src/hotspot/share/gc/g1/g1Analytics.cpp
author tschatzl
Fri, 29 Nov 2019 10:20:14 +0100
changeset 59319 9ee940f1de90
parent 59222 f4f60bb75ee4
child 59320 11ff4e485670
permissions -rw-r--r--
8227739: Merge cost predictions for scanning cards and log buffer entries Summary: Revamp the cost predictions for the changes in JDK-8200545 and JDK-8213108. Reviewed-by: sjohanss, kbarrett

/*
 * Copyright (c) 2016, 2019, 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/g1Analytics.hpp"
#include "gc/g1/g1Predictions.hpp"
#include "runtime/globals.hpp"
#include "runtime/os.hpp"
#include "utilities/debug.hpp"
#include "utilities/globalDefinitions.hpp"
#include "utilities/numberSeq.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_logged_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_card_merge_to_scan_ratio_defaults[] = {
  1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0
};

static double young_only_cost_per_card_scan_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
};

G1Analytics::G1Analytics(const G1Predictions* predictor) :
    _predictor(predictor),
    _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)),
    _concurrent_refine_rate_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
    _logged_cards_rate_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
    _young_card_merge_to_scan_ratio_seq(new TruncatedSeq(TruncatedSeqLength)),
    _mixed_card_merge_to_scan_ratio_seq(new TruncatedSeq(TruncatedSeqLength)),
    _young_cost_per_card_scan_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
    _mixed_cost_per_card_scan_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
    _young_cost_per_card_merge_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
    _mixed_cost_per_card_merge_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
    _copy_cost_per_byte_ms_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_length_seq(new TruncatedSeq(TruncatedSeqLength)),
    _cost_per_byte_ms_during_cm_seq(new TruncatedSeq(TruncatedSeqLength)),
    _recent_prev_end_times_for_all_gcs_sec(new TruncatedSeq(NumPrevPausesForHeuristics)),
    _recent_avg_pause_time_ratio(0.0),
    _last_pause_time_ratio(0.0) {

  // Seed sequences with initial values.
  _recent_prev_end_times_for_all_gcs_sec->add(os::elapsedTime());
  _prev_collection_pause_end_ms = os::elapsedTime() * 1000.0;

  int index = MIN2(ParallelGCThreads - 1, 7u);

  _rs_length_diff_seq->add(rs_length_diff_defaults[index]);
  // Start with inverse of maximum STW cost.
  _concurrent_refine_rate_ms_seq->add(1/cost_per_logged_card_ms_defaults[0]);
  // Some applications have very low rates for logging cards.
  _logged_cards_rate_ms_seq->add(0.0);
  _young_card_merge_to_scan_ratio_seq->add(young_card_merge_to_scan_ratio_defaults[index]);
  _young_cost_per_card_scan_ms_seq->add(young_only_cost_per_card_scan_ms_defaults[index]);

  _copy_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]);

  // start conservatively (around 50ms is about right)
  _concurrent_mark_remark_times_ms->add(0.05);
  _concurrent_mark_cleanup_times_ms->add(0.20);
}

bool G1Analytics::enough_samples_available(TruncatedSeq const* seq) const {
  return seq->num() >= 3;
}

double G1Analytics::get_new_prediction(TruncatedSeq const* seq) const {
  return _predictor->get_new_prediction(seq);
}

size_t G1Analytics::get_new_size_prediction(TruncatedSeq const* seq) const {
  return (size_t)get_new_prediction(seq);
}

int G1Analytics::num_alloc_rate_ms() const {
  return _alloc_rate_ms_seq->num();
}

void G1Analytics::report_concurrent_mark_remark_times_ms(double ms) {
  _concurrent_mark_remark_times_ms->add(ms);
}

void G1Analytics::report_alloc_rate_ms(double alloc_rate) {
  _alloc_rate_ms_seq->add(alloc_rate);
}

void G1Analytics::compute_pause_time_ratio(double interval_ms, double pause_time_ms) {
  _recent_avg_pause_time_ratio = _recent_gc_times_ms->sum() / interval_ms;

  // Clamp the result to [0.0 ... 1.0] to filter out nonsensical results due to bad input.
  _recent_avg_pause_time_ratio = clamp(_recent_avg_pause_time_ratio, 0.0, 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;
}

void G1Analytics::report_concurrent_refine_rate_ms(double cards_per_ms) {
  _concurrent_refine_rate_ms_seq->add(cards_per_ms);
}

void G1Analytics::report_logged_cards_rate_ms(double cards_per_ms) {
  _logged_cards_rate_ms_seq->add(cards_per_ms);
}

void G1Analytics::report_cost_per_card_scan_ms(double cost_per_card_ms, bool for_young_gc) {
  if (for_young_gc) {
    _young_cost_per_card_scan_ms_seq->add(cost_per_card_ms);
  } else {
    _mixed_cost_per_card_scan_ms_seq->add(cost_per_card_ms);
  }
}

void G1Analytics::report_cost_per_card_merge_ms(double cost_per_card_ms, bool for_young_gc) {
  if (for_young_gc) {
    _young_cost_per_card_merge_ms_seq->add(cost_per_card_ms);
  } else {
    _mixed_cost_per_card_merge_ms_seq->add(cost_per_card_ms);
  }
}

void G1Analytics::report_card_merge_to_scan_ratio(double merge_to_scan_ratio, bool for_young_gc) {
  if (for_young_gc) {
    _young_card_merge_to_scan_ratio_seq->add(merge_to_scan_ratio);
  } else {
    _mixed_card_merge_to_scan_ratio_seq->add(merge_to_scan_ratio);
  }
}

void G1Analytics::report_rs_length_diff(double rs_length_diff) {
  _rs_length_diff_seq->add(rs_length_diff);
}

void G1Analytics::report_cost_per_byte_ms(double cost_per_byte_ms, bool mark_or_rebuild_in_progress) {
  if (mark_or_rebuild_in_progress) {
    _cost_per_byte_ms_during_cm_seq->add(cost_per_byte_ms);
  } else {
    _copy_cost_per_byte_ms_seq->add(cost_per_byte_ms);
  }
}

void G1Analytics::report_young_other_cost_per_region_ms(double other_cost_per_region_ms) {
  _young_other_cost_per_region_ms_seq->add(other_cost_per_region_ms);
}

void G1Analytics::report_non_young_other_cost_per_region_ms(double other_cost_per_region_ms) {
  _non_young_other_cost_per_region_ms_seq->add(other_cost_per_region_ms);
}

void G1Analytics::report_constant_other_time_ms(double constant_other_time_ms) {
  _constant_other_time_ms_seq->add(constant_other_time_ms);
}

void G1Analytics::report_pending_cards(double pending_cards) {
  _pending_cards_seq->add(pending_cards);
}

void G1Analytics::report_rs_length(double rs_length) {
  _rs_length_seq->add(rs_length);
}

double G1Analytics::predict_alloc_rate_ms() const {
  return get_new_prediction(_alloc_rate_ms_seq);
}

double G1Analytics::predict_concurrent_refine_rate_ms() const {
  return get_new_prediction(_concurrent_refine_rate_ms_seq);
}

double G1Analytics::predict_logged_cards_rate_ms() const {
  return get_new_prediction(_logged_cards_rate_ms_seq);
}

double G1Analytics::predict_young_card_merge_to_scan_ratio() const {
  return get_new_prediction(_young_card_merge_to_scan_ratio_seq);
}

size_t G1Analytics::predict_scan_card_num(size_t rs_length, bool for_young_gc) const {
  if (for_young_gc || !enough_samples_available(_mixed_card_merge_to_scan_ratio_seq)) {
    return (size_t) (rs_length * predict_young_card_merge_to_scan_ratio());
  } else {
    return (size_t) (rs_length * get_new_prediction(_mixed_card_merge_to_scan_ratio_seq));
  }
}

double G1Analytics::predict_card_merge_time_ms(size_t card_num, bool for_young_gc) const {
  if (for_young_gc || !enough_samples_available(_mixed_cost_per_card_merge_ms_seq)) {
    return card_num * get_new_prediction(_young_cost_per_card_merge_ms_seq);
  } else {
    return card_num * get_new_prediction(_mixed_cost_per_card_merge_ms_seq);
  }
}

double G1Analytics::predict_card_scan_time_ms(size_t card_num, bool for_young_gc) const {
  if (for_young_gc || !enough_samples_available(_mixed_cost_per_card_scan_ms_seq)) {
    return card_num * get_new_prediction(_young_cost_per_card_scan_ms_seq);
  } else {
    return card_num * get_new_prediction(_mixed_cost_per_card_scan_ms_seq);
  }
}

double G1Analytics::predict_object_copy_time_ms_during_cm(size_t bytes_to_copy) const {
  if (!enough_samples_available(_cost_per_byte_ms_during_cm_seq)) {
    return (1.1 * bytes_to_copy) * get_new_prediction(_copy_cost_per_byte_ms_seq);
  } else {
    return bytes_to_copy * get_new_prediction(_cost_per_byte_ms_during_cm_seq);
  }
}

double G1Analytics::predict_object_copy_time_ms(size_t bytes_to_copy, bool during_concurrent_mark) const {
  if (during_concurrent_mark) {
    return predict_object_copy_time_ms_during_cm(bytes_to_copy);
  } else {
    return bytes_to_copy * get_new_prediction(_copy_cost_per_byte_ms_seq);
  }
}

double G1Analytics::predict_constant_other_time_ms() const {
  return get_new_prediction(_constant_other_time_ms_seq);
}

double G1Analytics::predict_young_other_time_ms(size_t young_num) const {
  return young_num * get_new_prediction(_young_other_cost_per_region_ms_seq);
}

double G1Analytics::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 G1Analytics::predict_remark_time_ms() const {
  return get_new_prediction(_concurrent_mark_remark_times_ms);
}

double G1Analytics::predict_cleanup_time_ms() const {
  return get_new_prediction(_concurrent_mark_cleanup_times_ms);
}

size_t G1Analytics::predict_rs_length() const {
  return get_new_size_prediction(_rs_length_seq) + get_new_prediction(_rs_length_diff_seq);
}

size_t G1Analytics::predict_pending_cards() const {
  return get_new_size_prediction(_pending_cards_seq);
}

double G1Analytics::last_known_gc_end_time_sec() const {
  return _recent_prev_end_times_for_all_gcs_sec->oldest();
}

void G1Analytics::update_recent_gc_times(double end_time_sec,
                                         double pause_time_ms) {
  _recent_gc_times_ms->add(pause_time_ms);
  _recent_prev_end_times_for_all_gcs_sec->add(end_time_sec);
  _prev_collection_pause_end_ms = end_time_sec * 1000.0;
}

void G1Analytics::report_concurrent_mark_cleanup_times_ms(double ms) {
  _concurrent_mark_cleanup_times_ms->add(ms);
}