2019-11-12 11:08:04 +01:00
; Parameter options of the compiler.
2022-01-03 10:42:10 +01:00
; Copyright (C) 2019-2022 Free Software Foundation, Inc.
2019-11-12 11:08:04 +01:00
;
; This file is part of GCC.
;
; GCC is free software; you can redistribute it and/or modify it under
; the terms of the GNU General Public License as published by the Free
; Software Foundation; either version 3, or (at your option) any later
; version.
;
; GCC is distributed in the hope that it will be useful, but WITHOUT ANY
; WARRANTY; without even the implied warranty of MERCHANTABILITY or
; FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
; for more details.
;
; You should have received a copy of the GNU General Public License
; along with GCC; see the file COPYING3. If not see
; <http://www.gnu.org/licenses/>.
; See the GCC internals manual (options.texi) for a description of this file's format.
; Please try to keep this file in ASCII collating order.
-param=align-loop-iterations=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_align_loop_iterations) Init(4) Param Optimization
2019-11-12 11:08:04 +01:00
Loops iterating at least selected number of iterations will get loop alignment.
-param=align-threshold=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_align_threshold) Init(100) IntegerRange(1, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
Select fraction of the maximal frequency of executions of basic block in function given basic block get alignment.
-param=asan-globals=
Common Joined UInteger Var(param_asan_globals) Init(1) IntegerRange(0, 1) Param
Enable asan globals protection.
-param=asan-instrument-allocas=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_asan_protect_allocas) Init(1) IntegerRange(0, 1) Param Optimization
2019-11-12 11:08:04 +01:00
Enable asan allocas/VLAs protection.
-param=asan-instrument-reads=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_asan_instrument_reads) Init(1) IntegerRange(0, 1) Param Optimization
2019-11-12 11:08:04 +01:00
Enable asan load operations protection.
-param=asan-instrument-writes=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_asan_instrument_writes) Init(1) IntegerRange(0, 1) Param Optimization
2019-11-12 11:08:04 +01:00
Enable asan store operations protection.
-param=asan-instrumentation-with-call-threshold=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_asan_instrumentation_with_call_threshold) Init(7000) Param Optimization
2019-11-12 11:08:04 +01:00
Use callbacks instead of inline code if number of accesses in function becomes greater or equal to this number.
-param=asan-memintrin=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_asan_memintrin) Init(1) IntegerRange(0, 1) Param Optimization
2019-11-12 11:08:04 +01:00
Enable asan builtin functions protection.
-param=asan-stack=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_asan_stack) Init(1) IntegerRange(0, 1) Param Optimization
2019-11-12 11:08:04 +01:00
Enable asan stack protection.
-param=asan-use-after-return=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_asan_use_after_return) Init(1) IntegerRange(0, 1) Param Optimization
2019-11-12 11:08:04 +01:00
Enable asan detection of use-after-return bugs.
2020-11-25 17:31:43 +01:00
-param=hwasan-instrument-stack=
Common Joined UInteger Var(param_hwasan_instrument_stack) Init(1) IntegerRange(0, 1) Param Optimization
Enable hwasan instrumentation of statically sized stack-allocated variables.
-param=hwasan-random-frame-tag=
Common Joined UInteger Var(param_hwasan_random_frame_tag) Init(1) IntegerRange(0, 1) Param Optimization
Use random base tag for each frame, as opposed to base always zero.
-param=hwasan-instrument-allocas=
Common Joined UInteger Var(param_hwasan_instrument_allocas) Init(1) IntegerRange(0, 1) Param Optimization
Enable hwasan instrumentation of allocas/VLAs.
-param=hwasan-instrument-reads=
Common Joined UInteger Var(param_hwasan_instrument_reads) Init(1) IntegerRange(0, 1) Param Optimization
Enable hwasan instrumentation of load operations.
-param=hwasan-instrument-writes=
Common Joined UInteger Var(param_hwasan_instrument_writes) Init(1) IntegerRange(0, 1) Param Optimization
Enable hwasan instrumentation of store operations.
-param=hwasan-instrument-mem-intrinsics=
Common Joined UInteger Var(param_hwasan_instrument_mem_intrinsics) Init(1) IntegerRange(0, 1) Param Optimization
Enable hwasan instrumentation of builtin functions.
2019-11-12 11:08:04 +01:00
-param=avg-loop-niter=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_avg_loop_niter) Init(10) IntegerRange(1, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
Average number of iterations of a loop.
-param=avoid-fma-max-bits=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_avoid_fma_max_bits) IntegerRange(0, 512) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of bits for which we avoid creating FMAs.
-param=builtin-expect-probability=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_builtin_expect_probability) Init(90) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
Set the estimated probability in percentage for builtin expect. The default value is 90% probability.
-param=builtin-string-cmp-inline-length=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_builtin_string_cmp_inline_length) Init(3) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum length of a constant string for a builtin string cmp call eligible for inlining. The default value is 3.
-param=case-values-threshold=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_case_values_threshold) Param Optimization
2019-11-12 11:08:04 +01:00
The smallest number of different values for which it is best to use a jump-table instead of a tree of conditional branches, if 0, use the default for the machine.
-param=comdat-sharing-probability=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_comdat_sharing_probability) Init(20) Param Optimization
2019-11-12 11:08:04 +01:00
Probability that COMDAT function will be shared with different compilation unit.
-param=cxx-max-namespaces-for-diagnostic-help=
Common Joined UInteger Var(param_cxx_max_namespaces_for_diagnostic_help) Init(1000) Param
Maximum number of namespaces to search for alternatives when name lookup fails.
-param=dse-max-alias-queries-per-store=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_dse_max_alias_queries_per_store) Init(256) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of queries into the alias oracle per store.
-param=dse-max-object-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_dse_max_object_size) Init(256) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum size (in bytes) of objects tracked bytewise by dead store elimination.
-param=early-inlining-insns=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_early_inlining_insns) Init(6) Optimization Param
Maximal estimated growth of function body caused by early inlining of single call.
2019-11-12 11:08:04 +01:00
2021-06-07 19:18:55 +02:00
-param=evrp-sparse-threshold=
Common Joined UInteger Var(param_evrp_sparse_threshold) Init(800) Optimization Param
Maximum number of basic blocks before EVRP uses a sparse cache.
2021-09-29 23:25:50 +02:00
-param=evrp-switch-limit=
Common Joined UInteger Var(param_evrp_switch_limit) Init(50) Optimization Param
Maximum number of outgoing edges in a switch before EVRP will not process it.
2019-11-12 11:08:04 +01:00
-param=fsm-scale-path-blocks=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_fsm_scale_path_blocks) Init(3) IntegerRange(1, 10) Param Optimization
2019-11-12 11:08:04 +01:00
Scale factor to apply to the number of blocks in a threading path when comparing to the number of (scaled) statements.
-param=fsm-scale-path-stmts=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_fsm_scale_path_stmts) Init(2) IntegerRange(1, 10) Param Optimization
2019-11-12 11:08:04 +01:00
Scale factor to apply to the number of statements in a threading path when comparing to the number of (scaled) blocks.
-param=gcse-after-reload-critical-fraction=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_gcse_after_reload_critical_fraction) Init(10) Param Optimization
2019-11-12 11:08:04 +01:00
The threshold ratio of critical edges execution count that permit performing redundancy elimination after reload.
-param=gcse-after-reload-partial-fraction=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_gcse_after_reload_partial_fraction) Init(3) Param Optimization
2019-11-12 11:08:04 +01:00
The threshold ratio for performing partial redundancy elimination after reload.
-param=gcse-cost-distance-ratio=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_gcse_cost_distance_ratio) Init(10) Param Optimization
2019-11-12 11:08:04 +01:00
Scaling factor in calculation of maximum distance an expression can be moved by GCSE optimizations.
-param=gcse-unrestricted-cost=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_gcse_unrestricted_cost) Init(3) Param Optimization
2019-11-12 11:08:04 +01:00
Cost at which GCSE optimizations will not constraint the distance an expression can travel.
-param=ggc-min-expand=
Common Joined UInteger Var(param_ggc_min_expand) Init(30) Param
Minimum heap expansion to trigger garbage collection, as a percentage of the total size of the heap.
-param=ggc-min-heapsize=
Common Joined UInteger Var(param_ggc_min_heapsize) Init(4096) Param
Minimum heap size before we start collecting garbage, in kilobytes.
-param=gimple-fe-computed-hot-bb-threshold=
Common Joined UInteger Var(param_gimple_fe_computed_hot_bb_threshold) Param
The number of executions of a basic block which is considered hot. The parameter is used only in GIMPLE FE.
-param=graphite-allow-codegen-errors=
Common Joined UInteger Var(param_graphite_allow_codegen_errors) IntegerRange(0, 1) Param
Whether codegen errors should be ICEs when -fchecking.
-param=graphite-max-arrays-per-scop=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_graphite_max_arrays_per_scop) Init(100) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of arrays per SCoP.
-param=graphite-max-nb-scop-params=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_graphite_max_nb_scop_params) Init(10) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of parameters in a SCoP.
-param=hash-table-verification-limit=
Common Joined UInteger Var(param_hash_table_verification_limit) Init(10) Param
The number of elements for which hash table verification is done for each searched element.
-param=hot-bb-count-fraction=
Common Joined UInteger Var(param_hot_bb_count_fraction) Init(10000) Param
The denominator n of fraction 1/n of the maximal execution count of a basic block in the entire program that a basic block needs to at least have in order to be considered hot (used in non-LTO mode).
-param=hot-bb-count-ws-permille=
Common Joined UInteger Var(param_hot_bb_count_ws_permille) Init(990) IntegerRange(0, 1000) Param
The number of most executed permilles of the profiled execution of the entire program to which the execution count of a basic block must be part of in order to be considered hot (used in LTO mode).
-param=hot-bb-frequency-fraction=
Common Joined UInteger Var(param_hot_bb_frequency_fraction) Init(1000) Param
The denominator n of fraction 1/n of the execution frequency of the entry block of a function that a basic block of this function needs to at least have in order to be considered hot.
-param=inline-heuristics-hint-percent=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_inline_heuristics_hint_percent) Init(200) Optimization IntegerRange(100, 1000000) Param
2019-11-12 11:08:04 +01:00
The scale (in percents) applied to inline-insns-single and auto limits when heuristics hints that inlining is very profitable.
-param=inline-min-speedup=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_inline_min_speedup) Init(30) Optimization IntegerRange(0, 100) Param
2019-11-12 11:08:04 +01:00
The minimal estimated speedup allowing inliner to ignore inline-insns-single and inline-insns-auto.
-param=inline-unit-growth=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_inline_unit_growth) Init(40) Optimization Param
2019-11-12 11:08:04 +01:00
How much can given compilation unit grow because of the inlining (in percent).
-param=integer-share-limit=
Common Joined UInteger Var(param_integer_share_limit) Init(251) IntegerRange(2, 65536) Param
The upper bound for sharing integer constants.
-param=ipa-cp-eval-threshold=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_ipa_cp_eval_threshold) Init(500) Param Optimization
2019-11-12 11:08:04 +01:00
Threshold ipa-cp opportunity evaluation that is still considered beneficial to clone.
-param=ipa-cp-loop-hint-bonus=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_ipa_cp_loop_hint_bonus) Init(64) Param Optimization
2019-11-12 11:08:04 +01:00
Compile-time bonus IPA-CP assigns to candidates which make loop bounds or strides known.
2019-12-02 07:37:30 +01:00
-param=ipa-cp-max-recursive-depth=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_ipa_cp_max_recursive_depth) Init(8) Param Optimization
2019-12-02 07:37:30 +01:00
Maximum depth of recursive cloning for self-recursive function.
-param=ipa-cp-min-recursive-probability=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_ipa_cp_min_recursive_probability) Init(2) Param Optimization
2019-12-02 07:37:30 +01:00
Recursive cloning only when the probability of call being executed exceeds the parameter.
ipa-cp: Propagation boost for recursion generated values
Recursive call graph edges, even when they are hot and important for
the compiled program, can never have frequency bigger than one, even
when the actual time savings in the next recursion call are not
realized just once but depend on the depth of recursion. The current
IPA-CP effect propagation code did not take that into account and just
used the frequency, thus severely underestimating the effect.
This patch artificially boosts values taking part in such calls. If a
value feeds into itself through a recursive call, the frequency of the
edge is multiplied by a parameter with default value of 6, basically
assuming that the recursion will take place 6 times. This value can
of course be subject to change.
Moreover, values which do not feed into themselves but which were
generated for a self-recursive call with an arithmetic
pass-function (aka the 548.exchange "hack" which however is generally
applicable for recursive functions which count the recursion depth in
a parameter) have the edge frequency multiplied as many times as there
are generated values in the chain. In essence, we will assume they
are all useful.
This patch partially fixes the current situation when we fail to
optimize 548.exchange with PGO. In the benchmark one recursive edge
count overwhelmingly dominates all other counts in the program and so
we fail to perform the first cloning (for the nonrecursive entry call)
because it looks totally insignificant.
gcc/ChangeLog:
2021-07-16 Martin Jambor <mjambor@suse.cz>
* params.opt (ipa-cp-recursive-freq-factor): New.
* ipa-cp.c (ipcp_value): Switch to inline initialization. New members
scc_no, self_recursion_generated_level, same_scc and
self_recursion_generated_p.
(ipcp_lattice::add_value): Replaced parameter unlimited with
same_lat_gen_level, usit it determine limit of values and store it to
the value.
(ipcp_lattice<valtype>::print): Dump the new fileds.
(allocate_and_init_ipcp_value): Take same_lat_gen_level as a new
parameter and store it to the new value.
(self_recursively_generated_p): Removed.
(propagate_vals_across_arith_jfunc): Use self_recursion_generated_p
instead of self_recursively_generated_p, store self generation level
to such values.
(value_topo_info<valtype>::add_val): Set scc_no.
(value_topo_info<valtype>::propagate_effects): Multiply frequencies of
recursively feeding values and self generated values by appropriate
new factors.
2021-10-14 14:02:49 +02:00
-param=ipa-cp-recursive-freq-factor=
Common Joined UInteger Var(param_ipa_cp_recursive_freq_factor) Init(6) Param Optimization
2022-03-08 11:32:15 +01:00
When propagating IPA-CP effect estimates, multiply frequencies of recursive edges that bring back an unchanged value by this factor.
ipa-cp: Propagation boost for recursion generated values
Recursive call graph edges, even when they are hot and important for
the compiled program, can never have frequency bigger than one, even
when the actual time savings in the next recursion call are not
realized just once but depend on the depth of recursion. The current
IPA-CP effect propagation code did not take that into account and just
used the frequency, thus severely underestimating the effect.
This patch artificially boosts values taking part in such calls. If a
value feeds into itself through a recursive call, the frequency of the
edge is multiplied by a parameter with default value of 6, basically
assuming that the recursion will take place 6 times. This value can
of course be subject to change.
Moreover, values which do not feed into themselves but which were
generated for a self-recursive call with an arithmetic
pass-function (aka the 548.exchange "hack" which however is generally
applicable for recursive functions which count the recursion depth in
a parameter) have the edge frequency multiplied as many times as there
are generated values in the chain. In essence, we will assume they
are all useful.
This patch partially fixes the current situation when we fail to
optimize 548.exchange with PGO. In the benchmark one recursive edge
count overwhelmingly dominates all other counts in the program and so
we fail to perform the first cloning (for the nonrecursive entry call)
because it looks totally insignificant.
gcc/ChangeLog:
2021-07-16 Martin Jambor <mjambor@suse.cz>
* params.opt (ipa-cp-recursive-freq-factor): New.
* ipa-cp.c (ipcp_value): Switch to inline initialization. New members
scc_no, self_recursion_generated_level, same_scc and
self_recursion_generated_p.
(ipcp_lattice::add_value): Replaced parameter unlimited with
same_lat_gen_level, usit it determine limit of values and store it to
the value.
(ipcp_lattice<valtype>::print): Dump the new fileds.
(allocate_and_init_ipcp_value): Take same_lat_gen_level as a new
parameter and store it to the new value.
(self_recursively_generated_p): Removed.
(propagate_vals_across_arith_jfunc): Use self_recursion_generated_p
instead of self_recursively_generated_p, store self generation level
to such values.
(value_topo_info<valtype>::add_val): Set scc_no.
(value_topo_info<valtype>::propagate_effects): Multiply frequencies of
recursively feeding values and self generated values by appropriate
new factors.
2021-10-14 14:02:49 +02:00
2019-11-12 11:08:04 +01:00
-param=ipa-cp-recursion-penalty=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_ipa_cp_recursion_penalty) Init(40) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
Percentage penalty the recursive functions will receive when they are evaluated for cloning.
-param=ipa-cp-single-call-penalty=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_ipa_cp_single_call_penalty) Init(15) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
Percentage penalty functions containing a single call to another function will receive when they are evaluated for cloning.
2020-01-09 12:29:23 +01:00
-param=ipa-cp-unit-growth=
2020-01-13 14:38:59 +01:00
Common Joined UInteger Var(param_ipa_cp_unit_growth) Init(10) Param Optimization
2020-01-09 12:29:23 +01:00
How much can given compilation unit grow because of the interprocedural constant propagation (in percent).
ipa-cp: Separate and increase the large-unit parameter
A previous patch in the series has taught IPA-CP to identify the
important cloning opportunities in 548.exchange2_r as worthwhile on
their own, but the optimization is still prevented from taking place
because of the overall unit-growh limit. This patches raises that
limit so that it takes place and the benchmark runs 30% faster (on AMD
Zen2 CPU at least).
Before this patch, IPA-CP uses the following formulae to arrive at the
overall_size limit:
base = MAX(orig_size, param_large_unit_insns)
unit_growth_limit = base + base * param_ipa_cp_unit_growth / 100
since param_ipa_cp_unit_growth has default 10, param_large_unit_insns
has default value 10000.
The problem with exchange2 (at least on zen2 but I have had a quick
look on aarch64 too) is that the original estimated unit size is 10513
and so param_large_unit_insns does not apply and the default limit is
therefore 11564 which is good enough only for one of the ideal 8
clonings, we need the limit to be at least 16291.
I would like to raise param_ipa_cp_unit_growth a little bit more soon
too, but most certainly not to 55. Therefore, the large_unit must be
increased. In this patch, I decided to decouple the inlining and
ipa-cp large-unit parameters. It also makes sense because IPA-CP uses
it only at -O3 while inlining also at -O2 (IIUC). But if we agree we
can try raising param_large_unit_insns to 13-14 thousand
"instructions," perhaps it is not necessary. But then again, it may
make sense to actually increase the IPA-CP limit further.
I plan to experiment with IPA-CP tuning on a larger set of programs.
Meanwhile, mainly to address the 548.exchange2_r regression, I'm
suggesting this simple change.
gcc/ChangeLog:
2020-09-07 Martin Jambor <mjambor@suse.cz>
* params.opt (ipa-cp-large-unit-insns): New parameter.
* ipa-cp.c (get_max_overall_size): Use the new parameter.
2020-10-02 18:41:35 +02:00
-param=ipa-cp-large-unit-insns=
Common Joined UInteger Var(param_ipa_cp_large_unit_insns) Optimization Init(16000) Param
The size of translation unit that IPA-CP pass considers large.
2019-11-12 11:08:04 +01:00
-param=ipa-cp-value-list-size=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_ipa_cp_value_list_size) Init(8) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum size of a list of values associated with each parameter for interprocedural constant propagation.
2021-10-27 14:49:01 +02:00
-param=ipa-cp-profile-count-base=
Common Joined UInteger Var(param_ipa_cp_profile_count_base) Init(10) IntegerRange(0, 100) Param Optimization
When using profile feedback, use the edge at this percentage position in frequncy histogram as the bases for IPA-CP heuristics.
2020-10-29 11:51:49 +01:00
-param=ipa-jump-function-lookups=
2020-10-14 11:44:30 +02:00
Common Joined UInteger Var(param_ipa_jump_function_lookups) Init(8) Param Optimization
2020-10-15 09:24:06 +02:00
Maximum number of statements visited during jump function offset discovery.
2020-10-14 11:44:30 +02:00
2019-11-12 11:08:04 +01:00
-param=ipa-max-aa-steps=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_ipa_max_aa_steps) Init(25000) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of statements that will be visited by IPA formal parameter analysis based on alias analysis in any given function.
-param=ipa-max-agg-items=
2020-01-10 10:57:59 +01:00
Common Joined UInteger Var(param_ipa_max_agg_items) Init(16) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of aggregate content items for a parameter in jump functions and lattices.
-param=ipa-max-param-expr-ops=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_ipa_max_param_expr_ops) Init(10) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of operations in a parameter expression that can be handled by IPA analysis.
2020-10-02 18:41:35 +02:00
-param=ipa-max-loop-predicates=
Common Joined UInteger Var(param_ipa_max_loop_predicates) Init(16) Param Optimization
Maximum number of different predicates used to track properties of loops in IPA analysis.
2019-11-12 11:08:04 +01:00
-param=ipa-max-switch-predicate-bounds=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_ipa_max_switch_predicate_bounds) Init(5) Param Optimization
2020-02-18 08:54:52 +01:00
Maximal number of boundary endpoints of case ranges of switch statement used during IPA function summary generation.
2019-11-12 11:08:04 +01:00
-param=ipa-sra-max-replacements=
2020-01-10 11:01:33 +01:00
Common Joined UInteger Var(param_ipa_sra_max_replacements) Optimization Init(8) IntegerRange(0, 16) Param
2019-11-12 11:08:04 +01:00
Maximum pieces that IPA-SRA tracks per formal parameter, as a consequence, also the maximum number of replacements of a formal parameter.
-param=ipa-sra-ptr-growth-factor=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_ipa_sra_ptr_growth_factor) Init(2) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum allowed growth of number and total size of new parameters that ipa-sra replaces a pointer to an aggregate with.
-param=ira-loop-reserved-regs=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_ira_loop_reserved_regs) Init(2) Param Optimization
2019-11-12 11:08:04 +01:00
The number of registers in each class kept unused by loop invariant motion.
-param=ira-max-conflict-table-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_ira_max_conflict_table_size) Init(1000) Param Optimization
2019-11-12 11:08:04 +01:00
Max size of conflict table in MB.
-param=ira-max-loops-num=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_ira_max_loops_num) Init(100) Param Optimization
2019-11-12 11:08:04 +01:00
Max loops number for regional RA.
2021-07-06 03:53:19 +02:00
-param=ira-consider-dup-in-all-alts=
Common Joined UInteger Var(param_ira_consider_dup_in_all_alts) Init(1) IntegerRange(0, 1) Param Optimization
Control ira to consider matching constraint (duplicated operand number) heavily in all available alternatives for preferred register class. If it is set as zero, it means ira only respects the matching constraint when it's in the only available alternative with an appropriate register class. Otherwise, it means ira will check all available alternatives for preferred register class even if it has found some choice with an appropriate register class and respect the found qualified matching constraint.
2019-11-12 11:08:04 +01:00
-param=iv-always-prune-cand-set-bound=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_iv_always_prune_cand_set_bound) Init(10) Param Optimization
2019-11-12 11:08:04 +01:00
If number of candidates in the set is smaller, we always try to remove unused ivs during its optimization.
-param=iv-consider-all-candidates-bound=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_iv_consider_all_candidates_bound) Init(40) Param Optimization
2019-11-12 11:08:04 +01:00
Bound on number of candidates below that all candidates are considered in iv optimizations.
-param=iv-max-considered-uses=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_iv_max_considered_uses) Init(250) Param Optimization
2019-11-12 11:08:04 +01:00
Bound on number of iv uses in loop optimized in iv optimizations.
-param=jump-table-max-growth-ratio-for-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_jump_table_max_growth_ratio_for_size) Init(300) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum code size growth ratio when expanding into a jump table (in percent). The parameter is used when optimizing for size.
-param=jump-table-max-growth-ratio-for-speed=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_jump_table_max_growth_ratio_for_speed) Init(800) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum code size growth ratio when expanding into a jump table (in percent). The parameter is used when optimizing for speed.
-param=l1-cache-line-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_l1_cache_line_size) Init(32) Param Optimization
2019-11-12 11:08:04 +01:00
The size of L1 cache line.
c++: implement C++17 hardware interference size
The last missing piece of the C++17 standard library is the hardware
intereference size constants. Much of the delay in implementing these has
been due to uncertainty about what the right values are, and even whether
there is a single constant value that is suitable; the destructive
interference size is intended to be used in structure layout, so program
ABIs will depend on it.
In principle, both of these values should be the same as the target's L1
cache line size. When compiling for a generic target that is intended to
support a range of target CPUs with different cache line sizes, the
constructive size should probably be the minimum size, and the destructive
size the maximum, unless you are constrained by ABI compatibility with
previous code.
From discussion on gcc-patches, I've come to the conclusion that the
solution to the difficulty of choosing stable values is to give up on it,
and instead encourage only uses where ABI stability is unimportant: in
particular, uses where the ABI is shared at most between translation units
built at the same time with the same flags.
To that end, I've added a warning for any use of the constant value of
std::hardware_destructive_interference_size in a header or module export.
Appropriate uses within a project can disable the warning.
A previous iteration of this patch included an -finterference-tune flag to
make the value vary with -mtune; this iteration makes that the default
behavior, which should be appropriate for all reasonable uses of the
variable. The previous default of "stable-ish" seems to me likely to have
been more of an attractive nuisance; since we can't promise actual
stability, we should instead make proper uses more convenient.
JF Bastien's implementation proposal is summarized at
https://github.com/itanium-cxx-abi/cxx-abi/issues/74
I implement this by adding new --params for the two sizes. Targets can
override these values in targetm.target_option.override() to support a range
of values for the generic target; otherwise, both will default to the L1
cache line size.
64 bytes still seems correct for all x86.
I'm not sure why he proposed 64/64 for generic 32-bit ARM, since the Cortex
A9 has a 32-byte cache line, so I'd think 32/64 would make more sense.
He proposed 64/128 for generic AArch64, but since the A64FX now has a 256B
cache line, I've changed that to 64/256.
Other arch maintainers are invited to set ranges for their generic targets
if that seems better than using the default cache line size for both values.
With the above choice to reject stability as a goal, getting these values
"right" is now just a matter of what we want the default optimization to be,
and we can feel free to adjust them as CPUs with different cache lines
become more and less common.
gcc/ChangeLog:
* params.opt: Add destructive-interference-size and
constructive-interference-size.
* doc/invoke.texi: Document them.
* config/aarch64/aarch64.c (aarch64_override_options_internal):
Set them.
* config/arm/arm.c (arm_option_override): Set them.
* config/i386/i386-options.c (ix86_option_override_internal):
Set them.
gcc/c-family/ChangeLog:
* c.opt: Add -Winterference-size.
* c-cppbuiltin.c (cpp_atomic_builtins): Add __GCC_DESTRUCTIVE_SIZE
and __GCC_CONSTRUCTIVE_SIZE.
gcc/cp/ChangeLog:
* constexpr.c (maybe_warn_about_constant_value):
Complain about std::hardware_destructive_interference_size.
(cxx_eval_constant_expression): Call it.
* decl.c (cxx_init_decl_processing): Check
--param *-interference-size values.
libstdc++-v3/ChangeLog:
* include/std/version: Define __cpp_lib_hardware_interference_size.
* libsupc++/new: Define hardware interference size variables.
gcc/testsuite/ChangeLog:
* g++.dg/warn/Winterference.H: New file.
* g++.dg/warn/Winterference.C: New test.
* g++.target/aarch64/interference.C: New test.
* g++.target/arm/interference.C: New test.
* g++.target/i386/interference.C: New test.
2021-07-15 21:30:17 +02:00
-param=destructive-interference-size=
Common Joined UInteger Var(param_destruct_interfere_size) Init(0) Param Optimization
The minimum recommended offset between two concurrently-accessed objects to
avoid additional performance degradation due to contention introduced by the
implementation. Typically the L1 cache line size, but can be larger to
accommodate a variety of target processors with different cache line sizes.
C++17 code might use this value in structure layout, but is strongly
discouraged from doing so in public ABIs.
-param=constructive-interference-size=
Common Joined UInteger Var(param_construct_interfere_size) Init(0) Param Optimization
The maximum recommended size of contiguous memory occupied by two objects
accessed with temporal locality by concurrent threads. Typically the L1 cache
line size, but can be smaller to accommodate a variety of target processors with
different cache line sizes.
2019-11-12 11:08:04 +01:00
-param=l1-cache-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_l1_cache_size) Init(64) Param Optimization
2019-11-12 11:08:04 +01:00
The size of L1 cache.
-param=l2-cache-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_l2_cache_size) Init(512) Param Optimization
2019-11-12 11:08:04 +01:00
The size of L2 cache.
-param=large-function-growth=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_large_function_growth) Optimization Init(100) Param
2019-11-12 11:08:04 +01:00
Maximal growth due to inlining of large function (in percent).
-param=large-function-insns=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_large_function_insns) Optimization Init(2700) Param
2019-11-12 11:08:04 +01:00
The size of function body to be considered large.
-param=large-stack-frame=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_large_stack_frame) Init(256) Optimization Param
2019-11-12 11:08:04 +01:00
The size of stack frame to be considered large.
-param=large-stack-frame-growth=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_stack_frame_growth) Optimization Init(1000) Param
2019-11-12 11:08:04 +01:00
Maximal stack frame growth due to inlining (in percent).
-param=large-unit-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_large_unit_insns) Optimization Init(10000) Param
2019-11-12 11:08:04 +01:00
The size of translation unit to be considered large.
2020-12-01 20:39:38 +01:00
-param=lazy-modules=
C++ Joined UInteger Var(param_lazy_modules) Init(32768) Param
Maximum number of concurrently open C++ module files when lazy loading.
2019-11-12 11:08:04 +01:00
-param=lim-expensive=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_lim_expensive) Init(20) Param Optimization
2019-11-12 11:08:04 +01:00
The minimum cost of an expensive expression in the loop invariant motion.
-param=logical-op-non-short-circuit=
2021-10-25 16:54:22 +02:00
Common Joined UInteger Var(param_logical_op_non_short_circuit) Init(-1) IntegerRange(0, 1) Param
2019-11-12 11:08:04 +01:00
True if a non-short-circuit operation is optimal.
-param=loop-block-tile-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_loop_block_tile_size) Init(51) Param Optimization
2019-11-12 11:08:04 +01:00
Size of tiles for loop blocking.
-param=loop-interchange-max-num-stmts=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_loop_interchange_max_num_stmts) Init(64) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of stmts in loop nest for loop interchange.
-param=loop-interchange-stride-ratio=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_loop_interchange_stride_ratio) Init(2) Param Optimization
2019-11-12 11:08:04 +01:00
The minimum stride ratio for loop interchange to be profitable.
-param=loop-invariant-max-bbs-in-loop=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_loop_invariant_max_bbs_in_loop) Init(10000) Param Optimization
2019-11-12 11:08:04 +01:00
Max basic blocks number in loop for loop invariant motion.
-param=loop-max-datarefs-for-datadeps=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_loop_max_datarefs_for_datadeps) Init(1000) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of datarefs in loop for building loop data dependencies.
-param=loop-versioning-max-inner-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_loop_versioning_max_inner_insns) Init(200) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of instructions in an inner loop that is being considered for versioning.
-param=loop-versioning-max-outer-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_loop_versioning_max_outer_insns) Init(100) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of instructions in an outer loop that is being considered for versioning, on top of the instructions in inner loops.
-param=lra-inheritance-ebb-probability-cutoff=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_lra_inheritance_ebb_probability_cutoff) Init(40) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
Minimal fall-through edge probability in percentage used to add BB to inheritance EBB in LRA.
-param=lra-max-considered-reload-pseudos=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_lra_max_considered_reload_pseudos) Init(500) Param Optimization
2019-11-12 11:08:04 +01:00
The max number of reload pseudos which are considered during spilling a non-reload pseudo.
-param=lto-max-partition=
Common Joined UInteger Var(param_max_partition_size) Init(1000000) Param
Maximal size of a partition for LTO (in estimated instructions).
-param=lto-max-streaming-parallelism=
Common Joined UInteger Var(param_max_lto_streaming_parallelism) Init(32) IntegerRange(1, 65536) Param
maximal number of LTO partitions streamed in parallel.
-param=lto-min-partition=
Common Joined UInteger Var(param_min_partition_size) Init(10000) Param
Minimal size of a partition for LTO (in estimated instructions).
-param=lto-partitions=
Common Joined UInteger Var(param_lto_partitions) Init(128) IntegerRange(1, 65536) Param
Number of partitions the program should be split to.
-param=max-average-unrolled-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_average_unrolled_insns) Init(80) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of instructions to consider to unroll in a loop on average.
-param=max-combine-insns=
2020-01-06 10:13:15 +01:00
Common Joined UInteger Var(param_max_combine_insns) Init(4) IntegerRange(2, 4) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of insns combine tries to combine.
-param=max-completely-peel-loop-nest-depth=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_unroll_iterations) Init(8) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum depth of a loop nest we completely peel.
-param=max-completely-peel-times=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_completely_peel_times) Init(16) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of peelings of a single loop that is peeled completely.
-param=max-completely-peeled-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_completely_peeled_insns) Init(200) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of insns of a completely peeled loop.
-param=max-crossjump-edges=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_crossjump_edges) Init(100) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of incoming edges to consider for crossjumping.
-param=max-cse-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_cse_insns) Init(1000) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum instructions CSE process before flushing.
-param=max-cse-path-length=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_cse_path_length) Init(10) IntegerRange(1, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum length of path considered in cse.
-param=max-cselib-memory-locations=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_cselib_memory_locations) Init(500) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum memory locations recorded by cselib.
-param=max-debug-marker-count=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_debug_marker_count) Init(100000) Param Optimization
2019-11-12 11:08:04 +01:00
Max. count of debug markers to expand or inline.
-param=max-delay-slot-insn-search=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_delay_slot_insn_search) Init(100) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of instructions to consider to fill a delay slot.
-param=max-delay-slot-live-search=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_delay_slot_live_search) Init(333) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of instructions to consider to find accurate live register information.
-param=max-dse-active-local-stores=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_dse_active_local_stores) Init(5000) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of active local stores in RTL dead store elimination.
-param=max-early-inliner-iterations=
2020-01-09 12:29:23 +01:00
Common Joined UInteger Var(param_early_inliner_max_iterations) Init(1) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of nested indirect inlining performed by early inliner.
-param=max-fields-for-field-sensitive=
Common Joined UInteger Var(param_max_fields_for_field_sensitive) Param
Maximum number of fields in a structure before pointer analysis treats the structure as a single variable.
-param=max-fsm-thread-path-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_fsm_thread_path_insns) Init(100) IntegerRange(1, 999999) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of instructions to copy when duplicating blocks on a finite state automaton jump thread path.
-param=max-gcse-insertion-ratio=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_gcse_insertion_ratio) Init(20) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum ratio of insertions to deletions of expressions in GCSE.
-param=max-gcse-memory=
2021-01-29 13:58:28 +01:00
Common Joined UInteger Var(param_max_gcse_memory) Init(131072) Param Optimization
The maximum amount of memory to be allocated by GCSE, in kilobytes.
2019-11-12 11:08:04 +01:00
-param=max-goto-duplication-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_goto_duplication_insns) Init(8) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of insns to duplicate when unfactoring computed gotos.
-param=max-grow-copy-bb-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_grow_copy_bb_insns) Init(8) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum expansion factor when copying basic blocks.
-param=max-hoist-depth=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_hoist_depth) Init(30) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum depth of search in the dominator tree for expressions to hoist.
2021-12-09 21:02:17 +01:00
-param=max-inline-functions-called-once-loop-depth=
Common Joined UInteger Var(param_inline_functions_called_once_loop_depth) Init(6) Optimization Param
2021-12-10 13:31:21 +01:00
Maximum loop depth of a call which is considered for inlining functions called once.
2021-12-09 21:02:17 +01:00
-param=max-inline-functions-called-once-insns=
Common Joined UInteger Var(param_inline_functions_called_once_insns) Init(4000) Optimization Param
2022-03-08 18:14:33 +01:00
Maximum combined size of caller and callee which is inlined if callee is called once.
2021-12-09 21:02:17 +01:00
2019-11-12 11:08:04 +01:00
-param=max-inline-insns-auto=
2019-11-14 13:08:57 +01:00
Common Joined UInteger Var(param_max_inline_insns_auto) Init(15) Optimization Param
2019-11-12 11:08:04 +01:00
The maximum number of instructions when automatically inlining.
-param=max-inline-insns-recursive=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_max_inline_insns_recursive) Optimization Init(450) Param
2019-11-12 11:08:04 +01:00
The maximum number of instructions inline function can grow to via recursive inlining.
-param=max-inline-insns-recursive-auto=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_max_inline_insns_recursive_auto) Optimization Init(450) Param
2019-11-12 11:08:04 +01:00
The maximum number of instructions non-inline function can grow to via recursive inlining.
-param=max-inline-insns-single=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_max_inline_insns_single) Optimization Init(70) Param
2019-11-12 11:08:04 +01:00
The maximum number of instructions in a single function eligible for inlining.
-param=max-inline-insns-size=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_max_inline_insns_size) Optimization Param
2019-11-12 11:08:04 +01:00
The maximum number of instructions when inlining for size.
-param=max-inline-insns-small=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_max_inline_insns_small) Optimization Param
2019-11-12 11:08:04 +01:00
The maximum number of instructions when automatically inlining small functions.
-param=max-inline-recursive-depth=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_max_inline_recursive_depth) Optimization Init(8) Param
2019-11-12 11:08:04 +01:00
The maximum depth of recursive inlining for inline functions.
-param=max-inline-recursive-depth-auto=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_max_inline_recursive_depth_auto) Optimization Init(8) Param
2019-11-12 11:08:04 +01:00
The maximum depth of recursive inlining for non-inline functions.
-param=max-isl-operations=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_isl_operations) Init(350000) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of isl operations, 0 means unlimited.
-param=max-iterations-computation-cost=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_iterations_computation_cost) Init(10) Param Optimization
2019-11-12 11:08:04 +01:00
Bound on the cost of an expression to compute the number of iterations.
-param=max-iterations-to-track=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_iterations_to_track) Init(1000) Param Optimization
2019-11-12 11:08:04 +01:00
Bound on the number of iterations the brute force # of iterations analysis algorithm evaluates.
-param=max-jump-thread-duplication-stmts=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_jump_thread_duplication_stmts) Init(15) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of statements allowed in a block that needs to be duplicated when threading jumps.
2022-08-08 12:20:04 +02:00
-param=max-jump-thread-paths=
Common Joined UInteger Var(param_max_jump_thread_paths) Init(64) IntegerRange(1, 65536) Param Optimization
Search space limit for the backwards jump threader.
2019-11-12 11:08:04 +01:00
-param=max-last-value-rtl=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_last_value_rtl) Init(10000) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of RTL nodes that can be recorded as combiner's last value.
-param=max-loop-header-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_loop_header_insns) Init(20) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of insns in loop header duplicated by the copy loop headers pass.
-param=max-modulo-backtrack-attempts=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_modulo_backtrack_attempts) Init(40) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of backtrack attempts the scheduler should make when modulo scheduling a loop.
2022-03-18 18:58:06 +01:00
-param=min-pagesize=
Common Joined UInteger Var(param_min_pagesize) Init(4096) Param Optimization
Minimum page size for warning purposes.
2019-11-12 11:08:04 +01:00
-param=max-partial-antic-length=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_partial_antic_length) Init(100) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum length of partial antic set when performing tree pre optimization.
-param=max-peel-branches=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_peel_branches) Init(32) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of branches on the path through the peeled sequence.
-param=max-peel-times=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_peel_times) Init(16) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of peelings of a single loop.
-param=max-peeled-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_peeled_insns) Init(100) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of insns of a peeled loop.
-param=max-pending-list-length=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_pending_list_length) Init(32) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum length of scheduling's pending operations list.
-param=max-pipeline-region-blocks=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_pipeline_region_blocks) Init(15) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of blocks in a region to be considered for interblock scheduling.
-param=max-pipeline-region-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_pipeline_region_insns) Init(200) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of insns in a region to be considered for interblock scheduling.
-param=max-pow-sqrt-depth=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_pow_sqrt_depth) Init(5) IntegerRange(1, 32) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum depth of sqrt chains to use when synthesizing exponentiation by a real constant.
-param=max-predicted-iterations=
2020-01-17 00:50:22 +01:00
Common Joined UInteger Var(param_max_predicted_iterations) Init(100) IntegerRange(0, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of loop iterations we predict statically.
-param=max-reload-search-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_reload_search_insns) Init(100) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of instructions to search backward when looking for equivalent reload.
-param=max-rtl-if-conversion-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_rtl_if_conversion_insns) Init(10) IntegerRange(0, 99) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of insns in a basic block to consider for RTL if-conversion.
-param=max-rtl-if-conversion-predictable-cost=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_rtl_if_conversion_predictable_cost) Init(20) IntegerRange(0, 200) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum permissible cost for the sequence that would be generated by the RTL if-conversion pass for a branch that is considered predictable.
-param=max-rtl-if-conversion-unpredictable-cost=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_rtl_if_conversion_unpredictable_cost) Init(40) IntegerRange(0, 200) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum permissible cost for the sequence that would be generated by the RTL if-conversion pass for a branch that is considered unpredictable.
-param=max-sched-extend-regions-iters=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_sched_extend_regions_iters) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of iterations through CFG to extend regions.
-param=max-sched-insn-conflict-delay=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_sched_insn_conflict_delay) Init(3) IntegerRange(1, 10) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum conflict delay for an insn to be considered for speculative motion.
-param=max-sched-ready-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_sched_ready_insns) Init(100) IntegerRange(1, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of instructions ready to be issued to be considered by the scheduler during the first scheduling pass.
-param=max-sched-region-blocks=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_sched_region_blocks) Init(10) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of blocks in a region to be considered for interblock scheduling.
-param=max-sched-region-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_sched_region_insns) Init(100) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of insns in a region to be considered for interblock scheduling.
-param=max-slsr-cand-scan=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_slsr_candidate_scan) Init(50) IntegerRange(1, 999999) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum length of candidate scans for straight-line strength reduction.
-param=max-speculative-devirt-maydefs=
2020-01-09 13:17:12 +01:00
Common Joined UInteger Var(param_max_speculative_devirt_maydefs) Init(50) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of may-defs visited when devirtualizing speculatively.
-param=max-ssa-name-query-depth=
Common Joined UInteger Var(param_max_ssa_name_query_depth) Init(3) IntegerRange(1, 10) Param
Maximum recursion depth allowed when querying a property of an SSA name.
-param=max-stores-to-merge=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_stores_to_merge) Init(64) IntegerRange(2, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of constant stores to merge in the store merging pass.
-param=max-stores-to-sink=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_stores_to_sink) Init(2) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of conditional store pairs that can be sunk.
2021-02-11 11:13:47 +01:00
-param=max-store-chains-to-track=
2021-02-15 15:09:04 +01:00
Common Joined UInteger Var(param_max_store_chains_to_track) Init(64) IntegerRange(1, 65536) Param
2021-02-11 11:13:47 +01:00
Maximum number of store chains to track at the same time in the store merging pass.
-param=max-stores-to-track=
2021-02-15 15:09:04 +01:00
Common Joined UInteger Var(param_max_stores_to_track) Init(1024) IntegerRange(2, 1048576) Param
2021-02-11 11:13:47 +01:00
Maximum number of store chains to track at the same time in the store merging pass.
2019-11-12 11:08:04 +01:00
-param=max-tail-merge-comparisons=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_tail_merge_comparisons) Init(10) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum amount of similar bbs to compare a bb with.
-param=max-tail-merge-iterations=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_tail_merge_iterations) Init(2) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum amount of iterations of the pass over a function.
-param=max-tracked-strlens=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_tracked_strlens) Init(10000) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of strings for which strlen optimization pass will track string lengths.
-param=max-tree-if-conversion-phi-args=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_tree_if_conversion_phi_args) Init(4) IntegerRange(2, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of arguments in a PHI supported by TREE if-conversion unless the loop is marked with simd pragma.
-param=max-unroll-times=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_unroll_times) Init(8) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of unrollings of a single loop.
-param=max-unrolled-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_unrolled_insns) Init(200) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of instructions to consider to unroll in a loop.
-param=max-unswitch-insns=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_unswitch_insns) Init(50) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of insns of an unswitched loop.
-param=max-variable-expansions-in-unroller=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_variable_expansions) Init(1) Param Optimization
2019-11-12 11:08:04 +01:00
If -fvariable-expansion-in-unroller is used, the maximum number of times that an individual variable will be expanded during loop unrolling.
-param=max-vartrack-expr-depth=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_vartrack_expr_depth) Init(12) Param Optimization
2019-11-12 11:08:04 +01:00
Max. recursion depth for expanding var tracking expressions.
-param=max-vartrack-reverse-op-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_vartrack_reverse_op_size) Init(50) Param Optimization
2019-11-12 11:08:04 +01:00
Max. size of loc list for which reverse ops should be added.
-param=max-vartrack-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_vartrack_size) Init(50000000) Param Optimization
2019-11-12 11:08:04 +01:00
Max. size of var tracking hash tables.
2020-03-09 13:38:23 +01:00
-param=max-find-base-term-values=
2020-04-02 14:34:42 +02:00
Common Joined UInteger Var(param_max_find_base_term_values) Init(200) Param Optimization
2020-03-09 13:38:23 +01:00
Maximum number of VALUEs handled during a single find_base_term call.
2019-11-12 11:08:04 +01:00
-param=max-vrp-switch-assertions=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_max_vrp_switch_assertions) Init(10) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of assertions to add along the default edge of a switch statement during VRP.
-param=min-crossjump-insns=
2020-01-07 10:10:37 +01:00
Common Joined UInteger Var(param_min_crossjump_insns) Init(5) IntegerRange(1, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
The minimum number of matching instructions to consider for crossjumping.
-param=min-inline-recursive-probability=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_min_inline_recursive_probability) Init(10) Optimization Param
2019-11-12 11:08:04 +01:00
Inline recursively only when the probability of call being executed exceeds the parameter.
-param=min-insn-to-prefetch-ratio=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_min_insn_to_prefetch_ratio) Init(9) Param Optimization
2019-11-12 11:08:04 +01:00
Min. ratio of insns to prefetches to enable prefetching for a loop with an unknown trip count.
-param=min-loop-cond-split-prob=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_min_loop_cond_split_prob) Init(30) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
The minimum threshold for probability of semi-invariant condition statement to trigger loop split.
-param=min-nondebug-insn-uid=
Common Joined UInteger Var(param_min_nondebug_insn_uid) Param
The minimum UID to be used for a nondebug insn.
-param=min-size-for-stack-sharing=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_min_size_for_stack_sharing) Init(32) Param Optimization
2019-11-12 11:08:04 +01:00
The minimum size of variables taking part in stack slot sharing when not optimizing.
-param=min-spec-prob=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_min_spec_prob) Init(40) Param Optimization
2019-11-12 11:08:04 +01:00
The minimum probability of reaching a source block for interblock speculative scheduling.
-param=min-vect-loop-bound=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_min_vect_loop_bound) Param Optimization
2019-11-12 11:08:04 +01:00
If -ftree-vectorize is used, the minimal loop bound of a loop to be considered for vectorization.
2021-04-19 10:24:49 +02:00
-param=openacc-kernels=
Common Joined Enum(openacc_kernels) Var(param_openacc_kernels) Init(OPENACC_KERNELS_PARLOOPS) Param
--param=openacc-kernels=[decompose|parloops] Specify mode of OpenACC 'kernels' constructs handling.
Enum
Name(openacc_kernels) Type(enum openacc_kernels)
EnumValue
Enum(openacc_kernels) String(decompose) Value(OPENACC_KERNELS_DECOMPOSE)
EnumValue
Enum(openacc_kernels) String(parloops) Value(OPENACC_KERNELS_PARLOOPS)
2021-05-20 16:11:37 +02:00
-param=openacc-privatization=
Common Joined Enum(openacc_privatization) Var(param_openacc_privatization) Init(OPENACC_PRIVATIZATION_QUIET) Param
--param=openacc-privatization=[quiet|noisy] Specify mode of OpenACC privatization diagnostics.
Enum
Name(openacc_privatization) Type(enum openacc_privatization)
EnumValue
Enum(openacc_privatization) String(quiet) Value(OPENACC_PRIVATIZATION_QUIET)
EnumValue
Enum(openacc_privatization) String(noisy) Value(OPENACC_PRIVATIZATION_NOISY)
2019-11-12 11:08:04 +01:00
-param=parloops-chunk-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_parloops_chunk_size) Param Optimization
2019-11-12 11:08:04 +01:00
Chunk size of omp schedule for loops parallelized by parloops.
-param=parloops-min-per-thread=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_parloops_min_per_thread) Init(100) IntegerRange(2, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
Minimum number of iterations per thread of an innermost parallelized loop.
-param=parloops-schedule=
2020-01-07 10:21:01 +01:00
Common Joined Var(param_parloops_schedule) Enum(parloops_schedule_type) Param Optimization
2019-11-12 11:08:04 +01:00
--param=parloops-schedule=[static|dynamic|guided|auto|runtime] Schedule type of omp schedule for loops parallelized by parloops.
Enum
Name(parloops_schedule_type) Type(int)
EnumValue
Enum(parloops_schedule_type) String(static) Value(PARLOOPS_SCHEDULE_STATIC)
EnumValue
Enum(parloops_schedule_type) String(dynamic) Value(PARLOOPS_SCHEDULE_DYNAMIC)
EnumValue
Enum(parloops_schedule_type) String(guided) Value(PARLOOPS_SCHEDULE_GUIDED)
EnumValue
Enum(parloops_schedule_type) String(auto) Value(PARLOOPS_SCHEDULE_AUTO)
EnumValue
Enum(parloops_schedule_type) String(runtime) Value(PARLOOPS_SCHEDULE_RUNTIME)
-param=partial-inlining-entry-probability=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_partial_inlining_entry_probability) Init(70) Optimization IntegerRange(0, 100) Param
2019-11-12 11:08:04 +01:00
Maximum probability of the entry BB of split region (in percent relative to entry BB of the function) to make partial inlining happen.
-param=predictable-branch-outcome=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_predictable_branch_outcome) Init(2) IntegerRange(0, 50) Param Optimization
2019-11-12 11:08:04 +01:00
Maximal estimated outcome of branch considered predictable.
-param=prefetch-dynamic-strides=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_prefetch_dynamic_strides) Init(1) IntegerRange(0, 1) Param Optimization
2019-11-12 11:08:04 +01:00
Whether software prefetch hints should be issued for non-constant strides.
-param=prefetch-latency=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_prefetch_latency) Init(200) Param Optimization
2019-11-12 11:08:04 +01:00
The number of insns executed before prefetch is completed.
-param=prefetch-min-insn-to-mem-ratio=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_prefetch_min_insn_to_mem_ratio) Init(3) Param Optimization
2019-11-12 11:08:04 +01:00
Min. ratio of insns to mem ops to enable prefetching in a loop.
-param=prefetch-minimum-stride=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_prefetch_minimum_stride) Init(-1) Param Optimization
2019-11-12 11:08:04 +01:00
The minimum constant stride beyond which we should use prefetch hints for.
-param=profile-func-internal-id=
Common Joined UInteger Var(param_profile_func_internal_id) IntegerRange(0, 1) Param
Use internal function id in profile lookup.
2021-10-19 20:09:51 +02:00
-param=ranger-debug=
Common Joined Var(param_ranger_debug) Enum(ranger_debug) Init(RANGER_DEBUG_NONE) Param Optimization
--param=ranger-debug=[none|trace|gori|cache|tracegori|all] Specifies the output mode for debugging ranger.
Enum
Name(ranger_debug) Type(enum ranger_debug) UnknownError(unknown ranger debug mode %qs)
EnumValue
Enum(ranger_debug) String(none) Value(RANGER_DEBUG_NONE)
EnumValue
Enum(ranger_debug) String(trace) Value(RANGER_DEBUG_TRACE)
EnumValue
2021-10-21 16:58:16 +02:00
Enum(ranger_debug) String(cache) Value(RANGER_DEBUG_TRACE_CACHE)
2021-10-19 20:09:51 +02:00
EnumValue
Enum(ranger_debug) String(gori) Value(RANGER_DEBUG_GORI)
EnumValue
Enum(ranger_debug) String(tracegori) Value(RANGER_DEBUG_TRACE_GORI)
EnumValue
Enum(ranger_debug) String(all) Value(RANGER_DEBUG_ALL)
-param=ranger-logical-depth=
Common Joined UInteger Var(param_ranger_logical_depth) Init(6) IntegerRange(1, 999) Param Optimization
Maximum depth of logical expression evaluation ranger will look through when
evaluating outgoing edge ranges.
2022-01-17 18:03:18 +01:00
-param=relation-block-limit=
Common Joined UInteger Var(param_relation_block_limit) Init(200) IntegerRange(0, 9999) Param Optimization
Maximum number of relations the oracle will register in a basic block.
2019-11-12 11:08:04 +01:00
-param=rpo-vn-max-loop-depth=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_rpo_vn_max_loop_depth) Init(7) IntegerRange(2, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum depth of a loop nest to fully value-number optimistically.
-param=sccvn-max-alias-queries-per-access=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sccvn_max_alias_queries_per_access) Init(1000) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of disambiguations to perform per memory access.
-param=scev-max-expr-complexity=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_scev_max_expr_complexity) Init(10) Param Optimization
2019-11-12 11:08:04 +01:00
Bound on the complexity of the expressions in the scalar evolutions analyzer.
-param=scev-max-expr-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_scev_max_expr_size) Init(100) Param Optimization
2019-11-12 11:08:04 +01:00
Bound on size of expressions used in the scalar evolutions analyzer.
-param=sched-autopref-queue-depth=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sched_autopref_queue_depth) Init(-1) Param Optimization
2019-11-12 11:08:04 +01:00
Hardware autoprefetcher scheduler model control flag. Number of lookahead cycles the model looks into, at '0' only enable instruction sorting heuristic. Disabled by default.
-param=sched-mem-true-dep-cost=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sched_mem_true_dep_cost) Init(1) Param Optimization
2019-11-12 11:08:04 +01:00
Minimal distance between possibly conflicting store and load.
-param=sched-pressure-algorithm=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sched_pressure_algorithm) Init(1) IntegerRange(1, 2) Param Optimization
2019-11-12 11:08:04 +01:00
Which -fsched-pressure algorithm to apply.
-param=sched-spec-prob-cutoff=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sched_spec_prob_cutoff) Init(40) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
The minimal probability of speculation success (in percents), so that speculative insn will be scheduled.
-param=sched-state-edge-prob-cutoff=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sched_state_edge_prob_cutoff) Init(10) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
The minimum probability an edge must have for the scheduler to save its state across it.
-param=selsched-insns-to-rename=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_selsched_insns_to_rename) Init(2) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of instructions in the ready list that are considered eligible for renaming.
-param=selsched-max-lookahead=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_selsched_max_lookahead) Init(50) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum size of the lookahead window of selective scheduling.
-param=selsched-max-sched-times=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_selsched_max_sched_times) Init(2) IntegerRange(1, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of times that an insn could be scheduled.
-param=simultaneous-prefetches=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_simultaneous_prefetches) Init(3) Param Optimization
2019-11-12 11:08:04 +01:00
The number of prefetches that can run at the same time.
-param=sink-frequency-threshold=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sink_frequency_threshold) Init(75) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
Target block's relative execution frequency (as a percentage) required to sink a statement.
-param=sms-dfa-history=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sms_dfa_history) IntegerRange(0, 16) Param Optimization
2019-11-12 11:08:04 +01:00
The number of cycles the swing modulo scheduler considers when checking conflicts using DFA.
-param=sms-loop-average-count-threshold=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sms_loop_average_count_threshold) Param Optimization
2019-11-12 11:08:04 +01:00
A threshold on the average loop count considered by the swing modulo scheduler.
-param=sms-max-ii-factor=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sms_max_ii_factor) Init(2) IntegerRange(1, 16) Param Optimization
2019-11-12 11:08:04 +01:00
A factor for tuning the upper bound that swing modulo scheduler uses for scheduling a loop.
-param=sms-min-sc=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sms_min_sc) Init(2) IntegerRange(1, 2) Param Optimization
2019-11-12 11:08:04 +01:00
The minimum value of stage count that swing modulo scheduler will generate.
-param=sra-max-scalarization-size-Osize=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sra_max_scalarization_size_size) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum size, in storage units, of an aggregate which should be considered for scalarization when compiling for size.
-param=sra-max-scalarization-size-Ospeed=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_sra_max_scalarization_size_speed) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum size, in storage units, of an aggregate which should be considered for scalarization when compiling for speed.
2020-03-21 00:21:02 +01:00
-param=sra-max-propagations=
Common Joined UInteger Var(param_sra_max_propagations) Param Optimization Init(32)
Maximum number of artificial accesses to enable forward propagation that Scalar Replacement of Aggregates will keep for one local variable.
2019-11-12 11:08:04 +01:00
-param=ssa-name-def-chain-limit=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_ssa_name_def_chain_limit) Init(512) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum number of SSA_NAME assignments to follow in determining a value.
-param=ssp-buffer-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_ssp_buffer_size) Init(8) IntegerRange(1, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
The lower bound for a buffer to be considered for stack smashing protection.
-param=stack-clash-protection-guard-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_stack_clash_protection_guard_size) Init(12) IntegerRange(12, 30) Param Optimization
2019-11-12 11:08:04 +01:00
Size of the stack guard expressed as a power of two in bytes.
-param=stack-clash-protection-probe-interval=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_stack_clash_protection_probe_interval) Init(12) IntegerRange(10, 16) Param Optimization
2019-11-12 11:08:04 +01:00
Interval in which to probe the stack expressed as a power of two in bytes.
-param=store-merging-allow-unaligned=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_store_merging_allow_unaligned) Init(1) IntegerRange(0, 1) Param Optimization
2019-11-12 11:08:04 +01:00
Allow the store merging pass to introduce unaligned stores if it is legal to do so.
-param=store-merging-max-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_store_merging_max_size) Init(65536) IntegerRange(1, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum size of a single store merging region in bytes.
-param=switch-conversion-max-branch-ratio=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_switch_conversion_branch_ratio) Init(8) IntegerRange(1, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
The maximum ratio between array size and switch branches for a switch conversion to take place.
2020-09-20 07:25:16 +02:00
-param=modref-max-bases=
2020-11-16 13:13:40 +01:00
Common Joined UInteger Var(param_modref_max_bases) Init(32) Param Optimization
2020-09-20 07:25:16 +02:00
Maximum number of bases stored in each modref tree.
-param=modref-max-refs=
2020-11-16 13:13:40 +01:00
Common Joined UInteger Var(param_modref_max_refs) Init(16) Param Optimization
Add access through parameter derference tracking to modref
re-add tracking of accesses which was unfinished in David's patch.
At the moment I only implemented tracking of the fact that access is based on
derefernece of the parameter (so we track THIS pointers).
Patch does not implement IPA propagation since it needs bit more work which
I will post shortly: ipa-fnsummary needs to track when parameter points to
local memory, summaries needs to be merged when function is inlined (because
jump functions are) and propagation needs to be turned into iterative dataflow
on SCC components.
Patch also adds documentation of -fipa-modref and params that was left uncommited
in my branch :(.
Even without this change it does lead to nice increase of disambiguations
for cc1plus build.
Alias oracle query stats:
refs_may_alias_p: 62758323 disambiguations, 72935683 queries
ref_maybe_used_by_call_p: 139511 disambiguations, 63654045 queries
call_may_clobber_ref_p: 23502 disambiguations, 29242 queries
nonoverlapping_component_refs_p: 0 disambiguations, 37654 queries
nonoverlapping_refs_since_match_p: 19417 disambiguations, 55555 must overlaps, 75721 queries
aliasing_component_refs_p: 54665 disambiguations, 752449 queries
TBAA oracle: 21917926 disambiguations 53054678 queries
15763411 are in alias set 0
10162238 queries asked about the same object
124 queries asked about the same alias set
0 access volatile
3681593 are dependent in the DAG
1529386 are aritificially in conflict with void *
Modref stats:
modref use: 8311 disambiguations, 32527 queries
modref clobber: 742126 disambiguations, 1036986 queries
1987054 tbaa queries (1.916182 per modref query)
125479 base compares (0.121004 per modref query)
PTA query stats:
pt_solution_includes: 968314 disambiguations, 13609584 queries
pt_solutions_intersect: 1019136 disambiguations, 13147139 queries
So compared to
https://gcc.gnu.org/pipermail/gcc-patches/2020-September/554605.html
we get 41% more use disambiguations (with similar number of queries) and 8% more
clobber disambiguations.
For tramp3d:
Alias oracle query stats:
refs_may_alias_p: 2052256 disambiguations, 2312703 queries
ref_maybe_used_by_call_p: 7122 disambiguations, 2089118 queries
call_may_clobber_ref_p: 234 disambiguations, 234 queries
nonoverlapping_component_refs_p: 0 disambiguations, 4299 queries
nonoverlapping_refs_since_match_p: 329 disambiguations, 10200 must overlaps, 10616 queries
aliasing_component_refs_p: 857 disambiguations, 34555 queries
TBAA oracle: 885546 disambiguations 1677080 queries
132105 are in alias set 0
469030 queries asked about the same object
0 queries asked about the same alias set
0 access volatile
190084 are dependent in the DAG
315 are aritificially in conflict with void *
Modref stats:
modref use: 426 disambiguations, 1881 queries
modref clobber: 10042 disambiguations, 16202 queries
19405 tbaa queries (1.197692 per modref query)
2775 base compares (0.171275 per modref query)
PTA query stats:
pt_solution_includes: 313908 disambiguations, 526183 queries
pt_solutions_intersect: 130510 disambiguations, 416084 queries
Here uses decrease by 4 disambiguations and clobber improve by 3.5%. I think
the difference is caused by fact that gcc has much more alias set 0 accesses
originating from gimple and tree unions as I mentioned in original mail.
After pushing out the IPA propagation I will re-add code to track offsets and
sizes that further improve disambiguation. On tramp3d it enables a lot of DSE
for structure fields not acessed by uninlined function.
gcc/
* doc/invoke.texi: Document -fipa-modref, ipa-modref-max-bases,
ipa-modref-max-refs, ipa-modref-max-accesses, ipa-modref-max-tests.
* ipa-modref-tree.c (test_insert_search_collapse): Update.
(test_merge): Update.
(gt_ggc_mx): New function.
* ipa-modref-tree.h (struct modref_access_node): New structure.
(struct modref_ref_node): Add every_access and accesses array.
(modref_ref_node::modref_ref_node): Update ctor.
(modref_ref_node::search): New member function.
(modref_ref_node::collapse): New member function.
(modref_ref_node::insert_access): New member function.
(modref_base_node::insert_ref): Do not collapse base if ref is 0.
(modref_base_node::collapse): Copllapse also refs.
(modref_tree): Add accesses.
(modref_tree::modref_tree): Initialize max_accesses.
(modref_tree::insert): Add access parameter.
(modref_tree::cleanup): New member function.
(modref_tree::merge): Add parm_map; merge accesses.
(modref_tree::copy_from): New member function.
(modref_tree::create_ggc): Add max_accesses.
* ipa-modref.c (dump_access): New function.
(dump_records): Dump accesses.
(dump_lto_records): Dump accesses.
(get_access): New function.
(record_access): Record access.
(record_access_lto): Record access.
(analyze_call): Compute parm_map.
(analyze_function): Update construction of modref records.
(modref_summaries::duplicate): Likewise; use copy_from.
(write_modref_records): Stream accesses.
(read_modref_records): Sream accesses.
(pass_ipa_modref::execute): Update call of merge.
* params.opt (-param=modref-max-accesses): New.
* tree-ssa-alias.c (alias_stats): Add modref_baseptr_tests.
(dump_alias_stats): Update.
(base_may_alias_with_dereference_p): New function.
(modref_may_conflict): Check accesses.
(ref_maybe_used_by_call_p_1): Update call to modref_may_conflict.
(call_may_clobber_ref_p_1): Update call to modref_may_conflict.
2020-09-24 15:09:17 +02:00
Maximum number of references stored in each modref base.
-param=modref-max-accesses=
2020-11-16 13:13:40 +01:00
Common Joined UInteger Var(param_modref_max_accesses) Init(16) Param Optimization
2021-02-19 17:29:23 +01:00
Maximum number of accesses stored in each modref reference.
2020-09-20 07:25:16 +02:00
-param=modref-max-tests=
2020-11-16 13:13:40 +01:00
Common Joined UInteger Var(param_modref_max_tests) Init(64) Param Optimization
2020-09-22 10:45:06 +02:00
Maximum number of tests performed by modref query.
2020-09-20 07:25:16 +02:00
2020-11-14 13:52:36 +01:00
-param=modref-max-depth=
2021-11-07 22:58:19 +01:00
Common Joined UInteger Var(param_modref_max_depth) Init(256) IntegerRange(1, 65536) Param Optimization
2020-11-16 13:23:16 +01:00
Maximum depth of DFS walk used by modref escape analysis.
2020-11-14 13:52:36 +01:00
2020-11-16 19:30:45 +01:00
-param=modref-max-escape-points=
Common Joined UInteger Var(param_modref_max_escape_points) Init(256) Param Optimization
2020-11-16 20:43:31 +01:00
Maximum number of escape points tracked by modref per SSA-name.
2020-11-16 19:30:45 +01:00
2021-08-25 21:43:07 +02:00
-param=modref-max-adjustments=
2021-11-07 22:58:19 +01:00
Common Joined UInteger Var(param_modref_max_adjustments) Init(8) IntegerRange(0, 254) Param Optimization
2021-08-26 16:01:53 +02:00
Maximum number of times a given range is adjusted during the dataflow.
2021-08-25 21:43:07 +02:00
2021-11-09 10:14:25 +01:00
-param=threader-debug=
Common Joined Var(param_threader_debug) Enum(threader_debug) Init(THREADER_DEBUG_NONE) Param Optimization
--param=threader-debug=[none|all] Enables verbose dumping of the threader solver.
Enum
Name(threader_debug) Type(enum threader_debug) UnknownError(unknown threader debug mode %qs)
EnumValue
Enum(threader_debug) String(none) Value(THREADER_DEBUG_NONE)
EnumValue
Enum(threader_debug) String(all) Value(THREADER_DEBUG_ALL)
2019-11-12 11:08:04 +01:00
-param=tm-max-aggregate-size=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_tm_max_aggregate_size) Init(9) Param Optimization
2019-11-12 11:08:04 +01:00
Size in bytes after which thread-local aggregates should be instrumented with the logging functions instead of save/restore pairs.
-param=tracer-dynamic-coverage=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_tracer_dynamic_coverage) Init(75) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
The percentage of function, weighted by execution frequency, that must be covered by trace formation. Used when profile feedback is not available.
-param=tracer-dynamic-coverage-feedback=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_tracer_dynamic_coverage_feedback) Init(95) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
The percentage of function, weighted by execution frequency, that must be covered by trace formation. Used when profile feedback is available.
-param=tracer-max-code-growth=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_tracer_max_code_growth) Init(100) Param Optimization
2019-11-12 11:08:04 +01:00
Maximal code growth caused by tail duplication (in percent).
-param=tracer-min-branch-probability=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_tracer_min_branch_probability) Init(50) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
Stop forward growth if the probability of best edge is less than this threshold (in percent). Used when profile feedback is not available.
-param=tracer-min-branch-probability-feedback=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_tracer_min_branch_probability_feedback) Init(80) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
Stop forward growth if the probability of best edge is less than this threshold (in percent). Used when profile feedback is available.
-param=tracer-min-branch-ratio=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_tracer_min_branch_ratio) Init(10) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
Stop reverse growth if the reverse probability of best edge is less than this threshold (in percent).
-param=tree-reassoc-width=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_tree_reassoc_width) Param Optimization
2019-11-12 11:08:04 +01:00
Set the maximum number of instructions executed in parallel in reassociated tree. If 0, use the target dependent heuristic.
tsan: Add optional support for distinguishing volatiles
Add support to optionally emit different instrumentation for accesses to
volatile variables. While the default TSAN runtime likely will never
require this feature, other runtimes for different environments that
have subtly different memory models or assumptions may require
distinguishing volatiles.
One such environment are OS kernels, where volatile is still used in
various places, and often declare volatile to be appropriate even in
multi-threaded contexts. One such example is the Linux kernel, which
implements various synchronization primitives using volatile
(READ_ONCE(), WRITE_ONCE()).
Here the Kernel Concurrency Sanitizer (KCSAN), is a runtime that uses
TSAN instrumentation but otherwise implements a very different approach
to race detection from TSAN:
https://github.com/google/ktsan/wiki/KCSAN
Due to recent changes in requirements by the Linux kernel, KCSAN
requires that the compiler supports tsan-distinguish-volatile (among
several new requirements):
https://lore.kernel.org/lkml/20200521142047.169334-7-elver@google.com/
gcc/
* params.opt: Define --param=tsan-distinguish-volatile=[0,1].
* sanitizer.def (BUILT_IN_TSAN_VOLATILE_READ1): Define new
builtin for volatile instrumentation of reads/writes.
(BUILT_IN_TSAN_VOLATILE_READ2): Likewise.
(BUILT_IN_TSAN_VOLATILE_READ4): Likewise.
(BUILT_IN_TSAN_VOLATILE_READ8): Likewise.
(BUILT_IN_TSAN_VOLATILE_READ16): Likewise.
(BUILT_IN_TSAN_VOLATILE_WRITE1): Likewise.
(BUILT_IN_TSAN_VOLATILE_WRITE2): Likewise.
(BUILT_IN_TSAN_VOLATILE_WRITE4): Likewise.
(BUILT_IN_TSAN_VOLATILE_WRITE8): Likewise.
(BUILT_IN_TSAN_VOLATILE_WRITE16): Likewise.
* tsan.c (get_memory_access_decl): Argument if access is
volatile. If param tsan-distinguish-volatile is non-zero, and
access if volatile, return volatile instrumentation decl.
(instrument_expr): Check if access is volatile.
gcc/testsuite/
* c-c++-common/tsan/volatile.c: New test.
2020-06-09 15:15:39 +02:00
-param=tsan-distinguish-volatile=
Common Joined UInteger Var(param_tsan_distinguish_volatile) IntegerRange(0, 1) Param
Emit special instrumentation for accesses to volatiles.
2020-06-10 16:21:21 +02:00
-param=tsan-instrument-func-entry-exit=
Common Joined UInteger Var(param_tsan_instrument_func_entry_exit) Init(1) IntegerRange(0, 1) Param
Emit instrumentation calls to __tsan_func_entry() and __tsan_func_exit().
2019-11-12 11:08:04 +01:00
-param=uninit-control-dep-attempts=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_uninit_control_dep_attempts) Init(1000) IntegerRange(1, 65536) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of nested calls to search for control dependencies during uninitialized variable analysis.
-param=uninlined-function-insns=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_uninlined_function_insns) Init(2) Optimization IntegerRange(0, 1000000) Param
2019-11-12 11:08:04 +01:00
Instruction accounted for function prologue, epilogue and other overhead.
-param=uninlined-function-time=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_uninlined_function_time) Optimization IntegerRange(0, 1000000) Param
2019-11-12 11:08:04 +01:00
Time accounted for function prologue, epilogue and other overhead.
-param=uninlined-thunk-insns=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_uninlined_function_thunk_insns) Optimization Init(2) IntegerRange(0, 1000000) Param
2019-11-12 11:08:04 +01:00
Instruction accounted for function thunk overhead.
-param=uninlined-thunk-time=
Convert inliner to function specific param infrastructure
This patch adds opt_for_fn for all cross module params used by inliner
so they can be modified at function granuality. With inlining almost always
there are three functions to consider (callee and caller of the inlined edge
and the outer function caller is inlined to).
I always use the outer function params since that is how local parameters
behave. I hope it is kind of what is also expected in most case: it is better
to inline agressively into -O3 compiled code rather than inline agressively -O3
functions into their callers.
New params infrastructure is nice. One drawback is that is very hard to
search for individual param uses since they all occupy global namespace.
With C++ world we had chance to do something like params.param_flag_name
or params::param_flag_name instead...
Bootstrapped/regtested x86_64-linux, comitted.
* cif-code.def (MAX_INLINE_INSNS_SINGLE_O2_LIMIT): Remove.
* doc/invoke.texi (max-inline-insns-single-O2,
inline-heuristics-hint-percent-O2, inline-min-speedup-O2,
early-inlining-insns-O2): Remove documentation.
* ipa-fnsummary.c (analyze_function_body,
compute_fn_summary): Use opt_for_fn when accessing parameters.
* ipa-inline.c (caller_growth_limits, can_inline_edge_p,
inline_insns_auto, can_inline_edge_by_limits_p,
want_early_inline_function_p, big_speedup_p,
want_inline_small_function_p, want_inline_self_recursive_call_p,
recursive_inlining, compute_max_insns, inline_small_functions):
Likewise.
* opts.c (default_options): Add -O3 defaults for
OPT__param_early_inlining_insns_,
OPT__param_inline_heuristics_hint_percent_,
OPT__param_inline_min_speedup_, OPT__param_max_inline_insns_single_.
* params.opt (-param=early-inlining-insns-O2=,
-param=inline-heuristics-hint-percent-O2=,
-param=inline-min-speedup-O2=, -param=max-inline-insns-single-O2=
-param=early-inlining-insns=, -param=inline-heuristics-hint-percent=,
-param=inline-min-speedup=, -param=inline-unit-growth=,
-param=large-function-growth=, -param=large-stack-frame=,
-param=large-stack-frame-growth=, -param=large-unit-insns=,
-param=max-inline-insns-recursive=,
-param=max-inline-insns-recursive-auto=,
-param=max-inline-insns-single=,
-param=max-inline-insns-size=, -param=max-inline-insns-small=,
-param=max-inline-recursive-depth=,
-param=max-inline-recursive-depth-auto=,
-param=min-inline-recursive-probability=,
-param=partial-inlining-entry-probability=,
-param=uninlined-function-insns=, -param=uninlined-function-time=,
-param=uninlined-thunk-insns=, -param=uninlined-thunk-time=): Add
Optimization.
* g++.dg/tree-ssa/pr53844.C: Drop -O2 from param name.
* g++.dg/tree-ssa/pr61034.C: Likewise.
* g++.dg/tree-ssa/pr8781.C: Likewise.
* g++.dg/warn/Wstringop-truncation-1.C: Likewise.
* gcc.dg/ipa/pr63416.c: Likewise.
* gcc.dg/tree-ssa/ssa-thread-12.c: Likewise.
* gcc.dg/vect/pr66142.c: Likewise.
* gcc.dg/winline-3.c: Likewise.
* gcc.target/powerpc/pr72804.c: Likewise.
From-SVN: r278644
2019-11-23 14:11:25 +01:00
Common Joined UInteger Var(param_uninlined_function_thunk_time) Optimization Init(2) IntegerRange(0, 1000000) Param
2019-11-12 11:08:04 +01:00
Time accounted for function thunk overhead.
-param=unlikely-bb-count-fraction=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_unlikely_bb_count_fraction) Init(20) Param Optimization
2019-11-12 11:08:04 +01:00
The denominator n of fraction 1/n of the number of profiled runs of the entire program below which the execution count of a basic block must be in order for the basic block to be considered unlikely.
-param=unroll-jam-max-unroll=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_unroll_jam_max_unroll) Init(4) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum unroll factor for the unroll-and-jam transformation.
-param=unroll-jam-min-percent=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_unroll_jam_min_percent) Init(1) IntegerRange(0, 100) Param Optimization
2019-11-12 11:08:04 +01:00
Minimum percentage of memrefs that must go away for unroll-and-jam to be considered profitable.
-param=use-after-scope-direct-emission-threshold=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_use_after_scope_direct_emission_threshold) Init(256) Param Optimization
2019-11-12 11:08:04 +01:00
Use direct poisoning/unpoisoning instructions for variables smaller or equal to this number.
-param=use-canonical-types=
Common Joined UInteger Var(param_use_canonical_types) Init(1) IntegerRange(0, 1) Param
Whether to use canonical types.
-param=vect-epilogues-nomask=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_vect_epilogues_nomask) Init(1) IntegerRange(0, 1) Param Optimization
2019-11-12 11:08:04 +01:00
Enable loop epilogue vectorization using smaller vector size.
-param=vect-max-peeling-for-alignment=
2021-10-25 16:54:22 +02:00
Common Joined UInteger Var(param_vect_max_peeling_for_alignment) Init(-1) IntegerRange(0, 64) Param Optimization
2019-11-12 11:08:04 +01:00
Maximum number of loop peels to enhance alignment of data references in a loop.
-param=vect-max-version-for-alias-checks=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_vect_max_version_for_alias_checks) Init(10) Param Optimization
2019-11-12 11:08:04 +01:00
Bound on number of runtime checks inserted by the vectorizer's loop versioning for alias check.
-param=vect-max-version-for-alignment-checks=
2020-01-07 10:21:01 +01:00
Common Joined UInteger Var(param_vect_max_version_for_alignment_checks) Init(6) Param Optimization
2019-11-12 11:08:04 +01:00
Bound on number of runtime checks inserted by the vectorizer's loop versioning for alignment check.
2020-07-20 03:40:10 +02:00
-param=vect-partial-vector-usage=
Common Joined UInteger Var(param_vect_partial_vector_usage) Init(2) IntegerRange(0, 2) Param Optimization
Controls how loop vectorizer uses partial vectors. 0 means never, 1 means only for loops whose need to iterate can be removed, 2 means for all loops. The default value is 2.
2021-05-19 12:42:51 +02:00
-param=vect-inner-loop-cost-factor=
2021-08-23 14:15:14 +02:00
Common Joined UInteger Var(param_vect_inner_loop_cost_factor) Init(50) IntegerRange(1, 10000) Param Optimization
The maximum factor which the loop vectorizer applies to the cost of statements in an inner loop relative to the loop being vectorized.
2021-05-19 12:42:51 +02:00
2022-03-08 12:07:07 +01:00
-param=vect-induction-float=
Common Joined UInteger Var(param_vect_induction_float) Init(1) IntegerRage(0, 1) Param Optimization
Enable loop vectorization of floating point inductions.
2021-10-15 18:06:27 +02:00
-param=vrp1-mode=
Common Joined Var(param_vrp1_mode) Enum(vrp_mode) Init(VRP_MODE_VRP) Param Optimization
--param=vrp1-mode=[vrp|ranger] Specifies the mode VRP1 should operate in.
-param=vrp2-mode=
2021-11-03 02:26:44 +01:00
Common Joined Var(param_vrp2_mode) Enum(vrp_mode) Init(VRP_MODE_RANGER) Param Optimization
2021-10-15 18:06:27 +02:00
--param=vrp2-mode=[vrp|ranger] Specifies the mode VRP2 should operate in.
Enum
Name(vrp_mode) Type(enum vrp_mode) UnknownError(unknown vrp mode %qs)
EnumValue
Enum(vrp_mode) String(vrp) Value(VRP_MODE_VRP)
EnumValue
Enum(vrp_mode) String(ranger) Value(VRP_MODE_RANGER)
2019-11-12 11:08:04 +01:00
; This comment is to ensure we retain the blank line above.
