stmt_info, tree vectype,
@@ -14752,8 +14808,7 @@ private:
void record_potential_advsimd_unrolling (loop_vec_info);
void analyze_loop_vinfo (loop_vec_info);
void count_ops (unsigned int, vect_cost_for_stmt, stmt_vec_info, tree,
- unsigned int, aarch64_vec_op_count *,
- const aarch64_base_vec_issue_info *, unsigned int);
+ aarch64_vec_op_count *, unsigned int);
fractional_cost adjust_body_cost_sve (const aarch64_vec_issue_info *,
fractional_cost, fractional_cost,
bool, unsigned int, unsigned int *,
@@ -14809,7 +14864,7 @@ private:
/* Used only when vectorizing loops with SVE. It estimates the number and
kind of operations that the SVE loop would contain. */
- aarch64_sve_op_count m_sve_ops;
+ aarch64_vec_op_count m_sve_ops;
/* Used to detect cases in which we end up costing the same load twice,
once to account for results that are actually used and once to account
@@ -14817,6 +14872,15 @@ private:
hash_map<nofree_ptr_hash<_stmt_vec_info>, unsigned int> m_seen_loads;
};
+aarch64_vector_costs::aarch64_vector_costs (vec_info *vinfo,
+ bool costing_for_scalar)
+ : vector_costs (vinfo, costing_for_scalar),
+ m_scalar_ops (aarch64_tune_params.vec_costs->issue_info, 0),
+ m_advsimd_ops (aarch64_tune_params.vec_costs->issue_info, VEC_ADVSIMD),
+ m_sve_ops (aarch64_tune_params.vec_costs->issue_info, VEC_ANY_SVE)
+{
+}
+
/* Implement TARGET_VECTORIZE_CREATE_COSTS. */
vector_costs *
aarch64_vectorize_create_costs (vec_info *vinfo, bool costing_for_scalar)
@@ -15484,36 +15548,21 @@ aarch64_adjust_stmt_cost (vect_cost_for_stmt kind, stmt_vec_info stmt_info,
body of a vector loop. Record issue information relating to the vector
operation in OPS, where OPS is one of m_scalar_ops, m_advsimd_ops
or m_sve_ops; see the comments above those variables for details.
- In addition:
- - VEC_FLAGS is zero if OPS is m_scalar_ops.
-
- - VEC_FLAGS & VEC_ADVSIMD is nonzero if OPS is m_advsimd_ops.
-
- - VEC_FLAGS & VEC_ANY_SVE is nonzero if OPS is m_sve_ops.
-
- ISSUE_INFO provides the scalar, Advanced SIMD or SVE issue information
- associated with OPS and VEC_FLAGS. FACTOR says how many iterations of
- the loop described by VEC_FLAGS would be needed to match one iteration
- of the vector loop in VINFO. */
+ FACTOR says how many iterations of the loop described by VEC_FLAGS would be
+ needed to match one iteration of the vector loop in VINFO. */
void
aarch64_vector_costs::count_ops (unsigned int count, vect_cost_for_stmt kind,
stmt_vec_info stmt_info, tree vectype,
- unsigned int vec_flags,
aarch64_vec_op_count *ops,
- const aarch64_base_vec_issue_info *issue_info,
unsigned int factor)
{
- if (!issue_info)
+ const aarch64_base_vec_issue_info *base_issue = ops->base_issue_info ();
+ if (!base_issue)
return;
-
- const aarch64_simd_vec_issue_info *simd_issue = nullptr;
- if (vec_flags)
- simd_issue = static_cast<const aarch64_simd_vec_issue_info *> (issue_info);
-
- const aarch64_sve_vec_issue_info *sve_issue = nullptr;
- if (vec_flags & VEC_ANY_SVE)
- sve_issue = static_cast<const aarch64_sve_vec_issue_info *> (issue_info);
+ const aarch64_simd_vec_issue_info *simd_issue = ops->simd_issue_info ();
+ const aarch64_sve_vec_issue_info *sve_issue = ops->sve_issue_info ();
+ unsigned int vec_flags = ops->vec_flags ();
/* Calculate the minimum cycles per iteration imposed by a reduction
operation. */
@@ -15608,7 +15657,7 @@ aarch64_vector_costs::count_ops (unsigned int count, vect_cost_for_stmt kind,
}
ops->loads += num_copies;
if (vec_flags || FLOAT_TYPE_P (vectype))
- ops->general_ops += issue_info->fp_simd_load_general_ops * num_copies;
+ ops->general_ops += base_issue->fp_simd_load_general_ops * num_copies;
break;
case vector_store:
@@ -15616,7 +15665,7 @@ aarch64_vector_costs::count_ops (unsigned int count, vect_cost_for_stmt kind,
case scalar_store:
ops->stores += num_copies;
if (vec_flags || FLOAT_TYPE_P (vectype))
- ops->general_ops += issue_info->fp_simd_store_general_ops * num_copies;
+ ops->general_ops += base_issue->fp_simd_store_general_ops * num_copies;
break;
}
@@ -15644,7 +15693,7 @@ aarch64_vector_costs::count_ops (unsigned int count, vect_cost_for_stmt kind,
unsigned int base = (FLOAT_TYPE_P (type)
? sve_issue->fp_cmp_pred_ops
: sve_issue->int_cmp_pred_ops);
- m_sve_ops.pred_ops += base * num_copies;
+ ops->pred_ops += base * num_copies;
}
/* Add any extra overhead associated with LD[234] and ST[234] operations. */
@@ -15670,7 +15719,7 @@ aarch64_vector_costs::count_ops (unsigned int count, vect_cost_for_stmt kind,
&& STMT_VINFO_MEMORY_ACCESS_TYPE (stmt_info) == VMAT_GATHER_SCATTER)
{
unsigned int pairs = CEIL (count, 2);
- m_sve_ops.pred_ops += sve_issue->gather_scatter_pair_pred_ops * pairs;
+ ops->pred_ops += sve_issue->gather_scatter_pair_pred_ops * pairs;
ops->general_ops += sve_issue->gather_scatter_pair_general_ops * pairs;
}
}
@@ -15740,9 +15789,7 @@ aarch64_vector_costs::add_stmt_cost (int count, vect_cost_for_stmt kind,
/* If we're recording a nonzero vector loop body cost for the
innermost loop, also estimate the operations that would need
to be issued by all relevant implementations of the loop. */
- auto *issue_info = aarch64_tune_params.vec_costs->issue_info;
if (loop_vinfo
- && issue_info
&& m_vec_flags
&& where == vect_body
&& (!LOOP_VINFO_LOOP (loop_vinfo)->inner || in_inner_loop_p)
@@ -15750,26 +15797,24 @@ aarch64_vector_costs::add_stmt_cost (int count, vect_cost_for_stmt kind,
&& stmt_cost != 0)
{
/* Record estimates for the scalar code. */
- count_ops (count, kind, stmt_info, vectype, 0, &m_scalar_ops,
- issue_info->scalar, vect_nunits_for_cost (vectype));
+ count_ops (count, kind, stmt_info, vectype, &m_scalar_ops,
+ vect_nunits_for_cost (vectype));
- if (aarch64_sve_mode_p (m_vinfo->vector_mode) && issue_info->sve)
+ if (aarch64_sve_mode_p (m_vinfo->vector_mode)
+ && m_sve_ops.base_issue_info ())
{
/* Record estimates for a possible Advanced SIMD version
of the SVE code. */
- count_ops (count, kind, stmt_info, vectype, VEC_ADVSIMD,
- &m_advsimd_ops, issue_info->advsimd,
+ count_ops (count, kind, stmt_info, vectype, &m_advsimd_ops,
aarch64_estimated_sve_vq ());
/* Record estimates for the SVE code itself. */
- count_ops (count, kind, stmt_info, vectype, VEC_ANY_SVE,
- &m_sve_ops, issue_info->sve, 1);
+ count_ops (count, kind, stmt_info, vectype, &m_sve_ops, 1);
}
else
/* Record estimates for the Advanced SIMD code. Treat SVE like
Advanced SIMD if the CPU has no specific SVE costs. */
- count_ops (count, kind, stmt_info, vectype, VEC_ADVSIMD,
- &m_advsimd_ops, issue_info->advsimd, 1);
+ count_ops (count, kind, stmt_info, vectype, &m_advsimd_ops, 1);
}
/* If we're applying the SVE vs. Advanced SIMD unrolling heuristic,
@@ -15793,19 +15838,13 @@ aarch64_vec_op_count::dump () const
" store operations = %d\n", stores);
dump_printf_loc (MSG_NOTE, vect_location,
" general operations = %d\n", general_ops);
+ if (sve_issue_info ())
+ dump_printf_loc (MSG_NOTE, vect_location,
+ " predicate operations = %d\n", pred_ops);
dump_printf_loc (MSG_NOTE, vect_location,
" reduction latency = %d\n", reduction_latency);
}
-/* Dump information about the structure. */
-void
-aarch64_sve_op_count::dump () const
-{
- aarch64_vec_op_count::dump ();
- dump_printf_loc (MSG_NOTE, vect_location,
- " predicate operations = %d\n", pred_ops);
-}
-
/* Use ISSUE_INFO to estimate the minimum number of cycles needed to issue
the operations described by OPS. This is a very simplistic model! */
static fractional_cost
--
2.25.1
More information about the Gcc-patches
mailing list