/*
* Copyright (c) 2001, 2016, Oracle and/or its affiliates. All rights reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
* This code is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License version 2 only, as
* published by the Free Software Foundation.
*
* This code 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
* version 2 for more details (a copy is included in the LICENSE file that
* accompanied this code).
*
* You should have received a copy of the GNU General Public License version
* 2 along with this work; if not, write to the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
*
* Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
* or visit www.oracle.com if you need additional information or have any
* questions.
*
*/
#include "precompiled.hpp"
#include "gc/g1/concurrentG1Refine.hpp"
#include "gc/g1/concurrentMarkThread.inline.hpp"
#include "gc/g1/g1CollectedHeap.inline.hpp"
#include "gc/g1/g1CollectorPolicy.hpp"
#include "gc/g1/g1ConcurrentMark.hpp"
#include "gc/g1/g1IHOPControl.hpp"
#include "gc/g1/g1GCPhaseTimes.hpp"
#include "gc/g1/heapRegion.inline.hpp"
#include "gc/g1/heapRegionRemSet.hpp"
#include "gc/shared/gcPolicyCounters.hpp"
#include "runtime/arguments.hpp"
#include "runtime/java.hpp"
#include "runtime/mutexLocker.hpp"
#include "utilities/debug.hpp"
#include "utilities/pair.hpp"
// Different defaults for different number of GC threads
// They were chosen by running GCOld and SPECjbb on debris with different
// numbers of GC threads and choosing them based on the results
// all the same
static double rs_length_diff_defaults[] = {
0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0
};
static double cost_per_card_ms_defaults[] = {
0.01, 0.005, 0.005, 0.003, 0.003, 0.002, 0.002, 0.0015
};
// all the same
static double young_cards_per_entry_ratio_defaults[] = {
1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0
};
static double cost_per_entry_ms_defaults[] = {
0.015, 0.01, 0.01, 0.008, 0.008, 0.0055, 0.0055, 0.005
};
static double cost_per_byte_ms_defaults[] = {
0.00006, 0.00003, 0.00003, 0.000015, 0.000015, 0.00001, 0.00001, 0.000009
};
// these should be pretty consistent
static double constant_other_time_ms_defaults[] = {
5.0, 5.0, 5.0, 5.0, 5.0, 5.0, 5.0, 5.0
};
static double young_other_cost_per_region_ms_defaults[] = {
0.3, 0.2, 0.2, 0.15, 0.15, 0.12, 0.12, 0.1
};
static double non_young_other_cost_per_region_ms_defaults[] = {
1.0, 0.7, 0.7, 0.5, 0.5, 0.42, 0.42, 0.30
};
G1CollectorPolicy::G1CollectorPolicy() :
_predictor(G1ConfidencePercent / 100.0),
_recent_gc_times_ms(new TruncatedSeq(NumPrevPausesForHeuristics)),
_concurrent_mark_remark_times_ms(new TruncatedSeq(NumPrevPausesForHeuristics)),
_concurrent_mark_cleanup_times_ms(new TruncatedSeq(NumPrevPausesForHeuristics)),
_alloc_rate_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
_prev_collection_pause_end_ms(0.0),
_rs_length_diff_seq(new TruncatedSeq(TruncatedSeqLength)),
_cost_per_card_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
_cost_scan_hcc_seq(new TruncatedSeq(TruncatedSeqLength)),
_young_cards_per_entry_ratio_seq(new TruncatedSeq(TruncatedSeqLength)),
_mixed_cards_per_entry_ratio_seq(new TruncatedSeq(TruncatedSeqLength)),
_cost_per_entry_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
_mixed_cost_per_entry_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
_cost_per_byte_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
_cost_per_byte_ms_during_cm_seq(new TruncatedSeq(TruncatedSeqLength)),
_constant_other_time_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
_young_other_cost_per_region_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
_non_young_other_cost_per_region_ms_seq(
new TruncatedSeq(TruncatedSeqLength)),
_pending_cards_seq(new TruncatedSeq(TruncatedSeqLength)),
_rs_lengths_seq(new TruncatedSeq(TruncatedSeqLength)),
_pause_time_target_ms((double) MaxGCPauseMillis),
_recent_prev_end_times_for_all_gcs_sec(
new TruncatedSeq(NumPrevPausesForHeuristics)),
_recent_avg_pause_time_ratio(0.0),
_rs_lengths_prediction(0),
_max_survivor_regions(0),
_eden_cset_region_length(0),
_survivor_cset_region_length(0),
_old_cset_region_length(0),
_collection_set(NULL),
_collection_set_bytes_used_before(0),
// Incremental CSet attributes
_inc_cset_build_state(Inactive),
_inc_cset_head(NULL),
_inc_cset_tail(NULL),
_inc_cset_bytes_used_before(0),
_inc_cset_recorded_rs_lengths(0),
_inc_cset_recorded_rs_lengths_diffs(0),
_inc_cset_predicted_elapsed_time_ms(0.0),
_inc_cset_predicted_elapsed_time_ms_diffs(0.0),
// add here any more surv rate groups
_recorded_survivor_regions(0),
_recorded_survivor_head(NULL),
_recorded_survivor_tail(NULL),
_survivors_age_table(true),
_gc_overhead_perc(0.0),
_bytes_allocated_in_old_since_last_gc(0),
_ihop_control(NULL),
_initial_mark_to_mixed() {
// SurvRateGroups below must be initialized after the predictor because they
// indirectly use it through this object passed to their constructor.
_short_lived_surv_rate_group =
new SurvRateGroup(&_predictor, "Short Lived", G1YoungSurvRateNumRegionsSummary);
_survivor_surv_rate_group =
new SurvRateGroup(&_predictor, "Survivor", G1YoungSurvRateNumRegionsSummary);
// Set up the region size and associated fields. Given that the
// policy is created before the heap, we have to set this up here,
// so it's done as soon as possible.
// It would have been natural to pass initial_heap_byte_size() and
// max_heap_byte_size() to setup_heap_region_size() but those have
// not been set up at this point since they should be aligned with
// the region size. So, there is a circular dependency here. We base
// the region size on the heap size, but the heap size should be
// aligned with the region size. To get around this we use the
// unaligned values for the heap.
HeapRegion::setup_heap_region_size(InitialHeapSize, MaxHeapSize);
HeapRegionRemSet::setup_remset_size();
_recent_prev_end_times_for_all_gcs_sec->add(os::elapsedTime());
_prev_collection_pause_end_ms = os::elapsedTime() * 1000.0;
clear_ratio_check_data();
_phase_times = new G1GCPhaseTimes(ParallelGCThreads);
int index = MIN2(ParallelGCThreads - 1, 7u);
_rs_length_diff_seq->add(rs_length_diff_defaults[index]);
_cost_per_card_ms_seq->add(cost_per_card_ms_defaults[index]);
_cost_scan_hcc_seq->add(0.0);
_young_cards_per_entry_ratio_seq->add(
young_cards_per_entry_ratio_defaults[index]);
_cost_per_entry_ms_seq->add(cost_per_entry_ms_defaults[index]);
_cost_per_byte_ms_seq->add(cost_per_byte_ms_defaults[index]);
_constant_other_time_ms_seq->add(constant_other_time_ms_defaults[index]);
_young_other_cost_per_region_ms_seq->add(
young_other_cost_per_region_ms_defaults[index]);
_non_young_other_cost_per_region_ms_seq->add(
non_young_other_cost_per_region_ms_defaults[index]);
// Below, we might need to calculate the pause time target based on
// the pause interval. When we do so we are going to give G1 maximum
// flexibility and allow it to do pauses when it needs to. So, we'll
// arrange that the pause interval to be pause time target + 1 to
// ensure that a) the pause time target is maximized with respect to
// the pause interval and b) we maintain the invariant that pause
// time target < pause interval. If the user does not want this
// maximum flexibility, they will have to set the pause interval
// explicitly.
// First make sure that, if either parameter is set, its value is
// reasonable.
if (!FLAG_IS_DEFAULT(MaxGCPauseMillis)) {
if (MaxGCPauseMillis < 1) {
vm_exit_during_initialization("MaxGCPauseMillis should be "
"greater than 0");
}
}
if (!FLAG_IS_DEFAULT(GCPauseIntervalMillis)) {
if (GCPauseIntervalMillis < 1) {
vm_exit_during_initialization("GCPauseIntervalMillis should be "
"greater than 0");
}
}
// Then, if the pause time target parameter was not set, set it to
// the default value.
if (FLAG_IS_DEFAULT(MaxGCPauseMillis)) {
if (FLAG_IS_DEFAULT(GCPauseIntervalMillis)) {
// The default pause time target in G1 is 200ms
FLAG_SET_DEFAULT(MaxGCPauseMillis, 200);
} else {
// We do not allow the pause interval to be set without the
// pause time target
vm_exit_during_initialization("GCPauseIntervalMillis cannot be set "
"without setting MaxGCPauseMillis");
}
}
// Then, if the interval parameter was not set, set it according to
// the pause time target (this will also deal with the case when the
// pause time target is the default value).
if (FLAG_IS_DEFAULT(GCPauseIntervalMillis)) {
FLAG_SET_DEFAULT(GCPauseIntervalMillis, MaxGCPauseMillis + 1);
}
// Finally, make sure that the two parameters are consistent.
if (MaxGCPauseMillis >= GCPauseIntervalMillis) {
char buffer[256];
jio_snprintf(buffer, 256,
"MaxGCPauseMillis (%u) should be less than "
"GCPauseIntervalMillis (%u)",
MaxGCPauseMillis, GCPauseIntervalMillis);
vm_exit_during_initialization(buffer);
}
double max_gc_time = (double) MaxGCPauseMillis / 1000.0;
double time_slice = (double) GCPauseIntervalMillis / 1000.0;
_mmu_tracker = new G1MMUTrackerQueue(time_slice, max_gc_time);
// start conservatively (around 50ms is about right)
_concurrent_mark_remark_times_ms->add(0.05);
_concurrent_mark_cleanup_times_ms->add(0.20);
_tenuring_threshold = MaxTenuringThreshold;
assert(GCTimeRatio > 0,
"we should have set it to a default value set_g1_gc_flags() "
"if a user set it to 0");
_gc_overhead_perc = 100.0 * (1.0 / (1.0 + GCTimeRatio));
uintx reserve_perc = G1ReservePercent;
// Put an artificial ceiling on this so that it's not set to a silly value.
if (reserve_perc > 50) {
reserve_perc = 50;
warning("G1ReservePercent is set to a value that is too large, "
"it's been updated to " UINTX_FORMAT, reserve_perc);
}
_reserve_factor = (double) reserve_perc / 100.0;
// This will be set when the heap is expanded
// for the first time during initialization.
_reserve_regions = 0;
_cset_chooser = new CollectionSetChooser();
}
G1CollectorPolicy::~G1CollectorPolicy() {
delete _ihop_control;
}
double G1CollectorPolicy::get_new_prediction(TruncatedSeq const* seq) const {
return _predictor.get_new_prediction(seq);
}
size_t G1CollectorPolicy::get_new_size_prediction(TruncatedSeq const* seq) const {
return (size_t)get_new_prediction(seq);
}
void G1CollectorPolicy::initialize_alignments() {
_space_alignment = HeapRegion::GrainBytes;
size_t card_table_alignment = CardTableRS::ct_max_alignment_constraint();
size_t page_size = UseLargePages ? os::large_page_size() : os::vm_page_size();
_heap_alignment = MAX3(card_table_alignment, _space_alignment, page_size);
}
G1CollectorState* G1CollectorPolicy::collector_state() const { return _g1->collector_state(); }
// There are three command line options related to the young gen size:
// NewSize, MaxNewSize and NewRatio (There is also -Xmn, but that is
// just a short form for NewSize==MaxNewSize). G1 will use its internal
// heuristics to calculate the actual young gen size, so these options
// basically only limit the range within which G1 can pick a young gen
// size. Also, these are general options taking byte sizes. G1 will
// internally work with a number of regions instead. So, some rounding
// will occur.
//
// If nothing related to the the young gen size is set on the command
// line we should allow the young gen to be between G1NewSizePercent
// and G1MaxNewSizePercent of the heap size. This means that every time
// the heap size changes, the limits for the young gen size will be
// recalculated.
//
// If only -XX:NewSize is set we should use the specified value as the
// minimum size for young gen. Still using G1MaxNewSizePercent of the
// heap as maximum.
//
// If only -XX:MaxNewSize is set we should use the specified value as the
// maximum size for young gen. Still using G1NewSizePercent of the heap
// as minimum.
//
// If -XX:NewSize and -XX:MaxNewSize are both specified we use these values.
// No updates when the heap size changes. There is a special case when
// NewSize==MaxNewSize. This is interpreted as "fixed" and will use a
// different heuristic for calculating the collection set when we do mixed
// collection.
//
// If only -XX:NewRatio is set we should use the specified ratio of the heap
// as both min and max. This will be interpreted as "fixed" just like the
// NewSize==MaxNewSize case above. But we will update the min and max
// every time the heap size changes.
//
// NewSize and MaxNewSize override NewRatio. So, NewRatio is ignored if it is
// combined with either NewSize or MaxNewSize. (A warning message is printed.)
class G1YoungGenSizer : public CHeapObj<mtGC> {
private:
enum SizerKind {
SizerDefaults,
SizerNewSizeOnly,
SizerMaxNewSizeOnly,
SizerMaxAndNewSize,
SizerNewRatio
};
SizerKind _sizer_kind;
uint _min_desired_young_length;
uint _max_desired_young_length;
bool _adaptive_size;
uint calculate_default_min_length(uint new_number_of_heap_regions);
uint calculate_default_max_length(uint new_number_of_heap_regions);
// Update the given values for minimum and maximum young gen length in regions
// given the number of heap regions depending on the kind of sizing algorithm.
void recalculate_min_max_young_length(uint number_of_heap_regions, uint* min_young_length, uint* max_young_length);
public:
G1YoungGenSizer();
// Calculate the maximum length of the young gen given the number of regions
// depending on the sizing algorithm.
uint max_young_length(uint number_of_heap_regions);
void heap_size_changed(uint new_number_of_heap_regions);
uint min_desired_young_length() {
return _min_desired_young_length;
}
uint max_desired_young_length() {
return _max_desired_young_length;
}
bool adaptive_young_list_length() const {
return _adaptive_size;
}
};
G1YoungGenSizer::G1YoungGenSizer() : _sizer_kind(SizerDefaults), _adaptive_size(true),
_min_desired_young_length(0), _max_desired_young_length(0) {
if (FLAG_IS_CMDLINE(NewRatio)) {
if (FLAG_IS_CMDLINE(NewSize) || FLAG_IS_CMDLINE(MaxNewSize)) {
warning("-XX:NewSize and -XX:MaxNewSize override -XX:NewRatio");
} else {
_sizer_kind = SizerNewRatio;
_adaptive_size = false;
return;
}
}
if (NewSize > MaxNewSize) {
if (FLAG_IS_CMDLINE(MaxNewSize)) {
warning("NewSize (" SIZE_FORMAT "k) is greater than the MaxNewSize (" SIZE_FORMAT "k). "
"A new max generation size of " SIZE_FORMAT "k will be used.",
NewSize/K, MaxNewSize/K, NewSize/K);
}
MaxNewSize = NewSize;
}
if (FLAG_IS_CMDLINE(NewSize)) {
_min_desired_young_length = MAX2((uint) (NewSize / HeapRegion::GrainBytes),
1U);
if (FLAG_IS_CMDLINE(MaxNewSize)) {
_max_desired_young_length =
MAX2((uint) (MaxNewSize / HeapRegion::GrainBytes),
1U);
_sizer_kind = SizerMaxAndNewSize;
_adaptive_size = _min_desired_young_length == _max_desired_young_length;
} else {
_sizer_kind = SizerNewSizeOnly;
}
} else if (FLAG_IS_CMDLINE(MaxNewSize)) {
_max_desired_young_length =
MAX2((uint) (MaxNewSize / HeapRegion::GrainBytes),
1U);
_sizer_kind = SizerMaxNewSizeOnly;
}
}
uint G1YoungGenSizer::calculate_default_min_length(uint new_number_of_heap_regions) {
uint default_value = (new_number_of_heap_regions * G1NewSizePercent) / 100;
return MAX2(1U, default_value);
}
uint G1YoungGenSizer::calculate_default_max_length(uint new_number_of_heap_regions) {
uint default_value = (new_number_of_heap_regions * G1MaxNewSizePercent) / 100;
return MAX2(1U, default_value);
}
void G1YoungGenSizer::recalculate_min_max_young_length(uint number_of_heap_regions, uint* min_young_length, uint* max_young_length) {
assert(number_of_heap_regions > 0, "Heap must be initialized");
switch (_sizer_kind) {
case SizerDefaults:
*min_young_length = calculate_default_min_length(number_of_heap_regions);
*max_young_length = calculate_default_max_length(number_of_heap_regions);
break;
case SizerNewSizeOnly:
*max_young_length = calculate_default_max_length(number_of_heap_regions);
*max_young_length = MAX2(*min_young_length, *max_young_length);
break;
case SizerMaxNewSizeOnly:
*min_young_length = calculate_default_min_length(number_of_heap_regions);
*min_young_length = MIN2(*min_young_length, *max_young_length);
break;
case SizerMaxAndNewSize:
// Do nothing. Values set on the command line, don't update them at runtime.
break;
case SizerNewRatio:
*min_young_length = number_of_heap_regions / (NewRatio + 1);
*max_young_length = *min_young_length;
break;
default:
ShouldNotReachHere();
}
assert(*min_young_length <= *max_young_length, "Invalid min/max young gen size values");
}
uint G1YoungGenSizer::max_young_length(uint number_of_heap_regions) {
// We need to pass the desired values because recalculation may not update these
// values in some cases.
uint temp = _min_desired_young_length;
uint result = _max_desired_young_length;
recalculate_min_max_young_length(number_of_heap_regions, &temp, &result);
return result;
}
void G1YoungGenSizer::heap_size_changed(uint new_number_of_heap_regions) {
recalculate_min_max_young_length(new_number_of_heap_regions, &_min_desired_young_length,
&_max_desired_young_length);
}
void G1CollectorPolicy::post_heap_initialize() {
uintx max_regions = G1CollectedHeap::heap()->max_regions();
size_t max_young_size = (size_t)_young_gen_sizer->max_young_length(max_regions) * HeapRegion::GrainBytes;
if (max_young_size != MaxNewSize) {
FLAG_SET_ERGO(size_t, MaxNewSize, max_young_size);
}
_ihop_control = create_ihop_control();
}
void G1CollectorPolicy::initialize_flags() {
if (G1HeapRegionSize != HeapRegion::GrainBytes) {
FLAG_SET_ERGO(size_t, G1HeapRegionSize, HeapRegion::GrainBytes);
}
if (SurvivorRatio < 1) {
vm_exit_during_initialization("Invalid survivor ratio specified");
}
CollectorPolicy::initialize_flags();
_young_gen_sizer = new G1YoungGenSizer(); // Must be after call to initialize_flags
}
void G1CollectorPolicy::init() {
// Set aside an initial future to_space.
_g1 = G1CollectedHeap::heap();
assert(Heap_lock->owned_by_self(), "Locking discipline.");
initialize_gc_policy_counters();
if (adaptive_young_list_length()) {
_young_list_fixed_length = 0;
} else {
_young_list_fixed_length = _young_gen_sizer->min_desired_young_length();
}
_free_regions_at_end_of_collection = _g1->num_free_regions();
update_young_list_max_and_target_length();
// We may immediately start allocating regions and placing them on the
// collection set list. Initialize the per-collection set info
start_incremental_cset_building();
}
void G1CollectorPolicy::note_gc_start(uint num_active_workers) {
phase_times()->note_gc_start(num_active_workers);
}
// Create the jstat counters for the policy.
void G1CollectorPolicy::initialize_gc_policy_counters() {
_gc_policy_counters = new GCPolicyCounters("GarbageFirst", 1, 3);
}
bool G1CollectorPolicy::predict_will_fit(uint young_length,
double base_time_ms,
uint base_free_regions,
double target_pause_time_ms) const {
if (young_length >= base_free_regions) {
// end condition 1: not enough space for the young regions
return false;
}
double accum_surv_rate = accum_yg_surv_rate_pred((int) young_length - 1);
size_t bytes_to_copy =
(size_t) (accum_surv_rate * (double) HeapRegion::GrainBytes);
double copy_time_ms = predict_object_copy_time_ms(bytes_to_copy);
double young_other_time_ms = predict_young_other_time_ms(young_length);
double pause_time_ms = base_time_ms + copy_time_ms + young_other_time_ms;
if (pause_time_ms > target_pause_time_ms) {
// end condition 2: prediction is over the target pause time
return false;
}
size_t free_bytes = (base_free_regions - young_length) * HeapRegion::GrainBytes;
// When copying, we will likely need more bytes free than is live in the region.
// Add some safety margin to factor in the confidence of our guess, and the
// natural expected waste.
// (100.0 / G1ConfidencePercent) is a scale factor that expresses the uncertainty
// of the calculation: the lower the confidence, the more headroom.
// (100 + TargetPLABWastePct) represents the increase in expected bytes during
// copying due to anticipated waste in the PLABs.
double safety_factor = (100.0 / G1ConfidencePercent) * (100 + TargetPLABWastePct) / 100.0;
size_t expected_bytes_to_copy = (size_t)(safety_factor * bytes_to_copy);
if (expected_bytes_to_copy > free_bytes) {
// end condition 3: out-of-space
return false;
}
// success!
return true;
}
void G1CollectorPolicy::record_new_heap_size(uint new_number_of_regions) {
// re-calculate the necessary reserve
double reserve_regions_d = (double) new_number_of_regions * _reserve_factor;
// We use ceiling so that if reserve_regions_d is > 0.0 (but
// smaller than 1.0) we'll get 1.
_reserve_regions = (uint) ceil(reserve_regions_d);
_young_gen_sizer->heap_size_changed(new_number_of_regions);
}
uint G1CollectorPolicy::calculate_young_list_desired_min_length(
uint base_min_length) const {
uint desired_min_length = 0;
if (adaptive_young_list_length()) {
if (_alloc_rate_ms_seq->num() > 3) {
double now_sec = os::elapsedTime();
double when_ms = _mmu_tracker->when_max_gc_sec(now_sec) * 1000.0;
double alloc_rate_ms = predict_alloc_rate_ms();
desired_min_length = (uint) ceil(alloc_rate_ms * when_ms);
} else {
// otherwise we don't have enough info to make the prediction
}
}
desired_min_length += base_min_length;
// make sure we don't go below any user-defined minimum bound
return MAX2(_young_gen_sizer->min_desired_young_length(), desired_min_length);
}
uint G1CollectorPolicy::calculate_young_list_desired_max_length() const {
// Here, we might want to also take into account any additional
// constraints (i.e., user-defined minimum bound). Currently, we
// effectively don't set this bound.
return _young_gen_sizer->max_desired_young_length();
}
uint G1CollectorPolicy::update_young_list_max_and_target_length() {
return update_young_list_max_and_target_length(get_new_size_prediction(_rs_lengths_seq));
}
uint G1CollectorPolicy::update_young_list_max_and_target_length(size_t rs_lengths) {
uint unbounded_target_length = update_young_list_target_length(rs_lengths);
update_max_gc_locker_expansion();
return unbounded_target_length;
}
uint G1CollectorPolicy::update_young_list_target_length(size_t rs_lengths) {
YoungTargetLengths young_lengths = young_list_target_lengths(rs_lengths);
_young_list_target_length = young_lengths.first;
return young_lengths.second;
}
G1CollectorPolicy::YoungTargetLengths G1CollectorPolicy::young_list_target_lengths(size_t rs_lengths) const {
YoungTargetLengths result;
// Calculate the absolute and desired min bounds first.
// This is how many young regions we already have (currently: the survivors).
uint base_min_length = recorded_survivor_regions();
uint desired_min_length = calculate_young_list_desired_min_length(base_min_length);
// This is the absolute minimum young length. Ensure that we
// will at least have one eden region available for allocation.
uint absolute_min_length = base_min_length + MAX2(_g1->young_list()->eden_length(), (uint)1);
// If we shrank the young list target it should not shrink below the current size.
desired_min_length = MAX2(desired_min_length, absolute_min_length);
// Calculate the absolute and desired max bounds.
uint desired_max_length = calculate_young_list_desired_max_length();
uint young_list_target_length = 0;
if (adaptive_young_list_length()) {
if (collector_state()->gcs_are_young()) {
young_list_target_length =
calculate_young_list_target_length(rs_lengths,
base_min_length,
desired_min_length,
desired_max_length);
} else {
// Don't calculate anything and let the code below bound it to
// the desired_min_length, i.e., do the next GC as soon as
// possible to maximize how many old regions we can add to it.
}
} else {
// The user asked for a fixed young gen so we'll fix the young gen
// whether the next GC is young or mixed.
young_list_target_length = _young_list_fixed_length;
}
result.second = young_list_target_length;
// We will try our best not to "eat" into the reserve.
uint absolute_max_length = 0;
if (_free_regions_at_end_of_collection > _reserve_regions) {
absolute_max_length = _free_regions_at_end_of_collection - _reserve_regions;
}
if (desired_max_length > absolute_max_length) {
desired_max_length = absolute_max_length;
}
// Make sure we don't go over the desired max length, nor under the
// desired min length. In case they clash, desired_min_length wins
// which is why that test is second.
if (young_list_target_length > desired_max_length) {
young_list_target_length = desired_max_length;
}
if (young_list_target_length < desired_min_length) {
young_list_target_length = desired_min_length;
}
assert(young_list_target_length > recorded_survivor_regions(),
"we should be able to allocate at least one eden region");
assert(young_list_target_length >= absolute_min_length, "post-condition");
result.first = young_list_target_length;
return result;
}
uint
G1CollectorPolicy::calculate_young_list_target_length(size_t rs_lengths,
uint base_min_length,
uint desired_min_length,
uint desired_max_length) const {
assert(adaptive_young_list_length(), "pre-condition");
assert(collector_state()->gcs_are_young(), "only call this for young GCs");
// In case some edge-condition makes the desired max length too small...
if (desired_max_length <= desired_min_length) {
return desired_min_length;
}
// We'll adjust min_young_length and max_young_length not to include
// the already allocated young regions (i.e., so they reflect the
// min and max eden regions we'll allocate). The base_min_length
// will be reflected in the predictions by the
// survivor_regions_evac_time prediction.
assert(desired_min_length > base_min_length, "invariant");
uint min_young_length = desired_min_length - base_min_length;
assert(desired_max_length > base_min_length, "invariant");
uint max_young_length = desired_max_length - base_min_length;
double target_pause_time_ms = _mmu_tracker->max_gc_time() * 1000.0;
double survivor_regions_evac_time = predict_survivor_regions_evac_time();
size_t pending_cards = get_new_size_prediction(_pending_cards_seq);
size_t adj_rs_lengths = rs_lengths + predict_rs_length_diff();
size_t scanned_cards = predict_young_card_num(adj_rs_lengths);
double base_time_ms =
predict_base_elapsed_time_ms(pending_cards, scanned_cards) +
survivor_regions_evac_time;
uint available_free_regions = _free_regions_at_end_of_collection;
uint base_free_regions = 0;
if (available_free_regions > _reserve_regions) {
base_free_regions = available_free_regions - _reserve_regions;
}
// Here, we will make sure that the shortest young length that
// makes sense fits within the target pause time.
if (predict_will_fit(min_young_length, base_time_ms,
base_free_regions, target_pause_time_ms)) {
// The shortest young length will fit into the target pause time;
// we'll now check whether the absolute maximum number of young
// regions will fit in the target pause time. If not, we'll do
// a binary search between min_young_length and max_young_length.
if (predict_will_fit(max_young_length, base_time_ms,
base_free_regions, target_pause_time_ms)) {
// The maximum young length will fit into the target pause time.
// We are done so set min young length to the maximum length (as
// the result is assumed to be returned in min_young_length).
min_young_length = max_young_length;
} else {
// The maximum possible number of young regions will not fit within
// the target pause time so we'll search for the optimal
// length. The loop invariants are:
//
// min_young_length < max_young_length
// min_young_length is known to fit into the target pause time
// max_young_length is known not to fit into the target pause time
//
// Going into the loop we know the above hold as we've just
// checked them. Every time around the loop we check whether
// the middle value between min_young_length and
// max_young_length fits into the target pause time. If it
// does, it becomes the new min. If it doesn't, it becomes
// the new max. This way we maintain the loop invariants.
assert(min_young_length < max_young_length, "invariant");
uint diff = (max_young_length - min_young_length) / 2;
while (diff > 0) {
uint young_length = min_young_length + diff;
if (predict_will_fit(young_length, base_time_ms,
base_free_regions, target_pause_time_ms)) {
min_young_length = young_length;
} else {
max_young_length = young_length;
}
assert(min_young_length < max_young_length, "invariant");
diff = (max_young_length - min_young_length) / 2;
}
// The results is min_young_length which, according to the
// loop invariants, should fit within the target pause time.
// These are the post-conditions of the binary search above:
assert(min_young_length < max_young_length,
"otherwise we should have discovered that max_young_length "
"fits into the pause target and not done the binary search");
assert(predict_will_fit(min_young_length, base_time_ms,
base_free_regions, target_pause_time_ms),
"min_young_length, the result of the binary search, should "
"fit into the pause target");
assert(!predict_will_fit(min_young_length + 1, base_time_ms,
base_free_regions, target_pause_time_ms),
"min_young_length, the result of the binary search, should be "
"optimal, so no larger length should fit into the pause target");
}
} else {
// Even the minimum length doesn't fit into the pause time
// target, return it as the result nevertheless.
}
return base_min_length + min_young_length;
}
double G1CollectorPolicy::predict_survivor_regions_evac_time() const {
double survivor_regions_evac_time = 0.0;
for (HeapRegion * r = _recorded_survivor_head;
r != NULL && r != _recorded_survivor_tail->get_next_young_region();
r = r->get_next_young_region()) {
survivor_regions_evac_time += predict_region_elapsed_time_ms(r, collector_state()->gcs_are_young());
}
return survivor_regions_evac_time;
}
void G1CollectorPolicy::revise_young_list_target_length_if_necessary(size_t rs_lengths) {
guarantee( adaptive_young_list_length(), "should not call this otherwise" );
if (rs_lengths > _rs_lengths_prediction) {
// add 10% to avoid having to recalculate often
size_t rs_lengths_prediction = rs_lengths * 1100 / 1000;
update_rs_lengths_prediction(rs_lengths_prediction);
update_young_list_max_and_target_length(rs_lengths_prediction);
}
}
void G1CollectorPolicy::update_rs_lengths_prediction() {
update_rs_lengths_prediction(get_new_size_prediction(_rs_lengths_seq));
}
void G1CollectorPolicy::update_rs_lengths_prediction(size_t prediction) {
if (collector_state()->gcs_are_young() && adaptive_young_list_length()) {
_rs_lengths_prediction = prediction;
}
}
#ifndef PRODUCT
bool G1CollectorPolicy::verify_young_ages() {
HeapRegion* head = _g1->young_list()->first_region();
return
verify_young_ages(head, _short_lived_surv_rate_group);
// also call verify_young_ages on any additional surv rate groups
}
bool
G1CollectorPolicy::verify_young_ages(HeapRegion* head,
SurvRateGroup *surv_rate_group) {
guarantee( surv_rate_group != NULL, "pre-condition" );
const char* name = surv_rate_group->name();
bool ret = true;
int prev_age = -1;
for (HeapRegion* curr = head;
curr != NULL;
curr = curr->get_next_young_region()) {
SurvRateGroup* group = curr->surv_rate_group();
if (group == NULL && !curr->is_survivor()) {
log_error(gc, verify)("## %s: encountered NULL surv_rate_group", name);
ret = false;
}
if (surv_rate_group == group) {
int age = curr->age_in_surv_rate_group();
if (age < 0) {
log_error(gc, verify)("## %s: encountered negative age", name);
ret = false;
}
if (age <= prev_age) {
log_error(gc, verify)("## %s: region ages are not strictly increasing (%d, %d)", name, age, prev_age);
ret = false;
}
prev_age = age;
}
}
return ret;
}
#endif // PRODUCT
void G1CollectorPolicy::record_full_collection_start() {
_full_collection_start_sec = os::elapsedTime();
// Release the future to-space so that it is available for compaction into.
collector_state()->set_full_collection(true);
}
void G1CollectorPolicy::record_full_collection_end() {
// Consider this like a collection pause for the purposes of allocation
// since last pause.
double end_sec = os::elapsedTime();
double full_gc_time_sec = end_sec - _full_collection_start_sec;
double full_gc_time_ms = full_gc_time_sec * 1000.0;
update_recent_gc_times(end_sec, full_gc_time_ms);
collector_state()->set_full_collection(false);
// "Nuke" the heuristics that control the young/mixed GC
// transitions and make sure we start with young GCs after the Full GC.
collector_state()->set_gcs_are_young(true);
collector_state()->set_last_young_gc(false);
collector_state()->set_initiate_conc_mark_if_possible(need_to_start_conc_mark("end of Full GC", 0));
collector_state()->set_during_initial_mark_pause(false);
collector_state()->set_in_marking_window(false);
collector_state()->set_in_marking_window_im(false);
_short_lived_surv_rate_group->start_adding_regions();
// also call this on any additional surv rate groups
record_survivor_regions(0, NULL, NULL);
_free_regions_at_end_of_collection = _g1->num_free_regions();
// Reset survivors SurvRateGroup.
_survivor_surv_rate_group->reset();
update_young_list_max_and_target_length();
update_rs_lengths_prediction();
cset_chooser()->clear();
_bytes_allocated_in_old_since_last_gc = 0;
record_pause(FullGC, _full_collection_start_sec, end_sec);
}
void G1CollectorPolicy::record_collection_pause_start(double start_time_sec) {
// We only need to do this here as the policy will only be applied
// to the GC we're about to start. so, no point is calculating this
// every time we calculate / recalculate the target young length.
update_survivors_policy();
assert(_g1->used() == _g1->recalculate_used(),
"sanity, used: " SIZE_FORMAT " recalculate_used: " SIZE_FORMAT,
_g1->used(), _g1->recalculate_used());
phase_times()->record_cur_collection_start_sec(start_time_sec);
_pending_cards = _g1->pending_card_num();
_collection_set_bytes_used_before = 0;
_bytes_copied_during_gc = 0;
collector_state()->set_last_gc_was_young(false);
// do that for any other surv rate groups
_short_lived_surv_rate_group->stop_adding_regions();
_survivors_age_table.clear();
assert( verify_young_ages(), "region age verification" );
}
void G1CollectorPolicy::record_concurrent_mark_init_end(double
mark_init_elapsed_time_ms) {
collector_state()->set_during_marking(true);
assert(!collector_state()->initiate_conc_mark_if_possible(), "we should have cleared it by now");
collector_state()->set_during_initial_mark_pause(false);
}
void G1CollectorPolicy::record_concurrent_mark_remark_start() {
_mark_remark_start_sec = os::elapsedTime();
collector_state()->set_during_marking(false);
}
void G1CollectorPolicy::record_concurrent_mark_remark_end() {
double end_time_sec = os::elapsedTime();
double elapsed_time_ms = (end_time_sec - _mark_remark_start_sec)*1000.0;
_concurrent_mark_remark_times_ms->add(elapsed_time_ms);
_prev_collection_pause_end_ms += elapsed_time_ms;
record_pause(Remark, _mark_remark_start_sec, end_time_sec);
}
void G1CollectorPolicy::record_concurrent_mark_cleanup_start() {
_mark_cleanup_start_sec = os::elapsedTime();
}
void G1CollectorPolicy::record_concurrent_mark_cleanup_completed() {
bool should_continue_with_reclaim = next_gc_should_be_mixed("request last young-only gc",
"skip last young-only gc");
collector_state()->set_last_young_gc(should_continue_with_reclaim);
// We skip the marking phase.
if (!should_continue_with_reclaim) {
abort_time_to_mixed_tracking();
}
collector_state()->set_in_marking_window(false);
}
double G1CollectorPolicy::average_time_ms(G1GCPhaseTimes::GCParPhases phase) const {
return phase_times()->average_time_ms(phase);
}
double G1CollectorPolicy::young_other_time_ms() const {
return phase_times()->young_cset_choice_time_ms() +
phase_times()->young_free_cset_time_ms();
}
double G1CollectorPolicy::non_young_other_time_ms() const {
return phase_times()->non_young_cset_choice_time_ms() +
phase_times()->non_young_free_cset_time_ms();
}
double G1CollectorPolicy::other_time_ms(double pause_time_ms) const {
return pause_time_ms -
average_time_ms(G1GCPhaseTimes::UpdateRS) -
average_time_ms(G1GCPhaseTimes::ScanRS) -
average_time_ms(G1GCPhaseTimes::ObjCopy) -
average_time_ms(G1GCPhaseTimes::Termination);
}
double G1CollectorPolicy::constant_other_time_ms(double pause_time_ms) const {
return other_time_ms(pause_time_ms) - young_other_time_ms() - non_young_other_time_ms();
}
bool G1CollectorPolicy::about_to_start_mixed_phase() const {
return _g1->concurrent_mark()->cmThread()->during_cycle() || collector_state()->last_young_gc();
}
bool G1CollectorPolicy::need_to_start_conc_mark(const char* source, size_t alloc_word_size) {
if (about_to_start_mixed_phase()) {
return false;
}
size_t marking_initiating_used_threshold = _ihop_control->get_conc_mark_start_threshold();
size_t cur_used_bytes = _g1->non_young_capacity_bytes();
size_t alloc_byte_size = alloc_word_size * HeapWordSize;
size_t marking_request_bytes = cur_used_bytes + alloc_byte_size;
bool result = false;
if (marking_request_bytes > marking_initiating_used_threshold) {
result = collector_state()->gcs_are_young() && !collector_state()->last_young_gc();
log_debug(gc, ergo, ihop)("%s occupancy: " SIZE_FORMAT "B allocation request: " SIZE_FORMAT "B threshold: " SIZE_FORMAT "B (%1.2f) source: %s",
result ? "Request concurrent cycle initiation (occupancy higher than threshold)" : "Do not request concurrent cycle initiation (still doing mixed collections)",
cur_used_bytes, alloc_byte_size, marking_initiating_used_threshold, (double) marking_initiating_used_threshold / _g1->capacity() * 100, source);
}
return result;
}
// Anything below that is considered to be zero
#define MIN_TIMER_GRANULARITY 0.0000001
void G1CollectorPolicy::record_collection_pause_end(double pause_time_ms, size_t cards_scanned, size_t heap_used_bytes_before_gc) {
double end_time_sec = os::elapsedTime();
size_t cur_used_bytes = _g1->used();
assert(cur_used_bytes == _g1->recalculate_used(), "It should!");
bool last_pause_included_initial_mark = false;
bool update_stats = !_g1->evacuation_failed();
NOT_PRODUCT(_short_lived_surv_rate_group->print());
record_pause(young_gc_pause_kind(), end_time_sec - pause_time_ms / 1000.0, end_time_sec);
last_pause_included_initial_mark = collector_state()->during_initial_mark_pause();
if (last_pause_included_initial_mark) {
record_concurrent_mark_init_end(0.0);
} else {
maybe_start_marking();
}
double app_time_ms = (phase_times()->cur_collection_start_sec() * 1000.0 - _prev_collection_pause_end_ms);
if (app_time_ms < MIN_TIMER_GRANULARITY) {
// This usually happens due to the timer not having the required
// granularity. Some Linuxes are the usual culprits.
// We'll just set it to something (arbitrarily) small.
app_time_ms = 1.0;
}
if (update_stats) {
// We maintain the invariant that all objects allocated by mutator
// threads will be allocated out of eden regions. So, we can use
// the eden region number allocated since the previous GC to
// calculate the application's allocate rate. The only exception
// to that is humongous objects that are allocated separately. But
// given that humongous object allocations do not really affect
// either the pause's duration nor when the next pause will take
// place we can safely ignore them here.
uint regions_allocated = eden_cset_region_length();
double alloc_rate_ms = (double) regions_allocated / app_time_ms;
_alloc_rate_ms_seq->add(alloc_rate_ms);
double interval_ms =
(end_time_sec - _recent_prev_end_times_for_all_gcs_sec->oldest()) * 1000.0;
update_recent_gc_times(end_time_sec, pause_time_ms);
_recent_avg_pause_time_ratio = _recent_gc_times_ms->sum()/interval_ms;
if (recent_avg_pause_time_ratio() < 0.0 ||
(recent_avg_pause_time_ratio() - 1.0 > 0.0)) {
// Clip ratio between 0.0 and 1.0, and continue. This will be fixed in
// CR 6902692 by redoing the manner in which the ratio is incrementally computed.
if (_recent_avg_pause_time_ratio < 0.0) {
_recent_avg_pause_time_ratio = 0.0;
} else {
assert(_recent_avg_pause_time_ratio - 1.0 > 0.0, "Ctl-point invariant");
_recent_avg_pause_time_ratio = 1.0;
}
}
// Compute the ratio of just this last pause time to the entire time range stored
// in the vectors. Comparing this pause to the entire range, rather than only the
// most recent interval, has the effect of smoothing over a possible transient 'burst'
// of more frequent pauses that don't really reflect a change in heap occupancy.
// This reduces the likelihood of a needless heap expansion being triggered.
_last_pause_time_ratio =
(pause_time_ms * _recent_prev_end_times_for_all_gcs_sec->num()) / interval_ms;
}
bool new_in_marking_window = collector_state()->in_marking_window();
bool new_in_marking_window_im = false;
if (last_pause_included_initial_mark) {
new_in_marking_window = true;
new_in_marking_window_im = true;
}
if (collector_state()->last_young_gc()) {
// This is supposed to to be the "last young GC" before we start
// doing mixed GCs. Here we decide whether to start mixed GCs or not.
assert(!last_pause_included_initial_mark, "The last young GC is not allowed to be an initial mark GC");
if (next_gc_should_be_mixed("start mixed GCs",
"do not start mixed GCs")) {
collector_state()->set_gcs_are_young(false);
} else {
// We aborted the mixed GC phase early.
abort_time_to_mixed_tracking();
}
collector_state()->set_last_young_gc(false);
}
if (!collector_state()->last_gc_was_young()) {
// This is a mixed GC. Here we decide whether to continue doing
// mixed GCs or not.
if (!next_gc_should_be_mixed("continue mixed GCs",
"do not continue mixed GCs")) {
collector_state()->set_gcs_are_young(true);
maybe_start_marking();
}
}
_short_lived_surv_rate_group->start_adding_regions();
// Do that for any other surv rate groups
double scan_hcc_time_ms = ConcurrentG1Refine::hot_card_cache_enabled() ? average_time_ms(G1GCPhaseTimes::ScanHCC) : 0.0;
if (update_stats) {
double cost_per_card_ms = 0.0;
if (_pending_cards > 0) {
cost_per_card_ms = (average_time_ms(G1GCPhaseTimes::UpdateRS) - scan_hcc_time_ms) / (double) _pending_cards;
_cost_per_card_ms_seq->add(cost_per_card_ms);
}
_cost_scan_hcc_seq->add(scan_hcc_time_ms);
double cost_per_entry_ms = 0.0;
if (cards_scanned > 10) {
cost_per_entry_ms = average_time_ms(G1GCPhaseTimes::ScanRS) / (double) cards_scanned;
if (collector_state()->last_gc_was_young()) {
_cost_per_entry_ms_seq->add(cost_per_entry_ms);
} else {
_mixed_cost_per_entry_ms_seq->add(cost_per_entry_ms);
}
}
if (_max_rs_lengths > 0) {
double cards_per_entry_ratio =
(double) cards_scanned / (double) _max_rs_lengths;
if (collector_state()->last_gc_was_young()) {
_young_cards_per_entry_ratio_seq->add(cards_per_entry_ratio);
} else {
_mixed_cards_per_entry_ratio_seq->add(cards_per_entry_ratio);
}
}
// This is defensive. For a while _max_rs_lengths could get
// smaller than _recorded_rs_lengths which was causing
// rs_length_diff to get very large and mess up the RSet length
// predictions. The reason was unsafe concurrent updates to the
// _inc_cset_recorded_rs_lengths field which the code below guards
// against (see CR 7118202). This bug has now been fixed (see CR
// 7119027). However, I'm still worried that
// _inc_cset_recorded_rs_lengths might still end up somewhat
// inaccurate. The concurrent refinement thread calculates an
// RSet's length concurrently with other CR threads updating it
// which might cause it to calculate the length incorrectly (if,
// say, it's in mid-coarsening). So I'll leave in the defensive
// conditional below just in case.
size_t rs_length_diff = 0;
if (_max_rs_lengths > _recorded_rs_lengths) {
rs_length_diff = _max_rs_lengths - _recorded_rs_lengths;
}
_rs_length_diff_seq->add((double) rs_length_diff);
size_t freed_bytes = heap_used_bytes_before_gc - cur_used_bytes;
size_t copied_bytes = _collection_set_bytes_used_before - freed_bytes;
double cost_per_byte_ms = 0.0;
if (copied_bytes > 0) {
cost_per_byte_ms = average_time_ms(G1GCPhaseTimes::ObjCopy) / (double) copied_bytes;
if (collector_state()->in_marking_window()) {
_cost_per_byte_ms_during_cm_seq->add(cost_per_byte_ms);
} else {
_cost_per_byte_ms_seq->add(cost_per_byte_ms);
}
}
if (young_cset_region_length() > 0) {
_young_other_cost_per_region_ms_seq->add(young_other_time_ms() /
young_cset_region_length());
}
if (old_cset_region_length() > 0) {
_non_young_other_cost_per_region_ms_seq->add(non_young_other_time_ms() /
old_cset_region_length());
}
_constant_other_time_ms_seq->add(constant_other_time_ms(pause_time_ms));
_pending_cards_seq->add((double) _pending_cards);
_rs_lengths_seq->add((double) _max_rs_lengths);
}
collector_state()->set_in_marking_window(new_in_marking_window);
collector_state()->set_in_marking_window_im(new_in_marking_window_im);
_free_regions_at_end_of_collection = _g1->num_free_regions();
// IHOP control wants to know the expected young gen length if it were not
// restrained by the heap reserve. Using the actual length would make the
// prediction too small and the limit the young gen every time we get to the
// predicted target occupancy.
size_t last_unrestrained_young_length = update_young_list_max_and_target_length();
update_rs_lengths_prediction();
update_ihop_prediction(app_time_ms / 1000.0,
_bytes_allocated_in_old_since_last_gc,
last_unrestrained_young_length * HeapRegion::GrainBytes);
_bytes_allocated_in_old_since_last_gc = 0;
_ihop_control->send_trace_event(_g1->gc_tracer_stw());
// Note that _mmu_tracker->max_gc_time() returns the time in seconds.
double update_rs_time_goal_ms = _mmu_tracker->max_gc_time() * MILLIUNITS * G1RSetUpdatingPauseTimePercent / 100.0;
if (update_rs_time_goal_ms < scan_hcc_time_ms) {
log_debug(gc, ergo, refine)("Adjust concurrent refinement thresholds (scanning the HCC expected to take longer than Update RS time goal)."
"Update RS time goal: %1.2fms Scan HCC time: %1.2fms",
update_rs_time_goal_ms, scan_hcc_time_ms);
update_rs_time_goal_ms = 0;
} else {
update_rs_time_goal_ms -= scan_hcc_time_ms;
}
adjust_concurrent_refinement(average_time_ms(G1GCPhaseTimes::UpdateRS) - scan_hcc_time_ms,
phase_times()->sum_thread_work_items(G1GCPhaseTimes::UpdateRS),
update_rs_time_goal_ms);
cset_chooser()->verify();
}
G1IHOPControl* G1CollectorPolicy::create_ihop_control() const {
if (G1UseAdaptiveIHOP) {
return new G1AdaptiveIHOPControl(InitiatingHeapOccupancyPercent,
G1CollectedHeap::heap()->max_capacity(),
&_predictor,
G1ReservePercent,
G1HeapWastePercent);
} else {
return new G1StaticIHOPControl(InitiatingHeapOccupancyPercent,
G1CollectedHeap::heap()->max_capacity());
}
}
void G1CollectorPolicy::update_ihop_prediction(double mutator_time_s,
size_t mutator_alloc_bytes,
size_t young_gen_size) {
// Always try to update IHOP prediction. Even evacuation failures give information
// about e.g. whether to start IHOP earlier next time.
// Avoid using really small application times that might create samples with
// very high or very low values. They may be caused by e.g. back-to-back gcs.
double const min_valid_time = 1e-6;
bool report = false;
double marking_to_mixed_time = -1.0;
if (!collector_state()->last_gc_was_young() && _initial_mark_to_mixed.has_result()) {
marking_to_mixed_time = _initial_mark_to_mixed.last_marking_time();
assert(marking_to_mixed_time > 0.0,
"Initial mark to mixed time must be larger than zero but is %.3f",
marking_to_mixed_time);
if (marking_to_mixed_time > min_valid_time) {
_ihop_control->update_marking_length(marking_to_mixed_time);
report = true;
}
}
// As an approximation for the young gc promotion rates during marking we use
// all of them. In many applications there are only a few if any young gcs during
// marking, which makes any prediction useless. This increases the accuracy of the
// prediction.
if (collector_state()->last_gc_was_young() && mutator_time_s > min_valid_time) {
_ihop_control->update_allocation_info(mutator_time_s, mutator_alloc_bytes, young_gen_size);
report = true;
}
if (report) {
report_ihop_statistics();
}
}
void G1CollectorPolicy::report_ihop_statistics() {
_ihop_control->print();
}
void G1CollectorPolicy::print_phases() {
phase_times()->print();
}
void G1CollectorPolicy::adjust_concurrent_refinement(double update_rs_time,
double update_rs_processed_buffers,
double goal_ms) {
DirtyCardQueueSet& dcqs = JavaThread::dirty_card_queue_set();
ConcurrentG1Refine *cg1r = G1CollectedHeap::heap()->concurrent_g1_refine();
if (G1UseAdaptiveConcRefinement) {
const int k_gy = 3, k_gr = 6;
const double inc_k = 1.1, dec_k = 0.9;
size_t g = cg1r->green_zone();
if (update_rs_time > goal_ms) {
g = (size_t)(g * dec_k); // Can become 0, that's OK. That would mean a mutator-only processing.
} else {
if (update_rs_time < goal_ms && update_rs_processed_buffers > g) {
g = (size_t)MAX2(g * inc_k, g + 1.0);
}
}
// Change the refinement threads params
cg1r->set_green_zone(g);
cg1r->set_yellow_zone(g * k_gy);
cg1r->set_red_zone(g * k_gr);
cg1r->reinitialize_threads();
size_t processing_threshold_delta = MAX2<size_t>(cg1r->green_zone() * _predictor.sigma(), 1);
size_t processing_threshold = MIN2(cg1r->green_zone() + processing_threshold_delta,
cg1r->yellow_zone());
// Change the barrier params
dcqs.set_process_completed_threshold((int)processing_threshold);
dcqs.set_max_completed_queue((int)cg1r->red_zone());
}
size_t curr_queue_size = dcqs.completed_buffers_num();
if (curr_queue_size >= cg1r->yellow_zone()) {
dcqs.set_completed_queue_padding(curr_queue_size);
} else {
dcqs.set_completed_queue_padding(0);
}
dcqs.notify_if_necessary();
}
size_t G1CollectorPolicy::predict_rs_length_diff() const {
return get_new_size_prediction(_rs_length_diff_seq);
}
double G1CollectorPolicy::predict_alloc_rate_ms() const {
return get_new_prediction(_alloc_rate_ms_seq);
}
double G1CollectorPolicy::predict_cost_per_card_ms() const {
return get_new_prediction(_cost_per_card_ms_seq);
}
double G1CollectorPolicy::predict_scan_hcc_ms() const {
return get_new_prediction(_cost_scan_hcc_seq);
}
double G1CollectorPolicy::predict_rs_update_time_ms(size_t pending_cards) const {
return pending_cards * predict_cost_per_card_ms() + predict_scan_hcc_ms();
}
double G1CollectorPolicy::predict_young_cards_per_entry_ratio() const {
return get_new_prediction(_young_cards_per_entry_ratio_seq);
}
double G1CollectorPolicy::predict_mixed_cards_per_entry_ratio() const {
if (_mixed_cards_per_entry_ratio_seq->num() < 2) {
return predict_young_cards_per_entry_ratio();
} else {
return get_new_prediction(_mixed_cards_per_entry_ratio_seq);
}
}
size_t G1CollectorPolicy::predict_young_card_num(size_t rs_length) const {
return (size_t) (rs_length * predict_young_cards_per_entry_ratio());
}
size_t G1CollectorPolicy::predict_non_young_card_num(size_t rs_length) const {
return (size_t)(rs_length * predict_mixed_cards_per_entry_ratio());
}
double G1CollectorPolicy::predict_rs_scan_time_ms(size_t card_num) const {
if (collector_state()->gcs_are_young()) {
return card_num * get_new_prediction(_cost_per_entry_ms_seq);
} else {
return predict_mixed_rs_scan_time_ms(card_num);
}
}
double G1CollectorPolicy::predict_mixed_rs_scan_time_ms(size_t card_num) const {
if (_mixed_cost_per_entry_ms_seq->num() < 3) {
return card_num * get_new_prediction(_cost_per_entry_ms_seq);
} else {
return card_num * get_new_prediction(_mixed_cost_per_entry_ms_seq);
}
}
double G1CollectorPolicy::predict_object_copy_time_ms_during_cm(size_t bytes_to_copy) const {
if (_cost_per_byte_ms_during_cm_seq->num() < 3) {
return (1.1 * bytes_to_copy) * get_new_prediction(_cost_per_byte_ms_seq);
} else {
return bytes_to_copy * get_new_prediction(_cost_per_byte_ms_during_cm_seq);
}
}
double G1CollectorPolicy::predict_object_copy_time_ms(size_t bytes_to_copy) const {
if (collector_state()->during_concurrent_mark()) {
return predict_object_copy_time_ms_during_cm(bytes_to_copy);
} else {
return bytes_to_copy * get_new_prediction(_cost_per_byte_ms_seq);
}
}
double G1CollectorPolicy::predict_constant_other_time_ms() const {
return get_new_prediction(_constant_other_time_ms_seq);
}
double G1CollectorPolicy::predict_young_other_time_ms(size_t young_num) const {
return young_num * get_new_prediction(_young_other_cost_per_region_ms_seq);
}
double G1CollectorPolicy::predict_non_young_other_time_ms(size_t non_young_num) const {
return non_young_num * get_new_prediction(_non_young_other_cost_per_region_ms_seq);
}
double G1CollectorPolicy::predict_remark_time_ms() const {
return get_new_prediction(_concurrent_mark_remark_times_ms);
}
double G1CollectorPolicy::predict_cleanup_time_ms() const {
return get_new_prediction(_concurrent_mark_cleanup_times_ms);
}
double G1CollectorPolicy::predict_yg_surv_rate(int age, SurvRateGroup* surv_rate_group) const {
TruncatedSeq* seq = surv_rate_group->get_seq(age);
guarantee(seq->num() > 0, "There should be some young gen survivor samples available. Tried to access with age %d", age);
double pred = get_new_prediction(seq);
if (pred > 1.0) {
pred = 1.0;
}
return pred;
}
double G1CollectorPolicy::predict_yg_surv_rate(int age) const {
return predict_yg_surv_rate(age, _short_lived_surv_rate_group);
}
double G1CollectorPolicy::accum_yg_surv_rate_pred(int age) const {
return _short_lived_surv_rate_group->accum_surv_rate_pred(age);
}
double G1CollectorPolicy::predict_base_elapsed_time_ms(size_t pending_cards,
size_t scanned_cards) const {
return
predict_rs_update_time_ms(pending_cards) +
predict_rs_scan_time_ms(scanned_cards) +
predict_constant_other_time_ms();
}
double G1CollectorPolicy::predict_base_elapsed_time_ms(size_t pending_cards) const {
size_t rs_length = predict_rs_length_diff();
size_t card_num;
if (collector_state()->gcs_are_young()) {
card_num = predict_young_card_num(rs_length);
} else {
card_num = predict_non_young_card_num(rs_length);
}
return predict_base_elapsed_time_ms(pending_cards, card_num);
}
size_t G1CollectorPolicy::predict_bytes_to_copy(HeapRegion* hr) const {
size_t bytes_to_copy;
if (hr->is_marked())
bytes_to_copy = hr->max_live_bytes();
else {
assert(hr->is_young() && hr->age_in_surv_rate_group() != -1, "invariant");
int age = hr->age_in_surv_rate_group();
double yg_surv_rate = predict_yg_surv_rate(age, hr->surv_rate_group());
bytes_to_copy = (size_t) (hr->used() * yg_surv_rate);
}
return bytes_to_copy;
}
double G1CollectorPolicy::predict_region_elapsed_time_ms(HeapRegion* hr,
bool for_young_gc) const {
size_t rs_length = hr->rem_set()->occupied();
size_t card_num;
// Predicting the number of cards is based on which type of GC
// we're predicting for.
if (for_young_gc) {
card_num = predict_young_card_num(rs_length);
} else {
card_num = predict_non_young_card_num(rs_length);
}
size_t bytes_to_copy = predict_bytes_to_copy(hr);
double region_elapsed_time_ms =
predict_rs_scan_time_ms(card_num) +
predict_object_copy_time_ms(bytes_to_copy);
// The prediction of the "other" time for this region is based
// upon the region type and NOT the GC type.
if (hr->is_young()) {
region_elapsed_time_ms += predict_young_other_time_ms(1);
} else {
region_elapsed_time_ms += predict_non_young_other_time_ms(1);
}
return region_elapsed_time_ms;
}
void G1CollectorPolicy::init_cset_region_lengths(uint eden_cset_region_length,
uint survivor_cset_region_length) {
_eden_cset_region_length = eden_cset_region_length;
_survivor_cset_region_length = survivor_cset_region_length;
_old_cset_region_length = 0;
}
void G1CollectorPolicy::set_recorded_rs_lengths(size_t rs_lengths) {
_recorded_rs_lengths = rs_lengths;
}
void G1CollectorPolicy::update_recent_gc_times(double end_time_sec,
double elapsed_ms) {
_recent_gc_times_ms->add(elapsed_ms);
_recent_prev_end_times_for_all_gcs_sec->add(end_time_sec);
_prev_collection_pause_end_ms = end_time_sec * 1000.0;
}
void G1CollectorPolicy::clear_ratio_check_data() {
_ratio_over_threshold_count = 0;
_ratio_over_threshold_sum = 0.0;
_pauses_since_start = 0;
}
size_t G1CollectorPolicy::expansion_amount() {
double recent_gc_overhead = recent_avg_pause_time_ratio() * 100.0;
double last_gc_overhead = _last_pause_time_ratio * 100.0;
double threshold = _gc_overhead_perc;
size_t expand_bytes = 0;
// If the heap is at less than half its maximum size, scale the threshold down,
// to a limit of 1. Thus the smaller the heap is, the more likely it is to expand,
// though the scaling code will likely keep the increase small.
if (_g1->capacity() <= _g1->max_capacity() / 2) {
threshold *= (double)_g1->capacity() / (double)(_g1->max_capacity() / 2);
threshold = MAX2(threshold, 1.0);
}
// If the last GC time ratio is over the threshold, increment the count of
// times it has been exceeded, and add this ratio to the sum of exceeded
// ratios.
if (last_gc_overhead > threshold) {
_ratio_over_threshold_count++;
_ratio_over_threshold_sum += last_gc_overhead;
}
// Check if we've had enough GC time ratio checks that were over the
// threshold to trigger an expansion. We'll also expand if we've
// reached the end of the history buffer and the average of all entries
// is still over the threshold. This indicates a smaller number of GCs were
// long enough to make the average exceed the threshold.
bool filled_history_buffer = _pauses_since_start == NumPrevPausesForHeuristics;
if ((_ratio_over_threshold_count == MinOverThresholdForGrowth) ||
(filled_history_buffer && (recent_gc_overhead > threshold))) {
size_t min_expand_bytes = HeapRegion::GrainBytes;
size_t reserved_bytes = _g1->max_capacity();
size_t committed_bytes = _g1->capacity();
size_t uncommitted_bytes = reserved_bytes - committed_bytes;
size_t expand_bytes_via_pct =
uncommitted_bytes * G1ExpandByPercentOfAvailable / 100;
double scale_factor = 1.0;
// If the current size is less than 1/4 of the Initial heap size, expand
// by half of the delta between the current and Initial sizes. IE, grow
// back quickly.
//
// Otherwise, take the current size, or G1ExpandByPercentOfAvailable % of
// the available expansion space, whichever is smaller, as the base
// expansion size. Then possibly scale this size according to how much the
// threshold has (on average) been exceeded by. If the delta is small
// (less than the StartScaleDownAt value), scale the size down linearly, but
// not by less than MinScaleDownFactor. If the delta is large (greater than
// the StartScaleUpAt value), scale up, but adding no more than MaxScaleUpFactor
// times the base size. The scaling will be linear in the range from
// StartScaleUpAt to (StartScaleUpAt + ScaleUpRange). In other words,
// ScaleUpRange sets the rate of scaling up.
if (committed_bytes < InitialHeapSize / 4) {
expand_bytes = (InitialHeapSize - committed_bytes) / 2;
} else {
double const MinScaleDownFactor = 0.2;
double const MaxScaleUpFactor = 2;
double const StartScaleDownAt = _gc_overhead_perc;
double const StartScaleUpAt = _gc_overhead_perc * 1.5;
double const ScaleUpRange = _gc_overhead_perc * 2.0;
double ratio_delta;
if (filled_history_buffer) {
ratio_delta = recent_gc_overhead - threshold;
} else {
ratio_delta = (_ratio_over_threshold_sum/_ratio_over_threshold_count) - threshold;
}
expand_bytes = MIN2(expand_bytes_via_pct, committed_bytes);
if (ratio_delta < StartScaleDownAt) {
scale_factor = ratio_delta / StartScaleDownAt;
scale_factor = MAX2(scale_factor, MinScaleDownFactor);
} else if (ratio_delta > StartScaleUpAt) {
scale_factor = 1 + ((ratio_delta - StartScaleUpAt) / ScaleUpRange);
scale_factor = MIN2(scale_factor, MaxScaleUpFactor);
}
}
log_debug(gc, ergo, heap)("Attempt heap expansion (recent GC overhead higher than threshold after GC) "
"recent GC overhead: %1.2f %% threshold: %1.2f %% uncommitted: " SIZE_FORMAT "B base expansion amount and scale: " SIZE_FORMAT "B (%1.2f%%)",
recent_gc_overhead, threshold, uncommitted_bytes, expand_bytes, scale_factor * 100);
expand_bytes = static_cast<size_t>(expand_bytes * scale_factor);
// Ensure the expansion size is at least the minimum growth amount
// and at most the remaining uncommitted byte size.
expand_bytes = MAX2(expand_bytes, min_expand_bytes);
expand_bytes = MIN2(expand_bytes, uncommitted_bytes);
clear_ratio_check_data();
} else {
// An expansion was not triggered. If we've started counting, increment
// the number of checks we've made in the current window. If we've
// reached the end of the window without resizing, clear the counters to
// start again the next time we see a ratio above the threshold.
if (_ratio_over_threshold_count > 0) {
_pauses_since_start++;
if (_pauses_since_start > NumPrevPausesForHeuristics) {
clear_ratio_check_data();
}
}
}
return expand_bytes;
}
void G1CollectorPolicy::print_yg_surv_rate_info() const {
#ifndef PRODUCT
_short_lived_surv_rate_group->print_surv_rate_summary();
// add this call for any other surv rate groups
#endif // PRODUCT
}
bool G1CollectorPolicy::is_young_list_full() const {
uint young_list_length = _g1->young_list()->length();
uint young_list_target_length = _young_list_target_length;
return young_list_length >= young_list_target_length;
}
bool G1CollectorPolicy::can_expand_young_list() const {
uint young_list_length = _g1->young_list()->length();
uint young_list_max_length = _young_list_max_length;
return young_list_length < young_list_max_length;
}
bool G1CollectorPolicy::adaptive_young_list_length() const {
return _young_gen_sizer->adaptive_young_list_length();
}
void G1CollectorPolicy::update_max_gc_locker_expansion() {
uint expansion_region_num = 0;
if (GCLockerEdenExpansionPercent > 0) {
double perc = (double) GCLockerEdenExpansionPercent / 100.0;
double expansion_region_num_d = perc * (double) _young_list_target_length;
// We use ceiling so that if expansion_region_num_d is > 0.0 (but
// less than 1.0) we'll get 1.
expansion_region_num = (uint) ceil(expansion_region_num_d);
} else {
assert(expansion_region_num == 0, "sanity");
}
_young_list_max_length = _young_list_target_length + expansion_region_num;
assert(_young_list_target_length <= _young_list_max_length, "post-condition");
}
// Calculates survivor space parameters.
void G1CollectorPolicy::update_survivors_policy() {
double max_survivor_regions_d =
(double) _young_list_target_length / (double) SurvivorRatio;
// We use ceiling so that if max_survivor_regions_d is > 0.0 (but
// smaller than 1.0) we'll get 1.
_max_survivor_regions = (uint) ceil(max_survivor_regions_d);
_tenuring_threshold = _survivors_age_table.compute_tenuring_threshold(
HeapRegion::GrainWords * _max_survivor_regions, counters());
}
bool G1CollectorPolicy::force_initial_mark_if_outside_cycle(GCCause::Cause gc_cause) {
// We actually check whether we are marking here and not if we are in a
// reclamation phase. This means that we will schedule a concurrent mark
// even while we are still in the process of reclaiming memory.
bool during_cycle = _g1->concurrent_mark()->cmThread()->during_cycle();
if (!during_cycle) {
log_debug(gc, ergo)("Request concurrent cycle initiation (requested by GC cause). GC cause: %s", GCCause::to_string(gc_cause));
collector_state()->set_initiate_conc_mark_if_possible(true);
return true;
} else {
log_debug(gc, ergo)("Do not request concurrent cycle initiation (concurrent cycle already in progress). GC cause: %s", GCCause::to_string(gc_cause));
return false;
}
}
void G1CollectorPolicy::initiate_conc_mark() {
collector_state()->set_during_initial_mark_pause(true);
collector_state()->set_initiate_conc_mark_if_possible(false);
}
void G1CollectorPolicy::decide_on_conc_mark_initiation() {
// We are about to decide on whether this pause will be an
// initial-mark pause.
// First, collector_state()->during_initial_mark_pause() should not be already set. We
// will set it here if we have to. However, it should be cleared by
// the end of the pause (it's only set for the duration of an
// initial-mark pause).
assert(!collector_state()->during_initial_mark_pause(), "pre-condition");
if (collector_state()->initiate_conc_mark_if_possible()) {
// We had noticed on a previous pause that the heap occupancy has
// gone over the initiating threshold and we should start a
// concurrent marking cycle. So we might initiate one.
if (!about_to_start_mixed_phase() && collector_state()->gcs_are_young()) {
// Initiate a new initial mark if there is no marking or reclamation going on.
initiate_conc_mark();
log_debug(gc, ergo)("Initiate concurrent cycle (concurrent cycle initiation requested)");
} else if (_g1->is_user_requested_concurrent_full_gc(_g1->gc_cause())) {
// Initiate a user requested initial mark. An initial mark must be young only
// GC, so the collector state must be updated to reflect this.
collector_state()->set_gcs_are_young(true);
collector_state()->set_last_young_gc(false);
abort_time_to_mixed_tracking();
initiate_conc_mark();
log_debug(gc, ergo)("Initiate concurrent cycle (user requested concurrent cycle)");
} else {
// The concurrent marking thread is still finishing up the
// previous cycle. If we start one right now the two cycles
// overlap. In particular, the concurrent marking thread might
// be in the process of clearing the next marking bitmap (which
// we will use for the next cycle if we start one). Starting a
// cycle now will be bad given that parts of the marking
// information might get cleared by the marking thread. And we
// cannot wait for the marking thread to finish the cycle as it
// periodically yields while clearing the next marking bitmap
// and, if it's in a yield point, it's waiting for us to
// finish. So, at this point we will not start a cycle and we'll
// let the concurrent marking thread complete the last one.
log_debug(gc, ergo)("Do not initiate concurrent cycle (concurrent cycle already in progress)");
}
}
}
class ParKnownGarbageHRClosure: public HeapRegionClosure {
G1CollectedHeap* _g1h;
CSetChooserParUpdater _cset_updater;
public:
ParKnownGarbageHRClosure(CollectionSetChooser* hrSorted,
uint chunk_size) :
_g1h(G1CollectedHeap::heap()),
_cset_updater(hrSorted, true /* parallel */, chunk_size) { }
bool doHeapRegion(HeapRegion* r) {
// Do we have any marking information for this region?
if (r->is_marked()) {
// We will skip any region that's currently used as an old GC
// alloc region (we should not consider those for collection
// before we fill them up).
if (_cset_updater.should_add(r) && !_g1h->is_old_gc_alloc_region(r)) {
_cset_updater.add_region(r);
}
}
return false;
}
};
class ParKnownGarbageTask: public AbstractGangTask {
CollectionSetChooser* _hrSorted;
uint _chunk_size;
G1CollectedHeap* _g1;
HeapRegionClaimer _hrclaimer;
public:
ParKnownGarbageTask(CollectionSetChooser* hrSorted, uint chunk_size, uint n_workers) :
AbstractGangTask("ParKnownGarbageTask"),
_hrSorted(hrSorted), _chunk_size(chunk_size),
_g1(G1CollectedHeap::heap()), _hrclaimer(n_workers) {}
void work(uint worker_id) {
ParKnownGarbageHRClosure parKnownGarbageCl(_hrSorted, _chunk_size);
_g1->heap_region_par_iterate(&parKnownGarbageCl, worker_id, &_hrclaimer);
}
};
uint G1CollectorPolicy::calculate_parallel_work_chunk_size(uint n_workers, uint n_regions) const {
assert(n_workers > 0, "Active gc workers should be greater than 0");
const uint overpartition_factor = 4;
const uint min_chunk_size = MAX2(n_regions / n_workers, 1U);
return MAX2(n_regions / (n_workers * overpartition_factor), min_chunk_size);
}
void G1CollectorPolicy::record_concurrent_mark_cleanup_end() {
cset_chooser()->clear();
WorkGang* workers = _g1->workers();
uint n_workers = workers->active_workers();
uint n_regions = _g1->num_regions();
uint chunk_size = calculate_parallel_work_chunk_size(n_workers, n_regions);
cset_chooser()->prepare_for_par_region_addition(n_workers, n_regions, chunk_size);
ParKnownGarbageTask par_known_garbage_task(cset_chooser(), chunk_size, n_workers);
workers->run_task(&par_known_garbage_task);
cset_chooser()->sort_regions();
double end_sec = os::elapsedTime();
double elapsed_time_ms = (end_sec - _mark_cleanup_start_sec) * 1000.0;
_concurrent_mark_cleanup_times_ms->add(elapsed_time_ms);
_prev_collection_pause_end_ms += elapsed_time_ms;
record_pause(Cleanup, _mark_cleanup_start_sec, end_sec);
}
// Add the heap region at the head of the non-incremental collection set
void G1CollectorPolicy::add_old_region_to_cset(HeapRegion* hr) {
assert(_inc_cset_build_state == Active, "Precondition");
assert(hr->is_old(), "the region should be old");
assert(!hr->in_collection_set(), "should not already be in the CSet");
_g1->register_old_region_with_cset(hr);
hr->set_next_in_collection_set(_collection_set);
_collection_set = hr;
_collection_set_bytes_used_before += hr->used();
size_t rs_length = hr->rem_set()->occupied();
_recorded_rs_lengths += rs_length;
_old_cset_region_length += 1;
}
// Initialize the per-collection-set information
void G1CollectorPolicy::start_incremental_cset_building() {
assert(_inc_cset_build_state == Inactive, "Precondition");
_inc_cset_head = NULL;
_inc_cset_tail = NULL;
_inc_cset_bytes_used_before = 0;
_inc_cset_recorded_rs_lengths = 0;
_inc_cset_recorded_rs_lengths_diffs = 0;
_inc_cset_predicted_elapsed_time_ms = 0.0;
_inc_cset_predicted_elapsed_time_ms_diffs = 0.0;
_inc_cset_build_state = Active;
}
void G1CollectorPolicy::finalize_incremental_cset_building() {
assert(_inc_cset_build_state == Active, "Precondition");
assert(SafepointSynchronize::is_at_safepoint(), "should be at a safepoint");
// The two "main" fields, _inc_cset_recorded_rs_lengths and
// _inc_cset_predicted_elapsed_time_ms, are updated by the thread
// that adds a new region to the CSet. Further updates by the
// concurrent refinement thread that samples the young RSet lengths
// are accumulated in the *_diffs fields. Here we add the diffs to
// the "main" fields.
if (_inc_cset_recorded_rs_lengths_diffs >= 0) {
_inc_cset_recorded_rs_lengths += _inc_cset_recorded_rs_lengths_diffs;
} else {
// This is defensive. The diff should in theory be always positive
// as RSets can only grow between GCs. However, given that we
// sample their size concurrently with other threads updating them
// it's possible that we might get the wrong size back, which
// could make the calculations somewhat inaccurate.
size_t diffs = (size_t) (-_inc_cset_recorded_rs_lengths_diffs);
if (_inc_cset_recorded_rs_lengths >= diffs) {
_inc_cset_recorded_rs_lengths -= diffs;
} else {
_inc_cset_recorded_rs_lengths = 0;
}
}
_inc_cset_predicted_elapsed_time_ms +=
_inc_cset_predicted_elapsed_time_ms_diffs;
_inc_cset_recorded_rs_lengths_diffs = 0;
_inc_cset_predicted_elapsed_time_ms_diffs = 0.0;
}
void G1CollectorPolicy::add_to_incremental_cset_info(HeapRegion* hr, size_t rs_length) {
// This routine is used when:
// * adding survivor regions to the incremental cset at the end of an
// evacuation pause,
// * adding the current allocation region to the incremental cset
// when it is retired, and
// * updating existing policy information for a region in the
// incremental cset via young list RSet sampling.
// Therefore this routine may be called at a safepoint by the
// VM thread, or in-between safepoints by mutator threads (when
// retiring the current allocation region) or a concurrent
// refine thread (RSet sampling).
double region_elapsed_time_ms = predict_region_elapsed_time_ms(hr, collector_state()->gcs_are_young());
size_t used_bytes = hr->used();
_inc_cset_recorded_rs_lengths += rs_length;
_inc_cset_predicted_elapsed_time_ms += region_elapsed_time_ms;
_inc_cset_bytes_used_before += used_bytes;
// Cache the values we have added to the aggregated information
// in the heap region in case we have to remove this region from
// the incremental collection set, or it is updated by the
// rset sampling code
hr->set_recorded_rs_length(rs_length);
hr->set_predicted_elapsed_time_ms(region_elapsed_time_ms);
}
void G1CollectorPolicy::update_incremental_cset_info(HeapRegion* hr,
size_t new_rs_length) {
// Update the CSet information that is dependent on the new RS length
assert(hr->is_young(), "Precondition");
assert(!SafepointSynchronize::is_at_safepoint(),
"should not be at a safepoint");
// We could have updated _inc_cset_recorded_rs_lengths and
// _inc_cset_predicted_elapsed_time_ms directly but we'd need to do
// that atomically, as this code is executed by a concurrent
// refinement thread, potentially concurrently with a mutator thread
// allocating a new region and also updating the same fields. To
// avoid the atomic operations we accumulate these updates on two
// separate fields (*_diffs) and we'll just add them to the "main"
// fields at the start of a GC.
ssize_t old_rs_length = (ssize_t) hr->recorded_rs_length();
ssize_t rs_lengths_diff = (ssize_t) new_rs_length - old_rs_length;
_inc_cset_recorded_rs_lengths_diffs += rs_lengths_diff;
double old_elapsed_time_ms = hr->predicted_elapsed_time_ms();
double new_region_elapsed_time_ms = predict_region_elapsed_time_ms(hr, collector_state()->gcs_are_young());
double elapsed_ms_diff = new_region_elapsed_time_ms - old_elapsed_time_ms;
_inc_cset_predicted_elapsed_time_ms_diffs += elapsed_ms_diff;
hr->set_recorded_rs_length(new_rs_length);
hr->set_predicted_elapsed_time_ms(new_region_elapsed_time_ms);
}
void G1CollectorPolicy::add_region_to_incremental_cset_common(HeapRegion* hr) {
assert(hr->is_young(), "invariant");
assert(hr->young_index_in_cset() > -1, "should have already been set");
assert(_inc_cset_build_state == Active, "Precondition");
// We need to clear and set the cached recorded/cached collection set
// information in the heap region here (before the region gets added
// to the collection set). An individual heap region's cached values
// are calculated, aggregated with the policy collection set info,
// and cached in the heap region here (initially) and (subsequently)
// by the Young List sampling code.
size_t rs_length = hr->rem_set()->occupied();
add_to_incremental_cset_info(hr, rs_length);
assert(!hr->in_collection_set(), "invariant");
_g1->register_young_region_with_cset(hr);
assert(hr->next_in_collection_set() == NULL, "invariant");
}
// Add the region at the RHS of the incremental cset
void G1CollectorPolicy::add_region_to_incremental_cset_rhs(HeapRegion* hr) {
// We should only ever be appending survivors at the end of a pause
assert(hr->is_survivor(), "Logic");
// Do the 'common' stuff
add_region_to_incremental_cset_common(hr);
// Now add the region at the right hand side
if (_inc_cset_tail == NULL) {
assert(_inc_cset_head == NULL, "invariant");
_inc_cset_head = hr;
} else {
_inc_cset_tail->set_next_in_collection_set(hr);
}
_inc_cset_tail = hr;
}
// Add the region to the LHS of the incremental cset
void G1CollectorPolicy::add_region_to_incremental_cset_lhs(HeapRegion* hr) {
// Survivors should be added to the RHS at the end of a pause
assert(hr->is_eden(), "Logic");
// Do the 'common' stuff
add_region_to_incremental_cset_common(hr);
// Add the region at the left hand side
hr->set_next_in_collection_set(_inc_cset_head);
if (_inc_cset_head == NULL) {
assert(_inc_cset_tail == NULL, "Invariant");
_inc_cset_tail = hr;
}
_inc_cset_head = hr;
}
#ifndef PRODUCT
void G1CollectorPolicy::print_collection_set(HeapRegion* list_head, outputStream* st) {
assert(list_head == inc_cset_head() || list_head == collection_set(), "must be");
st->print_cr("\nCollection_set:");
HeapRegion* csr = list_head;
while (csr != NULL) {
HeapRegion* next = csr->next_in_collection_set();
assert(csr->in_collection_set(), "bad CS");
st->print_cr(" " HR_FORMAT ", P: " PTR_FORMAT "N: " PTR_FORMAT ", age: %4d",
HR_FORMAT_PARAMS(csr),
p2i(csr->prev_top_at_mark_start()), p2i(csr->next_top_at_mark_start()),
csr->age_in_surv_rate_group_cond());
csr = next;
}
}
#endif // !PRODUCT
double G1CollectorPolicy::reclaimable_bytes_perc(size_t reclaimable_bytes) const {
// Returns the given amount of reclaimable bytes (that represents
// the amount of reclaimable space still to be collected) as a
// percentage of the current heap capacity.
size_t capacity_bytes = _g1->capacity();
return (double) reclaimable_bytes * 100.0 / (double) capacity_bytes;
}
void G1CollectorPolicy::maybe_start_marking() {
if (need_to_start_conc_mark("end of GC")) {
// Note: this might have already been set, if during the last
// pause we decided to start a cycle but at the beginning of
// this pause we decided to postpone it. That's OK.
collector_state()->set_initiate_conc_mark_if_possible(true);
}
}
G1CollectorPolicy::PauseKind G1CollectorPolicy::young_gc_pause_kind() const {
assert(!collector_state()->full_collection(), "must be");
if (collector_state()->during_initial_mark_pause()) {
assert(collector_state()->last_gc_was_young(), "must be");
assert(!collector_state()->last_young_gc(), "must be");
return InitialMarkGC;
} else if (collector_state()->last_young_gc()) {
assert(!collector_state()->during_initial_mark_pause(), "must be");
assert(collector_state()->last_gc_was_young(), "must be");
return LastYoungGC;
} else if (!collector_state()->last_gc_was_young()) {
assert(!collector_state()->during_initial_mark_pause(), "must be");
assert(!collector_state()->last_young_gc(), "must be");
return MixedGC;
} else {
assert(collector_state()->last_gc_was_young(), "must be");
assert(!collector_state()->during_initial_mark_pause(), "must be");
assert(!collector_state()->last_young_gc(), "must be");
return YoungOnlyGC;
}
}
void G1CollectorPolicy::record_pause(PauseKind kind, double start, double end) {
// Manage the MMU tracker. For some reason it ignores Full GCs.
if (kind != FullGC) {
_mmu_tracker->add_pause(start, end);
}
// Manage the mutator time tracking from initial mark to first mixed gc.
switch (kind) {
case FullGC:
abort_time_to_mixed_tracking();
break;
case Cleanup:
case Remark:
case YoungOnlyGC:
case LastYoungGC:
_initial_mark_to_mixed.add_pause(end - start);
break;
case InitialMarkGC:
_initial_mark_to_mixed.record_initial_mark_end(end);
break;
case MixedGC:
_initial_mark_to_mixed.record_mixed_gc_start(start);
break;
default:
ShouldNotReachHere();
}
}
void G1CollectorPolicy::abort_time_to_mixed_tracking() {
_initial_mark_to_mixed.reset();
}
bool G1CollectorPolicy::next_gc_should_be_mixed(const char* true_action_str,
const char* false_action_str) const {
if (cset_chooser()->is_empty()) {
log_debug(gc, ergo)("%s (candidate old regions not available)", false_action_str);
return false;
}
// Is the amount of uncollected reclaimable space above G1HeapWastePercent?
size_t reclaimable_bytes = cset_chooser()->remaining_reclaimable_bytes();
double reclaimable_perc = reclaimable_bytes_perc(reclaimable_bytes);
double threshold = (double) G1HeapWastePercent;
if (reclaimable_perc <= threshold) {
log_debug(gc, ergo)("%s (reclaimable percentage not over threshold). candidate old regions: %u reclaimable: " SIZE_FORMAT " (%1.2f) threshold: " UINTX_FORMAT,
false_action_str, cset_chooser()->remaining_regions(), reclaimable_bytes, reclaimable_perc, G1HeapWastePercent);
return false;
}
log_debug(gc, ergo)("%s (candidate old regions available). candidate old regions: %u reclaimable: " SIZE_FORMAT " (%1.2f) threshold: " UINTX_FORMAT,
true_action_str, cset_chooser()->remaining_regions(), reclaimable_bytes, reclaimable_perc, G1HeapWastePercent);
return true;
}
uint G1CollectorPolicy::calc_min_old_cset_length() const {
// The min old CSet region bound is based on the maximum desired
// number of mixed GCs after a cycle. I.e., even if some old regions
// look expensive, we should add them to the CSet anyway to make
// sure we go through the available old regions in no more than the
// maximum desired number of mixed GCs.
//
// The calculation is based on the number of marked regions we added
// to the CSet chooser in the first place, not how many remain, so
// that the result is the same during all mixed GCs that follow a cycle.
const size_t region_num = (size_t) cset_chooser()->length();
const size_t gc_num = (size_t) MAX2(G1MixedGCCountTarget, (uintx) 1);
size_t result = region_num / gc_num;
// emulate ceiling
if (result * gc_num < region_num) {
result += 1;
}
return (uint) result;
}
uint G1CollectorPolicy::calc_max_old_cset_length() const {
// The max old CSet region bound is based on the threshold expressed
// as a percentage of the heap size. I.e., it should bound the
// number of old regions added to the CSet irrespective of how many
// of them are available.
const G1CollectedHeap* g1h = G1CollectedHeap::heap();
const size_t region_num = g1h->num_regions();
const size_t perc = (size_t) G1OldCSetRegionThresholdPercent;
size_t result = region_num * perc / 100;
// emulate ceiling
if (100 * result < region_num * perc) {
result += 1;
}
return (uint) result;
}
double G1CollectorPolicy::finalize_young_cset_part(double target_pause_time_ms) {
double young_start_time_sec = os::elapsedTime();
YoungList* young_list = _g1->young_list();
finalize_incremental_cset_building();
guarantee(target_pause_time_ms > 0.0,
"target_pause_time_ms = %1.6lf should be positive", target_pause_time_ms);
guarantee(_collection_set == NULL, "Precondition");
double base_time_ms = predict_base_elapsed_time_ms(_pending_cards);
double time_remaining_ms = MAX2(target_pause_time_ms - base_time_ms, 0.0);
log_trace(gc, ergo, cset)("Start choosing CSet. pending cards: " SIZE_FORMAT " predicted base time: %1.2fms remaining time: %1.2fms target pause time: %1.2fms",
_pending_cards, base_time_ms, time_remaining_ms, target_pause_time_ms);
collector_state()->set_last_gc_was_young(collector_state()->gcs_are_young());
// The young list is laid with the survivor regions from the previous
// pause are appended to the RHS of the young list, i.e.
// [Newly Young Regions ++ Survivors from last pause].
uint survivor_region_length = young_list->survivor_length();
uint eden_region_length = young_list->eden_length();
init_cset_region_lengths(eden_region_length, survivor_region_length);
HeapRegion* hr = young_list->first_survivor_region();
while (hr != NULL) {
assert(hr->is_survivor(), "badly formed young list");
// There is a convention that all the young regions in the CSet
// are tagged as "eden", so we do this for the survivors here. We
// use the special set_eden_pre_gc() as it doesn't check that the
// region is free (which is not the case here).
hr->set_eden_pre_gc();
hr = hr->get_next_young_region();
}
// Clear the fields that point to the survivor list - they are all young now.
young_list->clear_survivors();
_collection_set = _inc_cset_head;
_collection_set_bytes_used_before = _inc_cset_bytes_used_before;
time_remaining_ms = MAX2(time_remaining_ms - _inc_cset_predicted_elapsed_time_ms, 0.0);
log_trace(gc, ergo, cset)("Add young regions to CSet. eden: %u regions, survivors: %u regions, predicted young region time: %1.2fms, target pause time: %1.2fms",
eden_region_length, survivor_region_length, _inc_cset_predicted_elapsed_time_ms, target_pause_time_ms);
// The number of recorded young regions is the incremental
// collection set's current size
set_recorded_rs_lengths(_inc_cset_recorded_rs_lengths);
double young_end_time_sec = os::elapsedTime();
phase_times()->record_young_cset_choice_time_ms((young_end_time_sec - young_start_time_sec) * 1000.0);
return time_remaining_ms;
}
void G1CollectorPolicy::finalize_old_cset_part(double time_remaining_ms) {
double non_young_start_time_sec = os::elapsedTime();
double predicted_old_time_ms = 0.0;
if (!collector_state()->gcs_are_young()) {
cset_chooser()->verify();
const uint min_old_cset_length = calc_min_old_cset_length();
const uint max_old_cset_length = calc_max_old_cset_length();
uint expensive_region_num = 0;
bool check_time_remaining = adaptive_young_list_length();
HeapRegion* hr = cset_chooser()->peek();
while (hr != NULL) {
if (old_cset_region_length() >= max_old_cset_length) {
// Added maximum number of old regions to the CSet.
log_debug(gc, ergo, cset)("Finish adding old regions to CSet (old CSet region num reached max). old %u regions, max %u regions",
old_cset_region_length(), max_old_cset_length);
break;
}
// Stop adding regions if the remaining reclaimable space is
// not above G1HeapWastePercent.
size_t reclaimable_bytes = cset_chooser()->remaining_reclaimable_bytes();
double reclaimable_perc = reclaimable_bytes_perc(reclaimable_bytes);
double threshold = (double) G1HeapWastePercent;
if (reclaimable_perc <= threshold) {
// We've added enough old regions that the amount of uncollected
// reclaimable space is at or below the waste threshold. Stop
// adding old regions to the CSet.
log_debug(gc, ergo, cset)("Finish adding old regions to CSet (reclaimable percentage not over threshold). "
"old %u regions, max %u regions, reclaimable: " SIZE_FORMAT "B (%1.2f%%) threshold: " UINTX_FORMAT "%%",
old_cset_region_length(), max_old_cset_length, reclaimable_bytes, reclaimable_perc, G1HeapWastePercent);
break;
}
double predicted_time_ms = predict_region_elapsed_time_ms(hr, collector_state()->gcs_are_young());
if (check_time_remaining) {
if (predicted_time_ms > time_remaining_ms) {
// Too expensive for the current CSet.
if (old_cset_region_length() >= min_old_cset_length) {
// We have added the minimum number of old regions to the CSet,
// we are done with this CSet.
log_debug(gc, ergo, cset)("Finish adding old regions to CSet (predicted time is too high). "
"predicted time: %1.2fms, remaining time: %1.2fms old %u regions, min %u regions",
predicted_time_ms, time_remaining_ms, old_cset_region_length(), min_old_cset_length);
break;
}
// We'll add it anyway given that we haven't reached the
// minimum number of old regions.
expensive_region_num += 1;
}
} else {
if (old_cset_region_length() >= min_old_cset_length) {
// In the non-auto-tuning case, we'll finish adding regions
// to the CSet if we reach the minimum.
log_debug(gc, ergo, cset)("Finish adding old regions to CSet (old CSet region num reached min). old %u regions, min %u regions",
old_cset_region_length(), min_old_cset_length);
break;
}
}
// We will add this region to the CSet.
time_remaining_ms = MAX2(time_remaining_ms - predicted_time_ms, 0.0);
predicted_old_time_ms += predicted_time_ms;
cset_chooser()->pop(); // already have region via peek()
_g1->old_set_remove(hr);
add_old_region_to_cset(hr);
hr = cset_chooser()->peek();
}
if (hr == NULL) {
log_debug(gc, ergo, cset)("Finish adding old regions to CSet (candidate old regions not available)");
}
if (expensive_region_num > 0) {
// We print the information once here at the end, predicated on
// whether we added any apparently expensive regions or not, to
// avoid generating output per region.
log_debug(gc, ergo, cset)("Added expensive regions to CSet (old CSet region num not reached min)."
"old: %u regions, expensive: %u regions, min: %u regions, remaining time: %1.2fms",
old_cset_region_length(), expensive_region_num, min_old_cset_length, time_remaining_ms);
}
cset_chooser()->verify();
}
stop_incremental_cset_building();
log_debug(gc, ergo, cset)("Finish choosing CSet. old: %u regions, predicted old region time: %1.2fms, time remaining: %1.2f",
old_cset_region_length(), predicted_old_time_ms, time_remaining_ms);
double non_young_end_time_sec = os::elapsedTime();
phase_times()->record_non_young_cset_choice_time_ms((non_young_end_time_sec - non_young_start_time_sec) * 1000.0);
}