a8e3eabb6d
Reviewed-by: sjohanss, pliden
1100 lines
45 KiB
C++
1100 lines
45 KiB
C++
/*
|
|
* Copyright (c) 2002, 2020, 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/parallel/parallelScavengeHeap.hpp"
|
|
#include "gc/parallel/psAdaptiveSizePolicy.hpp"
|
|
#include "gc/parallel/psGCAdaptivePolicyCounters.hpp"
|
|
#include "gc/parallel/psScavenge.hpp"
|
|
#include "gc/shared/gcCause.hpp"
|
|
#include "gc/shared/gcUtil.inline.hpp"
|
|
#include "gc/shared/gcPolicyCounters.hpp"
|
|
#include "logging/log.hpp"
|
|
#include "runtime/timer.hpp"
|
|
#include "utilities/align.hpp"
|
|
|
|
#include <math.h>
|
|
|
|
PSAdaptiveSizePolicy::PSAdaptiveSizePolicy(size_t init_eden_size,
|
|
size_t init_promo_size,
|
|
size_t init_survivor_size,
|
|
size_t space_alignment,
|
|
double gc_pause_goal_sec,
|
|
double gc_minor_pause_goal_sec,
|
|
uint gc_cost_ratio) :
|
|
AdaptiveSizePolicy(init_eden_size,
|
|
init_promo_size,
|
|
init_survivor_size,
|
|
gc_pause_goal_sec,
|
|
gc_cost_ratio),
|
|
_avg_major_pause(new AdaptivePaddedAverage(AdaptiveTimeWeight, PausePadding)),
|
|
_avg_base_footprint(new AdaptiveWeightedAverage(AdaptiveSizePolicyWeight)),
|
|
_gc_stats(),
|
|
_collection_cost_margin_fraction(AdaptiveSizePolicyCollectionCostMargin / 100.0),
|
|
_major_pause_old_estimator(new LinearLeastSquareFit(AdaptiveSizePolicyWeight)),
|
|
_major_pause_young_estimator(new LinearLeastSquareFit(AdaptiveSizePolicyWeight)),
|
|
_latest_major_mutator_interval_seconds(0),
|
|
_space_alignment(space_alignment),
|
|
_gc_minor_pause_goal_sec(gc_minor_pause_goal_sec),
|
|
_live_at_last_full_gc(init_promo_size),
|
|
_change_old_gen_for_min_pauses(0),
|
|
_change_young_gen_for_maj_pauses(0),
|
|
_young_gen_size_increment_supplement(YoungGenerationSizeSupplement),
|
|
_old_gen_size_increment_supplement(TenuredGenerationSizeSupplement)
|
|
{
|
|
// Start the timers
|
|
_major_timer.start();
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::calculate_free_based_on_live(size_t live, uintx ratio_as_percentage) {
|
|
// We want to calculate how much free memory there can be based on the
|
|
// amount of live data currently in the old gen. Using the formula:
|
|
// ratio * (free + live) = free
|
|
// Some equation solving later we get:
|
|
// free = (live * ratio) / (1 - ratio)
|
|
|
|
const double ratio = ratio_as_percentage / 100.0;
|
|
const double ratio_inverse = 1.0 - ratio;
|
|
const double tmp = live * ratio;
|
|
size_t free = (size_t)(tmp / ratio_inverse);
|
|
|
|
return free;
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::calculated_old_free_size_in_bytes() const {
|
|
size_t free_size = (size_t)(_promo_size + avg_promoted()->padded_average());
|
|
size_t live = ParallelScavengeHeap::heap()->old_gen()->used_in_bytes();
|
|
|
|
if (MinHeapFreeRatio != 0) {
|
|
size_t min_free = calculate_free_based_on_live(live, MinHeapFreeRatio);
|
|
free_size = MAX2(free_size, min_free);
|
|
}
|
|
|
|
if (MaxHeapFreeRatio != 100) {
|
|
size_t max_free = calculate_free_based_on_live(live, MaxHeapFreeRatio);
|
|
free_size = MIN2(max_free, free_size);
|
|
}
|
|
|
|
return free_size;
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::major_collection_begin() {
|
|
// Update the interval time
|
|
_major_timer.stop();
|
|
// Save most recent collection time
|
|
_latest_major_mutator_interval_seconds = _major_timer.seconds();
|
|
_major_timer.reset();
|
|
_major_timer.start();
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::update_minor_pause_old_estimator(
|
|
double minor_pause_in_ms) {
|
|
double promo_size_in_mbytes = ((double)_promo_size)/((double)M);
|
|
_minor_pause_old_estimator->update(promo_size_in_mbytes,
|
|
minor_pause_in_ms);
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::major_collection_end(size_t amount_live,
|
|
GCCause::Cause gc_cause) {
|
|
// Update the pause time.
|
|
_major_timer.stop();
|
|
|
|
if (should_update_promo_stats(gc_cause)) {
|
|
double major_pause_in_seconds = _major_timer.seconds();
|
|
double major_pause_in_ms = major_pause_in_seconds * MILLIUNITS;
|
|
|
|
// Sample for performance counter
|
|
_avg_major_pause->sample(major_pause_in_seconds);
|
|
|
|
// Cost of collection (unit-less)
|
|
double collection_cost = 0.0;
|
|
if ((_latest_major_mutator_interval_seconds > 0.0) &&
|
|
(major_pause_in_seconds > 0.0)) {
|
|
double interval_in_seconds =
|
|
_latest_major_mutator_interval_seconds + major_pause_in_seconds;
|
|
collection_cost =
|
|
major_pause_in_seconds / interval_in_seconds;
|
|
avg_major_gc_cost()->sample(collection_cost);
|
|
|
|
// Sample for performance counter
|
|
_avg_major_interval->sample(interval_in_seconds);
|
|
}
|
|
|
|
// Calculate variables used to estimate pause time vs. gen sizes
|
|
double eden_size_in_mbytes = ((double)_eden_size)/((double)M);
|
|
double promo_size_in_mbytes = ((double)_promo_size)/((double)M);
|
|
_major_pause_old_estimator->update(promo_size_in_mbytes,
|
|
major_pause_in_ms);
|
|
_major_pause_young_estimator->update(eden_size_in_mbytes,
|
|
major_pause_in_ms);
|
|
|
|
log_trace(gc, ergo)("psAdaptiveSizePolicy::major_collection_end: major gc cost: %f average: %f",
|
|
collection_cost,avg_major_gc_cost()->average());
|
|
log_trace(gc, ergo)(" major pause: %f major period %f",
|
|
major_pause_in_ms, _latest_major_mutator_interval_seconds * MILLIUNITS);
|
|
|
|
// Calculate variable used to estimate collection cost vs. gen sizes
|
|
assert(collection_cost >= 0.0, "Expected to be non-negative");
|
|
_major_collection_estimator->update(promo_size_in_mbytes,
|
|
collection_cost);
|
|
}
|
|
|
|
// Update the amount live at the end of a full GC
|
|
_live_at_last_full_gc = amount_live;
|
|
|
|
// Interval times use this timer to measure the interval that
|
|
// the mutator runs. Reset after the GC pause has been measured.
|
|
_major_timer.reset();
|
|
_major_timer.start();
|
|
}
|
|
|
|
// If the remaining free space in the old generation is less that
|
|
// that expected to be needed by the next collection, do a full
|
|
// collection now.
|
|
bool PSAdaptiveSizePolicy::should_full_GC(size_t old_free_in_bytes) {
|
|
|
|
// A similar test is done in the scavenge's should_attempt_scavenge(). If
|
|
// this is changed, decide if that test should also be changed.
|
|
bool result = padded_average_promoted_in_bytes() > (float) old_free_in_bytes;
|
|
log_trace(gc, ergo)("%s after scavenge average_promoted " SIZE_FORMAT " padded_average_promoted " SIZE_FORMAT " free in old gen " SIZE_FORMAT,
|
|
result ? "Full" : "No full",
|
|
(size_t) average_promoted_in_bytes(),
|
|
(size_t) padded_average_promoted_in_bytes(),
|
|
old_free_in_bytes);
|
|
return result;
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::clear_generation_free_space_flags() {
|
|
|
|
AdaptiveSizePolicy::clear_generation_free_space_flags();
|
|
|
|
set_change_old_gen_for_min_pauses(0);
|
|
|
|
set_change_young_gen_for_maj_pauses(0);
|
|
}
|
|
|
|
// If this is not a full GC, only test and modify the young generation.
|
|
|
|
void PSAdaptiveSizePolicy::compute_generations_free_space(
|
|
size_t young_live,
|
|
size_t eden_live,
|
|
size_t old_live,
|
|
size_t cur_eden,
|
|
size_t max_old_gen_size,
|
|
size_t max_eden_size,
|
|
bool is_full_gc) {
|
|
compute_eden_space_size(young_live,
|
|
eden_live,
|
|
cur_eden,
|
|
max_eden_size,
|
|
is_full_gc);
|
|
|
|
compute_old_gen_free_space(old_live,
|
|
cur_eden,
|
|
max_old_gen_size,
|
|
is_full_gc);
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::compute_eden_space_size(
|
|
size_t young_live,
|
|
size_t eden_live,
|
|
size_t cur_eden,
|
|
size_t max_eden_size,
|
|
bool is_full_gc) {
|
|
|
|
// Update statistics
|
|
// Time statistics are updated as we go, update footprint stats here
|
|
_avg_base_footprint->sample(BaseFootPrintEstimate);
|
|
avg_young_live()->sample(young_live);
|
|
avg_eden_live()->sample(eden_live);
|
|
|
|
// This code used to return if the policy was not ready , i.e.,
|
|
// policy_is_ready() returning false. The intent was that
|
|
// decisions below needed major collection times and so could
|
|
// not be made before two major collections. A consequence was
|
|
// adjustments to the young generation were not done until after
|
|
// two major collections even if the minor collections times
|
|
// exceeded the requested goals. Now let the young generation
|
|
// adjust for the minor collection times. Major collection times
|
|
// will be zero for the first collection and will naturally be
|
|
// ignored. Tenured generation adjustments are only made at the
|
|
// full collections so until the second major collection has
|
|
// been reached, no tenured generation adjustments will be made.
|
|
|
|
// Until we know better, desired promotion size uses the last calculation
|
|
size_t desired_promo_size = _promo_size;
|
|
|
|
// Start eden at the current value. The desired value that is stored
|
|
// in _eden_size is not bounded by constraints of the heap and can
|
|
// run away.
|
|
//
|
|
// As expected setting desired_eden_size to the current
|
|
// value of desired_eden_size as a starting point
|
|
// caused desired_eden_size to grow way too large and caused
|
|
// an overflow down stream. It may have improved performance in
|
|
// some case but is dangerous.
|
|
size_t desired_eden_size = cur_eden;
|
|
|
|
// Cache some values. There's a bit of work getting these, so
|
|
// we might save a little time.
|
|
const double major_cost = major_gc_cost();
|
|
const double minor_cost = minor_gc_cost();
|
|
|
|
// This method sets the desired eden size. That plus the
|
|
// desired survivor space sizes sets the desired young generation
|
|
// size. This methods does not know what the desired survivor
|
|
// size is but expects that other policy will attempt to make
|
|
// the survivor sizes compatible with the live data in the
|
|
// young generation. This limit is an estimate of the space left
|
|
// in the young generation after the survivor spaces have been
|
|
// subtracted out.
|
|
size_t eden_limit = max_eden_size;
|
|
|
|
const double gc_cost_limit = GCTimeLimit / 100.0;
|
|
|
|
// Which way should we go?
|
|
// if pause requirement is not met
|
|
// adjust size of any generation with average paus exceeding
|
|
// the pause limit. Adjust one pause at a time (the larger)
|
|
// and only make adjustments for the major pause at full collections.
|
|
// else if throughput requirement not met
|
|
// adjust the size of the generation with larger gc time. Only
|
|
// adjust one generation at a time.
|
|
// else
|
|
// adjust down the total heap size. Adjust down the larger of the
|
|
// generations.
|
|
|
|
// Add some checks for a threshold for a change. For example,
|
|
// a change less than the necessary alignment is probably not worth
|
|
// attempting.
|
|
|
|
|
|
if ((_avg_minor_pause->padded_average() > gc_pause_goal_sec()) ||
|
|
(_avg_major_pause->padded_average() > gc_pause_goal_sec())) {
|
|
//
|
|
// Check pauses
|
|
//
|
|
// Make changes only to affect one of the pauses (the larger)
|
|
// at a time.
|
|
adjust_eden_for_pause_time(is_full_gc, &desired_promo_size, &desired_eden_size);
|
|
|
|
} else if (_avg_minor_pause->padded_average() > gc_minor_pause_goal_sec()) {
|
|
// Adjust only for the minor pause time goal
|
|
adjust_eden_for_minor_pause_time(is_full_gc, &desired_eden_size);
|
|
|
|
} else if(adjusted_mutator_cost() < _throughput_goal) {
|
|
// This branch used to require that (mutator_cost() > 0.0 in 1.4.2.
|
|
// This sometimes resulted in skipping to the minimize footprint
|
|
// code. Change this to try and reduce GC time if mutator time is
|
|
// negative for whatever reason. Or for future consideration,
|
|
// bail out of the code if mutator time is negative.
|
|
//
|
|
// Throughput
|
|
//
|
|
assert(major_cost >= 0.0, "major cost is < 0.0");
|
|
assert(minor_cost >= 0.0, "minor cost is < 0.0");
|
|
// Try to reduce the GC times.
|
|
adjust_eden_for_throughput(is_full_gc, &desired_eden_size);
|
|
|
|
} else {
|
|
|
|
// Be conservative about reducing the footprint.
|
|
// Do a minimum number of major collections first.
|
|
// Have reasonable averages for major and minor collections costs.
|
|
if (UseAdaptiveSizePolicyFootprintGoal &&
|
|
young_gen_policy_is_ready() &&
|
|
avg_major_gc_cost()->average() >= 0.0 &&
|
|
avg_minor_gc_cost()->average() >= 0.0) {
|
|
size_t desired_sum = desired_eden_size + desired_promo_size;
|
|
desired_eden_size = adjust_eden_for_footprint(desired_eden_size, desired_sum);
|
|
}
|
|
}
|
|
|
|
// Note we make the same tests as in the code block below; the code
|
|
// seems a little easier to read with the printing in another block.
|
|
if (desired_eden_size > eden_limit) {
|
|
log_debug(gc, ergo)(
|
|
"PSAdaptiveSizePolicy::compute_eden_space_size limits:"
|
|
" desired_eden_size: " SIZE_FORMAT
|
|
" old_eden_size: " SIZE_FORMAT
|
|
" eden_limit: " SIZE_FORMAT
|
|
" cur_eden: " SIZE_FORMAT
|
|
" max_eden_size: " SIZE_FORMAT
|
|
" avg_young_live: " SIZE_FORMAT,
|
|
desired_eden_size, _eden_size, eden_limit, cur_eden,
|
|
max_eden_size, (size_t)avg_young_live()->average());
|
|
}
|
|
if (gc_cost() > gc_cost_limit) {
|
|
log_debug(gc, ergo)(
|
|
"PSAdaptiveSizePolicy::compute_eden_space_size: gc time limit"
|
|
" gc_cost: %f "
|
|
" GCTimeLimit: " UINTX_FORMAT,
|
|
gc_cost(), GCTimeLimit);
|
|
}
|
|
|
|
// Align everything and make a final limit check
|
|
desired_eden_size = align_up(desired_eden_size, _space_alignment);
|
|
desired_eden_size = MAX2(desired_eden_size, _space_alignment);
|
|
|
|
eden_limit = align_down(eden_limit, _space_alignment);
|
|
|
|
// And one last limit check, now that we've aligned things.
|
|
if (desired_eden_size > eden_limit) {
|
|
// If the policy says to get a larger eden but
|
|
// is hitting the limit, don't decrease eden.
|
|
// This can lead to a general drifting down of the
|
|
// eden size. Let the tenuring calculation push more
|
|
// into the old gen.
|
|
desired_eden_size = MAX2(eden_limit, cur_eden);
|
|
}
|
|
|
|
log_debug(gc, ergo)("PSAdaptiveSizePolicy::compute_eden_space_size: costs minor_time: %f major_cost: %f mutator_cost: %f throughput_goal: %f",
|
|
minor_gc_cost(), major_gc_cost(), mutator_cost(), _throughput_goal);
|
|
|
|
log_trace(gc, ergo)("Minor_pause: %f major_pause: %f minor_interval: %f major_interval: %fpause_goal: %f",
|
|
_avg_minor_pause->padded_average(),
|
|
_avg_major_pause->padded_average(),
|
|
_avg_minor_interval->average(),
|
|
_avg_major_interval->average(),
|
|
gc_pause_goal_sec());
|
|
|
|
log_debug(gc, ergo)("Live_space: " SIZE_FORMAT " free_space: " SIZE_FORMAT,
|
|
live_space(), free_space());
|
|
|
|
log_trace(gc, ergo)("Base_footprint: " SIZE_FORMAT " avg_young_live: " SIZE_FORMAT " avg_old_live: " SIZE_FORMAT,
|
|
(size_t)_avg_base_footprint->average(),
|
|
(size_t)avg_young_live()->average(),
|
|
(size_t)avg_old_live()->average());
|
|
|
|
log_debug(gc, ergo)("Old eden_size: " SIZE_FORMAT " desired_eden_size: " SIZE_FORMAT,
|
|
_eden_size, desired_eden_size);
|
|
|
|
set_eden_size(desired_eden_size);
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::compute_old_gen_free_space(
|
|
size_t old_live,
|
|
size_t cur_eden,
|
|
size_t max_old_gen_size,
|
|
bool is_full_gc) {
|
|
|
|
// Update statistics
|
|
// Time statistics are updated as we go, update footprint stats here
|
|
if (is_full_gc) {
|
|
// old_live is only accurate after a full gc
|
|
avg_old_live()->sample(old_live);
|
|
}
|
|
|
|
// This code used to return if the policy was not ready , i.e.,
|
|
// policy_is_ready() returning false. The intent was that
|
|
// decisions below needed major collection times and so could
|
|
// not be made before two major collections. A consequence was
|
|
// adjustments to the young generation were not done until after
|
|
// two major collections even if the minor collections times
|
|
// exceeded the requested goals. Now let the young generation
|
|
// adjust for the minor collection times. Major collection times
|
|
// will be zero for the first collection and will naturally be
|
|
// ignored. Tenured generation adjustments are only made at the
|
|
// full collections so until the second major collection has
|
|
// been reached, no tenured generation adjustments will be made.
|
|
|
|
// Until we know better, desired promotion size uses the last calculation
|
|
size_t desired_promo_size = _promo_size;
|
|
|
|
// Start eden at the current value. The desired value that is stored
|
|
// in _eden_size is not bounded by constraints of the heap and can
|
|
// run away.
|
|
//
|
|
// As expected setting desired_eden_size to the current
|
|
// value of desired_eden_size as a starting point
|
|
// caused desired_eden_size to grow way too large and caused
|
|
// an overflow down stream. It may have improved performance in
|
|
// some case but is dangerous.
|
|
size_t desired_eden_size = cur_eden;
|
|
|
|
// Cache some values. There's a bit of work getting these, so
|
|
// we might save a little time.
|
|
const double major_cost = major_gc_cost();
|
|
const double minor_cost = minor_gc_cost();
|
|
|
|
// Limits on our growth
|
|
size_t promo_limit = (size_t)(max_old_gen_size - avg_old_live()->average());
|
|
|
|
// But don't force a promo size below the current promo size. Otherwise,
|
|
// the promo size will shrink for no good reason.
|
|
promo_limit = MAX2(promo_limit, _promo_size);
|
|
|
|
const double gc_cost_limit = GCTimeLimit/100.0;
|
|
|
|
// Which way should we go?
|
|
// if pause requirement is not met
|
|
// adjust size of any generation with average paus exceeding
|
|
// the pause limit. Adjust one pause at a time (the larger)
|
|
// and only make adjustments for the major pause at full collections.
|
|
// else if throughput requirement not met
|
|
// adjust the size of the generation with larger gc time. Only
|
|
// adjust one generation at a time.
|
|
// else
|
|
// adjust down the total heap size. Adjust down the larger of the
|
|
// generations.
|
|
|
|
// Add some checks for a threshold for a change. For example,
|
|
// a change less than the necessary alignment is probably not worth
|
|
// attempting.
|
|
|
|
if ((_avg_minor_pause->padded_average() > gc_pause_goal_sec()) ||
|
|
(_avg_major_pause->padded_average() > gc_pause_goal_sec())) {
|
|
//
|
|
// Check pauses
|
|
//
|
|
// Make changes only to affect one of the pauses (the larger)
|
|
// at a time.
|
|
if (is_full_gc) {
|
|
set_decide_at_full_gc(decide_at_full_gc_true);
|
|
adjust_promo_for_pause_time(is_full_gc, &desired_promo_size, &desired_eden_size);
|
|
}
|
|
} else if (adjusted_mutator_cost() < _throughput_goal) {
|
|
// This branch used to require that (mutator_cost() > 0.0 in 1.4.2.
|
|
// This sometimes resulted in skipping to the minimize footprint
|
|
// code. Change this to try and reduce GC time if mutator time is
|
|
// negative for whatever reason. Or for future consideration,
|
|
// bail out of the code if mutator time is negative.
|
|
//
|
|
// Throughput
|
|
//
|
|
assert(major_cost >= 0.0, "major cost is < 0.0");
|
|
assert(minor_cost >= 0.0, "minor cost is < 0.0");
|
|
// Try to reduce the GC times.
|
|
if (is_full_gc) {
|
|
set_decide_at_full_gc(decide_at_full_gc_true);
|
|
adjust_promo_for_throughput(is_full_gc, &desired_promo_size);
|
|
}
|
|
} else {
|
|
|
|
// Be conservative about reducing the footprint.
|
|
// Do a minimum number of major collections first.
|
|
// Have reasonable averages for major and minor collections costs.
|
|
if (UseAdaptiveSizePolicyFootprintGoal &&
|
|
young_gen_policy_is_ready() &&
|
|
avg_major_gc_cost()->average() >= 0.0 &&
|
|
avg_minor_gc_cost()->average() >= 0.0) {
|
|
if (is_full_gc) {
|
|
set_decide_at_full_gc(decide_at_full_gc_true);
|
|
size_t desired_sum = desired_eden_size + desired_promo_size;
|
|
desired_promo_size = adjust_promo_for_footprint(desired_promo_size, desired_sum);
|
|
}
|
|
}
|
|
}
|
|
|
|
// Note we make the same tests as in the code block below; the code
|
|
// seems a little easier to read with the printing in another block.
|
|
if (desired_promo_size > promo_limit) {
|
|
// "free_in_old_gen" was the original value for used for promo_limit
|
|
size_t free_in_old_gen = (size_t)(max_old_gen_size - avg_old_live()->average());
|
|
log_debug(gc, ergo)(
|
|
"PSAdaptiveSizePolicy::compute_old_gen_free_space limits:"
|
|
" desired_promo_size: " SIZE_FORMAT
|
|
" promo_limit: " SIZE_FORMAT
|
|
" free_in_old_gen: " SIZE_FORMAT
|
|
" max_old_gen_size: " SIZE_FORMAT
|
|
" avg_old_live: " SIZE_FORMAT,
|
|
desired_promo_size, promo_limit, free_in_old_gen,
|
|
max_old_gen_size, (size_t) avg_old_live()->average());
|
|
}
|
|
if (gc_cost() > gc_cost_limit) {
|
|
log_debug(gc, ergo)(
|
|
"PSAdaptiveSizePolicy::compute_old_gen_free_space: gc time limit"
|
|
" gc_cost: %f "
|
|
" GCTimeLimit: " UINTX_FORMAT,
|
|
gc_cost(), GCTimeLimit);
|
|
}
|
|
|
|
// Align everything and make a final limit check
|
|
desired_promo_size = align_up(desired_promo_size, _space_alignment);
|
|
desired_promo_size = MAX2(desired_promo_size, _space_alignment);
|
|
|
|
promo_limit = align_down(promo_limit, _space_alignment);
|
|
|
|
// And one last limit check, now that we've aligned things.
|
|
desired_promo_size = MIN2(desired_promo_size, promo_limit);
|
|
|
|
// Timing stats
|
|
log_debug(gc, ergo)("PSAdaptiveSizePolicy::compute_old_gen_free_space: costs minor_time: %f major_cost: %f mutator_cost: %f throughput_goal: %f",
|
|
minor_gc_cost(), major_gc_cost(), mutator_cost(), _throughput_goal);
|
|
|
|
log_trace(gc, ergo)("Minor_pause: %f major_pause: %f minor_interval: %f major_interval: %f pause_goal: %f",
|
|
_avg_minor_pause->padded_average(),
|
|
_avg_major_pause->padded_average(),
|
|
_avg_minor_interval->average(),
|
|
_avg_major_interval->average(),
|
|
gc_pause_goal_sec());
|
|
|
|
// Footprint stats
|
|
log_debug(gc, ergo)("Live_space: " SIZE_FORMAT " free_space: " SIZE_FORMAT,
|
|
live_space(), free_space());
|
|
|
|
log_trace(gc, ergo)("Base_footprint: " SIZE_FORMAT " avg_young_live: " SIZE_FORMAT " avg_old_live: " SIZE_FORMAT,
|
|
(size_t)_avg_base_footprint->average(),
|
|
(size_t)avg_young_live()->average(),
|
|
(size_t)avg_old_live()->average());
|
|
|
|
log_debug(gc, ergo)("Old promo_size: " SIZE_FORMAT " desired_promo_size: " SIZE_FORMAT,
|
|
_promo_size, desired_promo_size);
|
|
|
|
set_promo_size(desired_promo_size);
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::decay_supplemental_growth(bool is_full_gc) {
|
|
// Decay the supplemental increment? Decay the supplement growth
|
|
// factor even if it is not used. It is only meant to give a boost
|
|
// to the initial growth and if it is not used, then it was not
|
|
// needed.
|
|
if (is_full_gc) {
|
|
// Don't wait for the threshold value for the major collections. If
|
|
// here, the supplemental growth term was used and should decay.
|
|
if ((_avg_major_pause->count() % TenuredGenerationSizeSupplementDecay)
|
|
== 0) {
|
|
_old_gen_size_increment_supplement =
|
|
_old_gen_size_increment_supplement >> 1;
|
|
}
|
|
} else {
|
|
if ((_avg_minor_pause->count() >= AdaptiveSizePolicyReadyThreshold) &&
|
|
(_avg_minor_pause->count() % YoungGenerationSizeSupplementDecay) == 0) {
|
|
_young_gen_size_increment_supplement =
|
|
_young_gen_size_increment_supplement >> 1;
|
|
}
|
|
}
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::adjust_eden_for_minor_pause_time(bool is_full_gc,
|
|
size_t* desired_eden_size_ptr) {
|
|
|
|
// Adjust the young generation size to reduce pause time of
|
|
// of collections.
|
|
//
|
|
// The AdaptiveSizePolicyInitializingSteps test is not used
|
|
// here. It has not seemed to be needed but perhaps should
|
|
// be added for consistency.
|
|
if (minor_pause_young_estimator()->decrement_will_decrease()) {
|
|
// reduce eden size
|
|
set_change_young_gen_for_min_pauses(
|
|
decrease_young_gen_for_min_pauses_true);
|
|
*desired_eden_size_ptr = *desired_eden_size_ptr -
|
|
eden_decrement_aligned_down(*desired_eden_size_ptr);
|
|
} else {
|
|
// EXPERIMENTAL ADJUSTMENT
|
|
// Only record that the estimator indicated such an action.
|
|
// *desired_eden_size_ptr = *desired_eden_size_ptr + eden_heap_delta;
|
|
set_change_young_gen_for_min_pauses(
|
|
increase_young_gen_for_min_pauses_true);
|
|
}
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::adjust_promo_for_pause_time(bool is_full_gc,
|
|
size_t* desired_promo_size_ptr,
|
|
size_t* desired_eden_size_ptr) {
|
|
|
|
size_t promo_heap_delta = 0;
|
|
// Add some checks for a threshold for a change. For example,
|
|
// a change less than the required alignment is probably not worth
|
|
// attempting.
|
|
|
|
if (_avg_minor_pause->padded_average() <= _avg_major_pause->padded_average() && is_full_gc) {
|
|
// Adjust for the major pause time only at full gc's because the
|
|
// affects of a change can only be seen at full gc's.
|
|
|
|
// Reduce old generation size to reduce pause?
|
|
if (major_pause_old_estimator()->decrement_will_decrease()) {
|
|
// reduce old generation size
|
|
set_change_old_gen_for_maj_pauses(decrease_old_gen_for_maj_pauses_true);
|
|
promo_heap_delta = promo_decrement_aligned_down(*desired_promo_size_ptr);
|
|
*desired_promo_size_ptr = _promo_size - promo_heap_delta;
|
|
} else {
|
|
// EXPERIMENTAL ADJUSTMENT
|
|
// Only record that the estimator indicated such an action.
|
|
// *desired_promo_size_ptr = _promo_size +
|
|
// promo_increment_aligned_up(*desired_promo_size_ptr);
|
|
set_change_old_gen_for_maj_pauses(increase_old_gen_for_maj_pauses_true);
|
|
}
|
|
}
|
|
|
|
log_trace(gc, ergo)(
|
|
"PSAdaptiveSizePolicy::adjust_promo_for_pause_time "
|
|
"adjusting gen sizes for major pause (avg %f goal %f). "
|
|
"desired_promo_size " SIZE_FORMAT " promo delta " SIZE_FORMAT,
|
|
_avg_major_pause->average(), gc_pause_goal_sec(),
|
|
*desired_promo_size_ptr, promo_heap_delta);
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::adjust_eden_for_pause_time(bool is_full_gc,
|
|
size_t* desired_promo_size_ptr,
|
|
size_t* desired_eden_size_ptr) {
|
|
|
|
size_t eden_heap_delta = 0;
|
|
// Add some checks for a threshold for a change. For example,
|
|
// a change less than the required alignment is probably not worth
|
|
// attempting.
|
|
if (_avg_minor_pause->padded_average() > _avg_major_pause->padded_average()) {
|
|
adjust_eden_for_minor_pause_time(is_full_gc, desired_eden_size_ptr);
|
|
}
|
|
log_trace(gc, ergo)(
|
|
"PSAdaptiveSizePolicy::adjust_eden_for_pause_time "
|
|
"adjusting gen sizes for major pause (avg %f goal %f). "
|
|
"desired_eden_size " SIZE_FORMAT " eden delta " SIZE_FORMAT,
|
|
_avg_major_pause->average(), gc_pause_goal_sec(),
|
|
*desired_eden_size_ptr, eden_heap_delta);
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::adjust_promo_for_throughput(bool is_full_gc,
|
|
size_t* desired_promo_size_ptr) {
|
|
|
|
// Add some checks for a threshold for a change. For example,
|
|
// a change less than the required alignment is probably not worth
|
|
// attempting.
|
|
|
|
if ((gc_cost() + mutator_cost()) == 0.0) {
|
|
return;
|
|
}
|
|
|
|
log_trace(gc, ergo)("PSAdaptiveSizePolicy::adjust_promo_for_throughput(is_full: %d, promo: " SIZE_FORMAT "): mutator_cost %f major_gc_cost %f minor_gc_cost %f",
|
|
is_full_gc, *desired_promo_size_ptr, mutator_cost(), major_gc_cost(), minor_gc_cost());
|
|
|
|
// Tenured generation
|
|
if (is_full_gc) {
|
|
// Calculate the change to use for the tenured gen.
|
|
size_t scaled_promo_heap_delta = 0;
|
|
// Can the increment to the generation be scaled?
|
|
if (gc_cost() >= 0.0 && major_gc_cost() >= 0.0) {
|
|
size_t promo_heap_delta =
|
|
promo_increment_with_supplement_aligned_up(*desired_promo_size_ptr);
|
|
double scale_by_ratio = major_gc_cost() / gc_cost();
|
|
scaled_promo_heap_delta =
|
|
(size_t) (scale_by_ratio * (double) promo_heap_delta);
|
|
log_trace(gc, ergo)("Scaled tenured increment: " SIZE_FORMAT " by %f down to " SIZE_FORMAT,
|
|
promo_heap_delta, scale_by_ratio, scaled_promo_heap_delta);
|
|
} else if (major_gc_cost() >= 0.0) {
|
|
// Scaling is not going to work. If the major gc time is the
|
|
// larger, give it a full increment.
|
|
if (major_gc_cost() >= minor_gc_cost()) {
|
|
scaled_promo_heap_delta =
|
|
promo_increment_with_supplement_aligned_up(*desired_promo_size_ptr);
|
|
}
|
|
} else {
|
|
// Don't expect to get here but it's ok if it does
|
|
// in the product build since the delta will be 0
|
|
// and nothing will change.
|
|
assert(false, "Unexpected value for gc costs");
|
|
}
|
|
|
|
switch (AdaptiveSizeThroughPutPolicy) {
|
|
case 1:
|
|
// Early in the run the statistics might not be good. Until
|
|
// a specific number of collections have been, use the heuristic
|
|
// that a larger generation size means lower collection costs.
|
|
if (major_collection_estimator()->increment_will_decrease() ||
|
|
(_old_gen_change_for_major_throughput
|
|
<= AdaptiveSizePolicyInitializingSteps)) {
|
|
// Increase tenured generation size to reduce major collection cost
|
|
if ((*desired_promo_size_ptr + scaled_promo_heap_delta) >
|
|
*desired_promo_size_ptr) {
|
|
*desired_promo_size_ptr = _promo_size + scaled_promo_heap_delta;
|
|
}
|
|
set_change_old_gen_for_throughput(
|
|
increase_old_gen_for_throughput_true);
|
|
_old_gen_change_for_major_throughput++;
|
|
} else {
|
|
// EXPERIMENTAL ADJUSTMENT
|
|
// Record that decreasing the old gen size would decrease
|
|
// the major collection cost but don't do it.
|
|
// *desired_promo_size_ptr = _promo_size -
|
|
// promo_decrement_aligned_down(*desired_promo_size_ptr);
|
|
set_change_old_gen_for_throughput(
|
|
decrease_old_gen_for_throughput_true);
|
|
}
|
|
|
|
break;
|
|
default:
|
|
// Simplest strategy
|
|
if ((*desired_promo_size_ptr + scaled_promo_heap_delta) >
|
|
*desired_promo_size_ptr) {
|
|
*desired_promo_size_ptr = *desired_promo_size_ptr +
|
|
scaled_promo_heap_delta;
|
|
}
|
|
set_change_old_gen_for_throughput(
|
|
increase_old_gen_for_throughput_true);
|
|
_old_gen_change_for_major_throughput++;
|
|
}
|
|
|
|
log_trace(gc, ergo)("Adjusting tenured gen for throughput (avg %f goal %f). desired_promo_size " SIZE_FORMAT " promo_delta " SIZE_FORMAT ,
|
|
mutator_cost(),
|
|
_throughput_goal,
|
|
*desired_promo_size_ptr, scaled_promo_heap_delta);
|
|
}
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::adjust_eden_for_throughput(bool is_full_gc,
|
|
size_t* desired_eden_size_ptr) {
|
|
|
|
// Add some checks for a threshold for a change. For example,
|
|
// a change less than the required alignment is probably not worth
|
|
// attempting.
|
|
|
|
if ((gc_cost() + mutator_cost()) == 0.0) {
|
|
return;
|
|
}
|
|
|
|
log_trace(gc, ergo)("PSAdaptiveSizePolicy::adjust_eden_for_throughput(is_full: %d, cur_eden: " SIZE_FORMAT "): mutator_cost %f major_gc_cost %f minor_gc_cost %f",
|
|
is_full_gc, *desired_eden_size_ptr, mutator_cost(), major_gc_cost(), minor_gc_cost());
|
|
|
|
// Young generation
|
|
size_t scaled_eden_heap_delta = 0;
|
|
// Can the increment to the generation be scaled?
|
|
if (gc_cost() >= 0.0 && minor_gc_cost() >= 0.0) {
|
|
size_t eden_heap_delta =
|
|
eden_increment_with_supplement_aligned_up(*desired_eden_size_ptr);
|
|
double scale_by_ratio = minor_gc_cost() / gc_cost();
|
|
assert(scale_by_ratio <= 1.0 && scale_by_ratio >= 0.0, "Scaling is wrong");
|
|
scaled_eden_heap_delta =
|
|
(size_t) (scale_by_ratio * (double) eden_heap_delta);
|
|
log_trace(gc, ergo)("Scaled eden increment: " SIZE_FORMAT " by %f down to " SIZE_FORMAT,
|
|
eden_heap_delta, scale_by_ratio, scaled_eden_heap_delta);
|
|
} else if (minor_gc_cost() >= 0.0) {
|
|
// Scaling is not going to work. If the minor gc time is the
|
|
// larger, give it a full increment.
|
|
if (minor_gc_cost() > major_gc_cost()) {
|
|
scaled_eden_heap_delta =
|
|
eden_increment_with_supplement_aligned_up(*desired_eden_size_ptr);
|
|
}
|
|
} else {
|
|
// Don't expect to get here but it's ok if it does
|
|
// in the product build since the delta will be 0
|
|
// and nothing will change.
|
|
assert(false, "Unexpected value for gc costs");
|
|
}
|
|
|
|
// Use a heuristic for some number of collections to give
|
|
// the averages time to settle down.
|
|
switch (AdaptiveSizeThroughPutPolicy) {
|
|
case 1:
|
|
if (minor_collection_estimator()->increment_will_decrease() ||
|
|
(_young_gen_change_for_minor_throughput
|
|
<= AdaptiveSizePolicyInitializingSteps)) {
|
|
// Expand young generation size to reduce frequency of
|
|
// of collections.
|
|
if ((*desired_eden_size_ptr + scaled_eden_heap_delta) >
|
|
*desired_eden_size_ptr) {
|
|
*desired_eden_size_ptr =
|
|
*desired_eden_size_ptr + scaled_eden_heap_delta;
|
|
}
|
|
set_change_young_gen_for_throughput(
|
|
increase_young_gen_for_througput_true);
|
|
_young_gen_change_for_minor_throughput++;
|
|
} else {
|
|
// EXPERIMENTAL ADJUSTMENT
|
|
// Record that decreasing the young gen size would decrease
|
|
// the minor collection cost but don't do it.
|
|
// *desired_eden_size_ptr = _eden_size -
|
|
// eden_decrement_aligned_down(*desired_eden_size_ptr);
|
|
set_change_young_gen_for_throughput(
|
|
decrease_young_gen_for_througput_true);
|
|
}
|
|
break;
|
|
default:
|
|
if ((*desired_eden_size_ptr + scaled_eden_heap_delta) >
|
|
*desired_eden_size_ptr) {
|
|
*desired_eden_size_ptr =
|
|
*desired_eden_size_ptr + scaled_eden_heap_delta;
|
|
}
|
|
set_change_young_gen_for_throughput(
|
|
increase_young_gen_for_througput_true);
|
|
_young_gen_change_for_minor_throughput++;
|
|
}
|
|
|
|
log_trace(gc, ergo)("Adjusting eden for throughput (avg %f goal %f). desired_eden_size " SIZE_FORMAT " eden delta " SIZE_FORMAT,
|
|
mutator_cost(), _throughput_goal, *desired_eden_size_ptr, scaled_eden_heap_delta);
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::adjust_promo_for_footprint(
|
|
size_t desired_promo_size, size_t desired_sum) {
|
|
assert(desired_promo_size <= desired_sum, "Inconsistent parameters");
|
|
set_decrease_for_footprint(decrease_old_gen_for_footprint_true);
|
|
|
|
size_t change = promo_decrement(desired_promo_size);
|
|
change = scale_down(change, desired_promo_size, desired_sum);
|
|
|
|
size_t reduced_size = desired_promo_size - change;
|
|
|
|
log_trace(gc, ergo)(
|
|
"AdaptiveSizePolicy::adjust_promo_for_footprint "
|
|
"adjusting tenured gen for footprint. "
|
|
"starting promo size " SIZE_FORMAT
|
|
" reduced promo size " SIZE_FORMAT
|
|
" promo delta " SIZE_FORMAT,
|
|
desired_promo_size, reduced_size, change );
|
|
|
|
assert(reduced_size <= desired_promo_size, "Inconsistent result");
|
|
return reduced_size;
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::adjust_eden_for_footprint(
|
|
size_t desired_eden_size, size_t desired_sum) {
|
|
assert(desired_eden_size <= desired_sum, "Inconsistent parameters");
|
|
set_decrease_for_footprint(decrease_young_gen_for_footprint_true);
|
|
|
|
size_t change = eden_decrement(desired_eden_size);
|
|
change = scale_down(change, desired_eden_size, desired_sum);
|
|
|
|
size_t reduced_size = desired_eden_size - change;
|
|
|
|
log_trace(gc, ergo)(
|
|
"AdaptiveSizePolicy::adjust_eden_for_footprint "
|
|
"adjusting eden for footprint. "
|
|
" starting eden size " SIZE_FORMAT
|
|
" reduced eden size " SIZE_FORMAT
|
|
" eden delta " SIZE_FORMAT,
|
|
desired_eden_size, reduced_size, change);
|
|
|
|
assert(reduced_size <= desired_eden_size, "Inconsistent result");
|
|
return reduced_size;
|
|
}
|
|
|
|
// Scale down "change" by the factor
|
|
// part / total
|
|
// Don't align the results.
|
|
|
|
size_t PSAdaptiveSizePolicy::scale_down(size_t change,
|
|
double part,
|
|
double total) {
|
|
assert(part <= total, "Inconsistent input");
|
|
size_t reduced_change = change;
|
|
if (total > 0) {
|
|
double fraction = part / total;
|
|
reduced_change = (size_t) (fraction * (double) change);
|
|
}
|
|
assert(reduced_change <= change, "Inconsistent result");
|
|
return reduced_change;
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::eden_increment(size_t cur_eden,
|
|
uint percent_change) {
|
|
size_t eden_heap_delta;
|
|
eden_heap_delta = cur_eden / 100 * percent_change;
|
|
return eden_heap_delta;
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::eden_increment(size_t cur_eden) {
|
|
return eden_increment(cur_eden, YoungGenerationSizeIncrement);
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::eden_increment_with_supplement_aligned_up(
|
|
size_t cur_eden) {
|
|
size_t result = eden_increment(cur_eden,
|
|
YoungGenerationSizeIncrement + _young_gen_size_increment_supplement);
|
|
return align_up(result, _space_alignment);
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::eden_decrement_aligned_down(size_t cur_eden) {
|
|
size_t eden_heap_delta = eden_decrement(cur_eden);
|
|
return align_down(eden_heap_delta, _space_alignment);
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::eden_decrement(size_t cur_eden) {
|
|
size_t eden_heap_delta = eden_increment(cur_eden) /
|
|
AdaptiveSizeDecrementScaleFactor;
|
|
return eden_heap_delta;
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::promo_increment(size_t cur_promo,
|
|
uint percent_change) {
|
|
size_t promo_heap_delta;
|
|
promo_heap_delta = cur_promo / 100 * percent_change;
|
|
return promo_heap_delta;
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::promo_increment(size_t cur_promo) {
|
|
return promo_increment(cur_promo, TenuredGenerationSizeIncrement);
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::promo_increment_with_supplement_aligned_up(
|
|
size_t cur_promo) {
|
|
size_t result = promo_increment(cur_promo,
|
|
TenuredGenerationSizeIncrement + _old_gen_size_increment_supplement);
|
|
return align_up(result, _space_alignment);
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::promo_decrement_aligned_down(size_t cur_promo) {
|
|
size_t promo_heap_delta = promo_decrement(cur_promo);
|
|
return align_down(promo_heap_delta, _space_alignment);
|
|
}
|
|
|
|
size_t PSAdaptiveSizePolicy::promo_decrement(size_t cur_promo) {
|
|
size_t promo_heap_delta = promo_increment(cur_promo);
|
|
promo_heap_delta = promo_heap_delta / AdaptiveSizeDecrementScaleFactor;
|
|
return promo_heap_delta;
|
|
}
|
|
|
|
uint PSAdaptiveSizePolicy::compute_survivor_space_size_and_threshold(
|
|
bool is_survivor_overflow,
|
|
uint tenuring_threshold,
|
|
size_t survivor_limit) {
|
|
assert(survivor_limit >= _space_alignment,
|
|
"survivor_limit too small");
|
|
assert(is_aligned(survivor_limit, _space_alignment),
|
|
"survivor_limit not aligned");
|
|
|
|
// This method is called even if the tenuring threshold and survivor
|
|
// spaces are not adjusted so that the averages are sampled above.
|
|
if (!UsePSAdaptiveSurvivorSizePolicy ||
|
|
!young_gen_policy_is_ready()) {
|
|
return tenuring_threshold;
|
|
}
|
|
|
|
// We'll decide whether to increase or decrease the tenuring
|
|
// threshold based partly on the newly computed survivor size
|
|
// (if we hit the maximum limit allowed, we'll always choose to
|
|
// decrement the threshold).
|
|
bool incr_tenuring_threshold = false;
|
|
bool decr_tenuring_threshold = false;
|
|
|
|
set_decrement_tenuring_threshold_for_gc_cost(false);
|
|
set_increment_tenuring_threshold_for_gc_cost(false);
|
|
set_decrement_tenuring_threshold_for_survivor_limit(false);
|
|
|
|
if (!is_survivor_overflow) {
|
|
// Keep running averages on how much survived
|
|
|
|
// We use the tenuring threshold to equalize the cost of major
|
|
// and minor collections.
|
|
// ThresholdTolerance is used to indicate how sensitive the
|
|
// tenuring threshold is to differences in cost between the
|
|
// collection types.
|
|
|
|
// Get the times of interest. This involves a little work, so
|
|
// we cache the values here.
|
|
const double major_cost = major_gc_cost();
|
|
const double minor_cost = minor_gc_cost();
|
|
|
|
if (minor_cost > major_cost * _threshold_tolerance_percent) {
|
|
// Minor times are getting too long; lower the threshold so
|
|
// less survives and more is promoted.
|
|
decr_tenuring_threshold = true;
|
|
set_decrement_tenuring_threshold_for_gc_cost(true);
|
|
} else if (major_cost > minor_cost * _threshold_tolerance_percent) {
|
|
// Major times are too long, so we want less promotion.
|
|
incr_tenuring_threshold = true;
|
|
set_increment_tenuring_threshold_for_gc_cost(true);
|
|
}
|
|
|
|
} else {
|
|
// Survivor space overflow occurred, so promoted and survived are
|
|
// not accurate. We'll make our best guess by combining survived
|
|
// and promoted and count them as survivors.
|
|
//
|
|
// We'll lower the tenuring threshold to see if we can correct
|
|
// things. Also, set the survivor size conservatively. We're
|
|
// trying to avoid many overflows from occurring if defnew size
|
|
// is just too small.
|
|
|
|
decr_tenuring_threshold = true;
|
|
}
|
|
|
|
// The padded average also maintains a deviation from the average;
|
|
// we use this to see how good of an estimate we have of what survived.
|
|
// We're trying to pad the survivor size as little as possible without
|
|
// overflowing the survivor spaces.
|
|
size_t target_size = align_up((size_t)_avg_survived->padded_average(),
|
|
_space_alignment);
|
|
target_size = MAX2(target_size, _space_alignment);
|
|
|
|
if (target_size > survivor_limit) {
|
|
// Target size is bigger than we can handle. Let's also reduce
|
|
// the tenuring threshold.
|
|
target_size = survivor_limit;
|
|
decr_tenuring_threshold = true;
|
|
set_decrement_tenuring_threshold_for_survivor_limit(true);
|
|
}
|
|
|
|
// Finally, increment or decrement the tenuring threshold, as decided above.
|
|
// We test for decrementing first, as we might have hit the target size
|
|
// limit.
|
|
if (decr_tenuring_threshold && !(AlwaysTenure || NeverTenure)) {
|
|
if (tenuring_threshold > 1) {
|
|
tenuring_threshold--;
|
|
}
|
|
} else if (incr_tenuring_threshold && !(AlwaysTenure || NeverTenure)) {
|
|
if (tenuring_threshold < MaxTenuringThreshold) {
|
|
tenuring_threshold++;
|
|
}
|
|
}
|
|
|
|
// We keep a running average of the amount promoted which is used
|
|
// to decide when we should collect the old generation (when
|
|
// the amount of old gen free space is less than what we expect to
|
|
// promote).
|
|
|
|
log_trace(gc, ergo)("avg_survived: %f avg_deviation: %f", _avg_survived->average(), _avg_survived->deviation());
|
|
log_debug(gc, ergo)("avg_survived_padded_avg: %f", _avg_survived->padded_average());
|
|
|
|
log_trace(gc, ergo)("avg_promoted_avg: %f avg_promoted_dev: %f", avg_promoted()->average(), avg_promoted()->deviation());
|
|
log_debug(gc, ergo)("avg_promoted_padded_avg: %f avg_pretenured_padded_avg: %f tenuring_thresh: %d target_size: " SIZE_FORMAT,
|
|
avg_promoted()->padded_average(),
|
|
_avg_pretenured->padded_average(),
|
|
tenuring_threshold, target_size);
|
|
|
|
set_survivor_size(target_size);
|
|
|
|
return tenuring_threshold;
|
|
}
|
|
|
|
void PSAdaptiveSizePolicy::update_averages(bool is_survivor_overflow,
|
|
size_t survived,
|
|
size_t promoted) {
|
|
// Update averages
|
|
if (!is_survivor_overflow) {
|
|
// Keep running averages on how much survived
|
|
_avg_survived->sample(survived);
|
|
} else {
|
|
size_t survived_guess = survived + promoted;
|
|
_avg_survived->sample(survived_guess);
|
|
}
|
|
avg_promoted()->sample(promoted);
|
|
|
|
log_trace(gc, ergo)("AdaptiveSizePolicy::update_averages: survived: " SIZE_FORMAT " promoted: " SIZE_FORMAT " overflow: %s",
|
|
survived, promoted, is_survivor_overflow ? "true" : "false");
|
|
}
|
|
|
|
bool PSAdaptiveSizePolicy::print() const {
|
|
|
|
if (!UseAdaptiveSizePolicy) {
|
|
return false;
|
|
}
|
|
|
|
if (AdaptiveSizePolicy::print()) {
|
|
AdaptiveSizePolicy::print_tenuring_threshold(PSScavenge::tenuring_threshold());
|
|
return true;
|
|
}
|
|
|
|
return false;
|
|
}
|