Hi!
While my r15-1001-g4cf2de9b5268224 PCH PIE power fix change decreased the
.data section sizes (219792 -> 189336), it increased the size of already
huge rs6000_init_generated_builtins generated function, from 218328
to 228668 bytes. That is because there are thousands of array references
to global arrays and we keep constructing the addresses of the arrays
again and again.
Ideally some optimization would figure out we have a single function which
has
461 rs6000_overload_info
1257 rs6000_builtin_info_fntype
1768 rs6000_builtin_decls
2548 rs6000_instance_info_fntype
array references and that maybe it might be a good idea to just preload
the addresses of those arrays into some register if it decreases code size
and doesn't slow things down.
The function actually is called just once and is huge, so code size is even
more important than speed, which is dominated by all the GC allocations
anyway.
Until that is done, here is a slightly cleaner version of the hack, which
makes the function noipa (so that LTO doesn't undo it) for GCC 8.1+ and
passes the 4 arrays as arguments to the function from the caller.
This decreases the function size from 228668 bytes to 207572 bytes.
Bootstrapped/regtested on powerpc64le-linux, ok for trunk?
2024-06-17 Jakub Jelinek <ja...@redhat.com>
PR target/115324
* config/rs6000/rs6000-gen-builtins.cc (write_decls): Change
declaration of rs6000_init_generated_builtins from no arguments
to 4 pointer arguments.
(write_init_bif_table): Change rs6000_builtin_info_fntype to
builtin_info_fntype and rs6000_builtin_decls to builtin_decls.
(write_init_ovld_table): Change rs6000_instance_info_fntype to
instance_info_fntype, rs6000_builtin_decls to builtin_decls and
rs6000_overload_info to overload_info.
(write_init_file): Add __noipa__ attribute to
rs6000_init_generated_builtins for GCC 8.1+ and change the function
from no arguments to 4 pointer arguments. Change rs6000_builtin_decls
to builtin_decls.
* config/rs6000/rs6000-builtin.cc (rs6000_init_builtins): Adjust
rs6000_init_generated_builtins caller.
--- gcc/config/rs6000/rs6000-gen-builtins.cc.jj 2024-06-03 23:11:02.662631144
+0200
+++ gcc/config/rs6000/rs6000-gen-builtins.cc 2024-06-03 23:38:31.727620920
+0200
@@ -2376,7 +2376,10 @@ write_decls (void)
"rs6000_instance_info_fntype[RS6000_INST_MAX];\n");
fprintf (header_file, "extern ovldrecord rs6000_overload_info[];\n\n");
- fprintf (header_file, "extern void rs6000_init_generated_builtins ();\n\n");
+ fprintf (header_file,
+ "extern void rs6000_init_generated_builtins (tree *, tree *,\n");
+ fprintf (header_file,
+ "\t\t\t\t\t ovldrecord *, tree *);\n\n");
fprintf (header_file,
"extern bool rs6000_builtin_is_supported (rs6000_gen_builtins);\n");
fprintf (header_file,
@@ -2651,7 +2654,7 @@ write_init_bif_table (void)
for (int i = 0; i <= curr_bif; i++)
{
fprintf (init_file,
- " rs6000_builtin_info_fntype[RS6000_BIF_%s]"
+ " builtin_info_fntype[RS6000_BIF_%s]"
"\n = %s;\n",
bifs[i].idname, bifs[i].fndecl);
@@ -2678,7 +2681,7 @@ write_init_bif_table (void)
}
fprintf (init_file,
- " rs6000_builtin_decls[(int)RS6000_BIF_%s] = t\n",
+ " builtin_decls[(int)RS6000_BIF_%s] = t\n",
bifs[i].idname);
fprintf (init_file,
" = add_builtin_function (\"%s\",\n",
@@ -2719,7 +2722,7 @@ write_init_bif_table (void)
fprintf (init_file, " }\n");
fprintf (init_file, " else\n");
fprintf (init_file, " {\n");
- fprintf (init_file, " rs6000_builtin_decls"
+ fprintf (init_file, " builtin_decls"
"[(int)RS6000_BIF_%s] = NULL_TREE;\n", bifs[i].idname);
fprintf (init_file, " }\n");
}
@@ -2740,7 +2743,7 @@ write_init_ovld_table (void)
for (int i = 0; i <= curr_ovld; i++)
{
fprintf (init_file,
- " rs6000_instance_info_fntype[RS6000_INST_%s]"
+ " instance_info_fntype[RS6000_INST_%s]"
"\n = %s;\n",
ovlds[i].ovld_id_name, ovlds[i].fndecl);
@@ -2772,7 +2775,7 @@ write_init_ovld_table (void)
}
fprintf (init_file,
- " rs6000_builtin_decls[(int)RS6000_OVLD_%s] = t\n",
+ " builtin_decls[(int)RS6000_OVLD_%s] = t\n",
stanza->stanza_id);
fprintf (init_file,
" = add_builtin_function (\"%s\",\n",
@@ -2793,7 +2796,7 @@ write_init_ovld_table (void)
fprintf (init_file, "\n");
fprintf (init_file,
- " rs6000_overload_info[RS6000_OVLD_%s - base]"
+ " overload_info[RS6000_OVLD_%s - base]"
".first_instance\n",
stanza->stanza_id);
fprintf (init_file,
@@ -2826,19 +2829,30 @@ write_init_file (void)
write_bif_static_init ();
write_ovld_static_init ();
+ /* The reason to pass pointers to the function instead of accessing
+ the rs6000_{{builtin,instance}_info_fntype,overload_info,builtin_decls}
+ arrays directly is to decrease size of the already large function and
+ noipa prevents the compiler with LTO to undo that optimization. */
+ fprintf (init_file, "#if GCC_VERSION >= 8001\n");
+ fprintf (init_file, "__attribute__((__noipa__))\n");
+ fprintf (init_file, "#endif\n");
fprintf (init_file, "void\n");
- fprintf (init_file, "rs6000_init_generated_builtins ()\n");
+ fprintf (init_file,
+ "rs6000_init_generated_builtins (tree *builtin_info_fntype,\n");
+ fprintf (init_file, "\t\t\t\ttree *instance_info_fntype,\n");
+ fprintf (init_file, "\t\t\t\tovldrecord *overload_info,\n");
+ fprintf (init_file, "\t\t\t\ttree *builtin_decls)\n");
fprintf (init_file, "{\n");
fprintf (init_file, " tree t;\n");
rbt_inorder_callback (&fntype_rbt, fntype_rbt.rbt_root, write_fntype_init);
fprintf (init_file, "\n");
fprintf (init_file,
- " rs6000_builtin_decls[RS6000_BIF_NONE] = NULL_TREE;\n");
+ " builtin_decls[RS6000_BIF_NONE] = NULL_TREE;\n");
fprintf (init_file,
- " rs6000_builtin_decls[RS6000_BIF_MAX] = NULL_TREE;\n");
+ " builtin_decls[RS6000_BIF_MAX] = NULL_TREE;\n");
fprintf (init_file,
- " rs6000_builtin_decls[RS6000_OVLD_NONE] = NULL_TREE;\n\n");
+ " builtin_decls[RS6000_OVLD_NONE] = NULL_TREE;\n\n");
write_init_bif_table ();
write_init_ovld_table ();
--- gcc/config/rs6000/rs6000-builtin.cc.jj 2024-06-03 23:11:02.662631144
+0200
+++ gcc/config/rs6000/rs6000-builtin.cc 2024-06-03 23:39:39.699745339 +0200
@@ -835,7 +835,10 @@ rs6000_init_builtins (void)
TYPE_QUAL_CONST));
/* Execute the autogenerated initialization code for builtins. */
- rs6000_init_generated_builtins ();
+ rs6000_init_generated_builtins (rs6000_builtin_info_fntype,
+ rs6000_instance_info_fntype,
+ rs6000_overload_info,
+ rs6000_builtin_decls);
if (TARGET_DEBUG_BUILTIN)
{
Jakub
