OpenE2K/gcc
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Fall back to masked_gather_load/masked_scatter_store

Browse Source 
This adds a fallback to the masked_ variants for gather_load
and scatter_store if the latter are not available.

2021-08-17  Richard Biener  <rguenther@suse.de>

	* optabs-query.c (supports_vec_gather_load_p): Also check
	for masked optabs.
	(supports_vec_scatter_store_p): Likewise.
	* tree-vect-data-refs.c (vect_gather_scatter_fn_p): Fall
	back to masked variants if non-masked are not supported.
	* tree-vect-patterns.c (vect_recog_gather_scatter_pattern):
	When we need to use masked gather/scatter but do not have
	a mask set up a constant true one.
	* tree-vect-stmts.c (vect_check_scalar_mask): Also allow
	non-SSA_NAME masks.
This commit is contained in:
Richard Biener 2021-08-17 15:50:31 +02:00
parent 817766f4dd
commit e7e17be3be
4 changed files with 28 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
gcc/optabs-query.c
View File

@ -740,7 +740,8 @@ supports_vec_gather_load_p ()
this_fn_optabs->supports_vec_gather_load_cached = true;
this_fn_optabs->supports_vec_gather_load
= supports_vec_convert_optab_p (gather_load_optab);
= (supports_vec_convert_optab_p (gather_load_optab)
|| supports_vec_convert_optab_p (mask_gather_load_optab));
return this_fn_optabs->supports_vec_gather_load;
}
@ -757,7 +758,8 @@ supports_vec_scatter_store_p ()
this_fn_optabs->supports_vec_scatter_store_cached = true;
this_fn_optabs->supports_vec_scatter_store
= supports_vec_convert_optab_p (scatter_store_optab);
= (supports_vec_convert_optab_p (scatter_store_optab)
|| supports_vec_convert_optab_p (mask_scatter_store_optab));
return this_fn_optabs->supports_vec_scatter_store;
}

22
gcc/tree-vect-data-refs.c
View File

@ -3735,11 +3735,17 @@ vect_gather_scatter_fn_p (vec_info *vinfo, bool read_p, bool masked_p,
return false;
/* Work out which function we need. */
internal_fn ifn;
internal_fn ifn, alt_ifn;
if (read_p)
ifn = masked_p ? IFN_MASK_GATHER_LOAD : IFN_GATHER_LOAD;
{
ifn = masked_p ? IFN_MASK_GATHER_LOAD : IFN_GATHER_LOAD;
alt_ifn = IFN_MASK_GATHER_LOAD;
}
else
ifn = masked_p ? IFN_MASK_SCATTER_STORE : IFN_SCATTER_STORE;
{
ifn = masked_p ? IFN_MASK_SCATTER_STORE : IFN_SCATTER_STORE;
alt_ifn = IFN_MASK_SCATTER_STORE;
}
for (;;)
{
@ -3755,6 +3761,16 @@ vect_gather_scatter_fn_p (vec_info *vinfo, bool read_p, bool masked_p,
*offset_vectype_out = offset_vectype;
return true;
}
else if (!masked_p
&& internal_gather_scatter_fn_supported_p (alt_ifn, vectype,
memory_type,
offset_vectype,
scale))
{
*ifn_out = alt_ifn;
*offset_vectype_out = offset_vectype;
return true;
}
if (TYPE_PRECISION (offset_type) >= POINTER_SIZE
&& TYPE_PRECISION (offset_type) >= element_bits)

7
gcc/tree-vect-patterns.c
View File

@ -4820,6 +4820,9 @@ vect_recog_gather_scatter_pattern (vec_info *vinfo,
if (mask)
mask = vect_convert_mask_for_vectype (mask, gs_vectype, stmt_info,
loop_vinfo);
else if (gs_info.ifn == IFN_MASK_SCATTER_STORE
|| gs_info.ifn == IFN_MASK_GATHER_LOAD)
mask = build_int_cst (TREE_TYPE (truth_type_for (gs_vectype)), -1);
/* Get the invariant base and non-invariant offset, converting the
latter to the same width as the vector elements. */
@ -4847,11 +4850,11 @@ vect_recog_gather_scatter_pattern (vec_info *vinfo,
{
tree rhs = vect_get_store_rhs (stmt_info);
if (mask != NULL)
pattern_stmt = gimple_build_call_internal (IFN_MASK_SCATTER_STORE, 5,
pattern_stmt = gimple_build_call_internal (gs_info.ifn, 5,
base, offset, scale, rhs,
mask);
else
pattern_stmt = gimple_build_call_internal (IFN_SCATTER_STORE, 4,
pattern_stmt = gimple_build_call_internal (gs_info.ifn, 4,
base, offset, scale, rhs);
}
gimple_call_set_nothrow (pattern_stmt, true);

8
gcc/tree-vect-stmts.c
View File

@ -2510,14 +2510,6 @@ vect_check_scalar_mask (vec_info *vinfo, stmt_vec_info stmt_info,
return false;
}
if (TREE_CODE (*mask) != SSA_NAME)
{
if (dump_enabled_p ())
dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
"mask argument is not an SSA name.\n");
return false;
}
/* If the caller is not prepared for adjusting an external/constant
SLP mask vector type fail. */
if (slp_node