From 53aedcce09b10467e60b9717223191f28e0360a3 Mon Sep 17 00:00:00 2001 From: Michael Ploujnikov Date: Fri, 30 Nov 2018 22:27:11 +0000 Subject: [PATCH] Minimize clone counter memory usage in create_virtual_clone. Based on Martin Jambour's suggestion: https://gcc.gnu.org/ml/gcc-patches/2018-09/msg00111.html gcc: * cgraph.h (clone_function_name): Add a variant that takes a tree decl. * cgraph.h (cgraph_node::create_virtual_clone): Add a new argument: num_suffix. * cgraphclones.c (cgraph_node::create_virtual_clone): Pass num_suffix to clone_function_name. (clone_function_name): Add a variant that takes a tree decl. * ipa-cp.c (create_specialized_node): Keep track of clone counters in clone_num_suffixes hash map. (ipcp_driver): Free the counter hash map. * ipa-hsa.c (process_hsa_functions): Creates at most one hsa clone per function. From-SVN: r266692 --- gcc/ChangeLog | 16 ++++++++++++++++ gcc/cgraph.h | 10 +++++++--- gcc/cgraphclones.c | 25 ++++++++++++++++++++----- gcc/ipa-cp.c | 10 +++++++++- gcc/ipa-hsa.c | 4 ++-- 5 files changed, 54 insertions(+), 11 deletions(-) diff --git a/gcc/ChangeLog b/gcc/ChangeLog index e47e2b17cb8..873d1c8d25f 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,3 +1,19 @@ +2018-11-30 Michael Ploujnikov + + Minimize clone counter memory usage in create_virtual_clone. + * cgraph.h (clone_function_name): Add a variant that takes a + tree decl. + * cgraph.h (cgraph_node::create_virtual_clone): Add a new + argument: num_suffix. + * cgraphclones.c (cgraph_node::create_virtual_clone): Pass + num_suffix to clone_function_name. + (clone_function_name): Add a variant that takes a tree decl. + * ipa-cp.c (create_specialized_node): Keep track of clone + counters in clone_num_suffixes hash map. + (ipcp_driver): Free the counter hash map. + * ipa-hsa.c (process_hsa_functions): Creates at most one hsa + clone per function. + 2018-11-30 Michael Ploujnikov Make function assembly more independent. diff --git a/gcc/cgraph.h b/gcc/cgraph.h index dd1e8fdbe4a..b8e23cc338a 100644 --- a/gcc/cgraph.h +++ b/gcc/cgraph.h @@ -968,11 +968,13 @@ public: cgraph_node *new_inlined_to, bitmap args_to_skip, const char *suffix = NULL); - /* Create callgraph node clone with new declaration. The actual body will - be copied later at compilation stage. */ + /* Create callgraph node clone with new declaration. The actual body will be + copied later at compilation stage. The name of the new clone will be + constructed from the name of the original node, SUFFIX and NUM_SUFFIX. */ cgraph_node *create_virtual_clone (vec redirect_callers, vec *tree_map, - bitmap args_to_skip, const char * suffix); + bitmap args_to_skip, const char * suffix, + unsigned num_suffix); /* cgraph node being removed from symbol table; see if its entry can be replaced by other inline clone. */ @@ -2386,6 +2388,8 @@ tree clone_function_name_numbered (const char *name, const char *suffix); tree clone_function_name_numbered (tree decl, const char *suffix); tree clone_function_name (const char *name, const char *suffix, unsigned long number); +tree clone_function_name (tree decl, const char *suffix, + unsigned long number); tree clone_function_name (tree decl, const char *suffix); void tree_function_versioning (tree, tree, vec *, diff --git a/gcc/cgraphclones.c b/gcc/cgraphclones.c index a45722f1fc7..f8076920fff 100644 --- a/gcc/cgraphclones.c +++ b/gcc/cgraphclones.c @@ -572,6 +572,19 @@ clone_function_name (const char *name, const char *suffix, return get_identifier (tmp_name); } +/* Return a new assembler name for a clone of DECL. Apart from the + string SUFFIX, the new name will end with the specified NUMBER. If + clone numbering is not needed then the two argument + clone_function_name should be used instead. */ + +tree +clone_function_name (tree decl, const char *suffix, + unsigned long number) +{ + return clone_function_name ( + IDENTIFIER_POINTER (DECL_ASSEMBLER_NAME (decl)), suffix, number); +} + /* Return a new assembler name ending with the string SUFFIX for a clone of DECL. */ @@ -599,8 +612,9 @@ clone_function_name (tree decl, const char *suffix) } -/* Create callgraph node clone with new declaration. The actual body will - be copied later at compilation stage. +/* Create callgraph node clone with new declaration. The actual body will be + copied later at compilation stage. The name of the new clone will be + constructed from the name of the original node, SUFFIX and NUM_SUFFIX. TODO: after merging in ipa-sra use function call notes instead of args_to_skip bitmap interface. @@ -608,7 +622,8 @@ clone_function_name (tree decl, const char *suffix) cgraph_node * cgraph_node::create_virtual_clone (vec redirect_callers, vec *tree_map, - bitmap args_to_skip, const char * suffix) + bitmap args_to_skip, const char * suffix, + unsigned num_suffix) { tree old_decl = decl; cgraph_node *new_node = NULL; @@ -643,8 +658,8 @@ cgraph_node::create_virtual_clone (vec redirect_callers, strcpy (name + len + 1, suffix); name[len] = '.'; DECL_NAME (new_decl) = get_identifier (name); - SET_DECL_ASSEMBLER_NAME (new_decl, clone_function_name_numbered (old_decl, - suffix)); + SET_DECL_ASSEMBLER_NAME (new_decl, + clone_function_name (old_decl, suffix, num_suffix)); SET_DECL_RTL (new_decl, NULL); new_node = create_clone (new_decl, count, false, diff --git a/gcc/ipa-cp.c b/gcc/ipa-cp.c index 6b9dc8cb08f..84ca688fce3 100644 --- a/gcc/ipa-cp.c +++ b/gcc/ipa-cp.c @@ -376,6 +376,9 @@ static profile_count max_count; static long overall_size, max_new_size; +/* Node to unique clone suffix number map. */ +static hash_map *clone_num_suffixes; + /* Return the param lattices structure corresponding to the Ith formal parameter of the function described by INFO. */ static inline struct ipcp_param_lattices * @@ -3829,8 +3832,11 @@ create_specialized_node (struct cgraph_node *node, } } + unsigned &suffix_counter = clone_num_suffixes->get_or_insert (node); new_node = node->create_virtual_clone (callers, replace_trees, - args_to_skip, "constprop"); + args_to_skip, "constprop", + suffix_counter); + suffix_counter++; bool have_self_recursive_calls = !self_recursive_calls.is_empty (); for (unsigned j = 0; j < self_recursive_calls.length (); j++) @@ -5044,6 +5050,7 @@ ipcp_driver (void) ipa_check_create_node_params (); ipa_check_create_edge_args (); + clone_num_suffixes = new hash_map; if (dump_file) { @@ -5065,6 +5072,7 @@ ipcp_driver (void) ipcp_store_vr_results (); /* Free all IPCP structures. */ + delete clone_num_suffixes; free_toporder_info (&topo); delete edge_clone_summaries; edge_clone_summaries = NULL; diff --git a/gcc/ipa-hsa.c b/gcc/ipa-hsa.c index 7c6ceaab8fe..63b41eabc9a 100644 --- a/gcc/ipa-hsa.c +++ b/gcc/ipa-hsa.c @@ -88,7 +88,7 @@ process_hsa_functions (void) continue; cgraph_node *clone = node->create_virtual_clone (vec (), - NULL, NULL, "hsa"); + NULL, NULL, "hsa", 0); TREE_PUBLIC (clone->decl) = TREE_PUBLIC (node->decl); clone->externally_visible = node->externally_visible; @@ -109,7 +109,7 @@ process_hsa_functions (void) continue; cgraph_node *clone = node->create_virtual_clone (vec (), - NULL, NULL, "hsa"); + NULL, NULL, "hsa", 0); TREE_PUBLIC (clone->decl) = TREE_PUBLIC (node->decl); clone->externally_visible = node->externally_visible; -- 2.30.2