1 /* 2 * Copyright (c) 2001, 2012, Oracle and/or its affiliates. All rights reserved. 3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. 4 * 5 * This code is free software; you can redistribute it and/or modify it 6 * under the terms of the GNU General Public License version 2 only, as 7 * published by the Free Software Foundation. 8 * 9 * This code is distributed in the hope that it will be useful, but WITHOUT 10 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or 11 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License 12 * version 2 for more details (a copy is included in the LICENSE file that 13 * accompanied this code). 14 * 15 * You should have received a copy of the GNU General Public License version 16 * 2 along with this work; if not, write to the Free Software Foundation, 17 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. 18 * 19 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA 20 * or visit www.oracle.com if you need additional information or have any 21 * questions. 22 * 23 */ 24 25 #include "precompiled.hpp" 26 #include "gc_implementation/g1/concurrentG1Refine.hpp" 27 #include "gc_implementation/g1/concurrentMark.hpp" 28 #include "gc_implementation/g1/concurrentMarkThread.inline.hpp" 29 #include "gc_implementation/g1/g1CollectedHeap.inline.hpp" 30 #include "gc_implementation/g1/g1CollectorPolicy.hpp" 31 #include "gc_implementation/g1/g1ErgoVerbose.hpp" 32 #include "gc_implementation/g1/g1GCPhaseTimes.hpp" 33 #include "gc_implementation/g1/g1Log.hpp" 34 #include "gc_implementation/g1/heapRegionRemSet.hpp" 35 #include "gc_implementation/shared/gcPolicyCounters.hpp" 36 #include "runtime/arguments.hpp" 37 #include "runtime/java.hpp" 38 #include "runtime/mutexLocker.hpp" 39 #include "utilities/debug.hpp" 40 41 // Different defaults for different number of GC threads 42 // They were chosen by running GCOld and SPECjbb on debris with different 43 // numbers of GC threads and choosing them based on the results 44 45 // all the same 46 static double rs_length_diff_defaults[] = { 47 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 48 }; 49 50 static double cost_per_card_ms_defaults[] = { 51 0.01, 0.005, 0.005, 0.003, 0.003, 0.002, 0.002, 0.0015 52 }; 53 54 // all the same 55 static double young_cards_per_entry_ratio_defaults[] = { 56 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0 57 }; 58 59 static double cost_per_entry_ms_defaults[] = { 60 0.015, 0.01, 0.01, 0.008, 0.008, 0.0055, 0.0055, 0.005 61 }; 62 63 static double cost_per_byte_ms_defaults[] = { 64 0.00006, 0.00003, 0.00003, 0.000015, 0.000015, 0.00001, 0.00001, 0.000009 65 }; 66 67 // these should be pretty consistent 68 static double constant_other_time_ms_defaults[] = { 69 5.0, 5.0, 5.0, 5.0, 5.0, 5.0, 5.0, 5.0 70 }; 71 72 73 static double young_other_cost_per_region_ms_defaults[] = { 74 0.3, 0.2, 0.2, 0.15, 0.15, 0.12, 0.12, 0.1 75 }; 76 77 static double non_young_other_cost_per_region_ms_defaults[] = { 78 1.0, 0.7, 0.7, 0.5, 0.5, 0.42, 0.42, 0.30 79 }; 80 81 G1CollectorPolicy::G1CollectorPolicy() : 82 _parallel_gc_threads(G1CollectedHeap::use_parallel_gc_threads() 83 ? ParallelGCThreads : 1), 84 85 _recent_gc_times_ms(new TruncatedSeq(NumPrevPausesForHeuristics)), 86 _stop_world_start(0.0), 87 88 _concurrent_mark_remark_times_ms(new TruncatedSeq(NumPrevPausesForHeuristics)), 89 _concurrent_mark_cleanup_times_ms(new TruncatedSeq(NumPrevPausesForHeuristics)), 90 91 _alloc_rate_ms_seq(new TruncatedSeq(TruncatedSeqLength)), 92 _prev_collection_pause_end_ms(0.0), 93 _rs_length_diff_seq(new TruncatedSeq(TruncatedSeqLength)), 94 _cost_per_card_ms_seq(new TruncatedSeq(TruncatedSeqLength)), 95 _young_cards_per_entry_ratio_seq(new TruncatedSeq(TruncatedSeqLength)), 96 _mixed_cards_per_entry_ratio_seq(new TruncatedSeq(TruncatedSeqLength)), 97 _cost_per_entry_ms_seq(new TruncatedSeq(TruncatedSeqLength)), 98 _mixed_cost_per_entry_ms_seq(new TruncatedSeq(TruncatedSeqLength)), 99 _cost_per_byte_ms_seq(new TruncatedSeq(TruncatedSeqLength)), 100 _cost_per_byte_ms_during_cm_seq(new TruncatedSeq(TruncatedSeqLength)), 101 _constant_other_time_ms_seq(new TruncatedSeq(TruncatedSeqLength)), 102 _young_other_cost_per_region_ms_seq(new TruncatedSeq(TruncatedSeqLength)), 103 _non_young_other_cost_per_region_ms_seq( 104 new TruncatedSeq(TruncatedSeqLength)), 105 106 _pending_cards_seq(new TruncatedSeq(TruncatedSeqLength)), 107 _rs_lengths_seq(new TruncatedSeq(TruncatedSeqLength)), 108 109 _pause_time_target_ms((double) MaxGCPauseMillis), 110 111 _gcs_are_young(true), 112 113 _during_marking(false), 114 _in_marking_window(false), 115 _in_marking_window_im(false), 116 117 _recent_prev_end_times_for_all_gcs_sec( 118 new TruncatedSeq(NumPrevPausesForHeuristics)), 119 120 _recent_avg_pause_time_ratio(0.0), 121 122 _initiate_conc_mark_if_possible(false), 123 _during_initial_mark_pause(false), 124 _last_young_gc(false), 125 _last_gc_was_young(false), 126 127 _eden_bytes_before_gc(0), 128 _survivor_bytes_before_gc(0), 129 _capacity_before_gc(0), 130 131 _eden_cset_region_length(0), 132 _survivor_cset_region_length(0), 133 _old_cset_region_length(0), 134 135 _collection_set(NULL), 136 _collection_set_bytes_used_before(0), 137 138 // Incremental CSet attributes 139 _inc_cset_build_state(Inactive), 140 _inc_cset_head(NULL), 141 _inc_cset_tail(NULL), 142 _inc_cset_bytes_used_before(0), 143 _inc_cset_max_finger(NULL), 144 _inc_cset_recorded_rs_lengths(0), 145 _inc_cset_recorded_rs_lengths_diffs(0), 146 _inc_cset_predicted_elapsed_time_ms(0.0), 147 _inc_cset_predicted_elapsed_time_ms_diffs(0.0), 148 149 #ifdef _MSC_VER // the use of 'this' below gets a warning, make it go away 150 #pragma warning( disable:4355 ) // 'this' : used in base member initializer list 151 #endif // _MSC_VER 152 153 _short_lived_surv_rate_group(new SurvRateGroup(this, "Short Lived", 154 G1YoungSurvRateNumRegionsSummary)), 155 _survivor_surv_rate_group(new SurvRateGroup(this, "Survivor", 156 G1YoungSurvRateNumRegionsSummary)), 157 // add here any more surv rate groups 158 _recorded_survivor_regions(0), 159 _recorded_survivor_head(NULL), 160 _recorded_survivor_tail(NULL), 161 _survivors_age_table(true), 162 163 _gc_overhead_perc(0.0) { 164 165 // Set up the region size and associated fields. Given that the 166 // policy is created before the heap, we have to set this up here, 167 // so it's done as soon as possible. 168 HeapRegion::setup_heap_region_size(Arguments::min_heap_size()); 169 HeapRegionRemSet::setup_remset_size(); 170 171 G1ErgoVerbose::initialize(); 172 if (PrintAdaptiveSizePolicy) { 173 // Currently, we only use a single switch for all the heuristics. 174 G1ErgoVerbose::set_enabled(true); 175 // Given that we don't currently have a verboseness level 176 // parameter, we'll hardcode this to high. This can be easily 177 // changed in the future. 178 G1ErgoVerbose::set_level(ErgoHigh); 179 } else { 180 G1ErgoVerbose::set_enabled(false); 181 } 182 183 // Verify PLAB sizes 184 const size_t region_size = HeapRegion::GrainWords; 185 if (YoungPLABSize > region_size || OldPLABSize > region_size) { 186 char buffer[128]; 187 jio_snprintf(buffer, sizeof(buffer), "%sPLABSize should be at most "SIZE_FORMAT, 188 OldPLABSize > region_size ? "Old" : "Young", region_size); 189 vm_exit_during_initialization(buffer); 190 } 191 192 _recent_prev_end_times_for_all_gcs_sec->add(os::elapsedTime()); 193 _prev_collection_pause_end_ms = os::elapsedTime() * 1000.0; 194 195 _phase_times = new G1GCPhaseTimes(_parallel_gc_threads); 196 197 int index = MIN2(_parallel_gc_threads - 1, 7); 198 199 _rs_length_diff_seq->add(rs_length_diff_defaults[index]); 200 _cost_per_card_ms_seq->add(cost_per_card_ms_defaults[index]); 201 _young_cards_per_entry_ratio_seq->add( 202 young_cards_per_entry_ratio_defaults[index]); 203 _cost_per_entry_ms_seq->add(cost_per_entry_ms_defaults[index]); 204 _cost_per_byte_ms_seq->add(cost_per_byte_ms_defaults[index]); 205 _constant_other_time_ms_seq->add(constant_other_time_ms_defaults[index]); 206 _young_other_cost_per_region_ms_seq->add( 207 young_other_cost_per_region_ms_defaults[index]); 208 _non_young_other_cost_per_region_ms_seq->add( 209 non_young_other_cost_per_region_ms_defaults[index]); 210 211 // Below, we might need to calculate the pause time target based on 212 // the pause interval. When we do so we are going to give G1 maximum 213 // flexibility and allow it to do pauses when it needs to. So, we'll 214 // arrange that the pause interval to be pause time target + 1 to 215 // ensure that a) the pause time target is maximized with respect to 216 // the pause interval and b) we maintain the invariant that pause 217 // time target < pause interval. If the user does not want this 218 // maximum flexibility, they will have to set the pause interval 219 // explicitly. 220 221 // First make sure that, if either parameter is set, its value is 222 // reasonable. 223 if (!FLAG_IS_DEFAULT(MaxGCPauseMillis)) { 224 if (MaxGCPauseMillis < 1) { 225 vm_exit_during_initialization("MaxGCPauseMillis should be " 226 "greater than 0"); 227 } 228 } 229 if (!FLAG_IS_DEFAULT(GCPauseIntervalMillis)) { 230 if (GCPauseIntervalMillis < 1) { 231 vm_exit_during_initialization("GCPauseIntervalMillis should be " 232 "greater than 0"); 233 } 234 } 235 236 // Then, if the pause time target parameter was not set, set it to 237 // the default value. 238 if (FLAG_IS_DEFAULT(MaxGCPauseMillis)) { 239 if (FLAG_IS_DEFAULT(GCPauseIntervalMillis)) { 240 // The default pause time target in G1 is 200ms 241 FLAG_SET_DEFAULT(MaxGCPauseMillis, 200); 242 } else { 243 // We do not allow the pause interval to be set without the 244 // pause time target 245 vm_exit_during_initialization("GCPauseIntervalMillis cannot be set " 246 "without setting MaxGCPauseMillis"); 247 } 248 } 249 250 // Then, if the interval parameter was not set, set it according to 251 // the pause time target (this will also deal with the case when the 252 // pause time target is the default value). 253 if (FLAG_IS_DEFAULT(GCPauseIntervalMillis)) { 254 FLAG_SET_DEFAULT(GCPauseIntervalMillis, MaxGCPauseMillis + 1); 255 } 256 257 // Finally, make sure that the two parameters are consistent. 258 if (MaxGCPauseMillis >= GCPauseIntervalMillis) { 259 char buffer[256]; 260 jio_snprintf(buffer, 256, 261 "MaxGCPauseMillis (%u) should be less than " 262 "GCPauseIntervalMillis (%u)", 263 MaxGCPauseMillis, GCPauseIntervalMillis); 264 vm_exit_during_initialization(buffer); 265 } 266 267 double max_gc_time = (double) MaxGCPauseMillis / 1000.0; 268 double time_slice = (double) GCPauseIntervalMillis / 1000.0; 269 _mmu_tracker = new G1MMUTrackerQueue(time_slice, max_gc_time); 270 _sigma = (double) G1ConfidencePercent / 100.0; 271 272 // start conservatively (around 50ms is about right) 273 _concurrent_mark_remark_times_ms->add(0.05); 274 _concurrent_mark_cleanup_times_ms->add(0.20); 275 _tenuring_threshold = MaxTenuringThreshold; 276 // _max_survivor_regions will be calculated by 277 // update_young_list_target_length() during initialization. 278 _max_survivor_regions = 0; 279 280 assert(GCTimeRatio > 0, 281 "we should have set it to a default value set_g1_gc_flags() " 282 "if a user set it to 0"); 283 _gc_overhead_perc = 100.0 * (1.0 / (1.0 + GCTimeRatio)); 284 285 uintx reserve_perc = G1ReservePercent; 286 // Put an artificial ceiling on this so that it's not set to a silly value. 287 if (reserve_perc > 50) { 288 reserve_perc = 50; 289 warning("G1ReservePercent is set to a value that is too large, " 290 "it's been updated to %u", reserve_perc); 291 } 292 _reserve_factor = (double) reserve_perc / 100.0; 293 // This will be set when the heap is expanded 294 // for the first time during initialization. 295 _reserve_regions = 0; 296 297 initialize_all(); 298 _collectionSetChooser = new CollectionSetChooser(); 299 _young_gen_sizer = new G1YoungGenSizer(); // Must be after call to initialize_flags 300 } 301 302 void G1CollectorPolicy::initialize_flags() { 303 set_min_alignment(HeapRegion::GrainBytes); 304 set_max_alignment(GenRemSet::max_alignment_constraint(rem_set_name())); 305 if (SurvivorRatio < 1) { 306 vm_exit_during_initialization("Invalid survivor ratio specified"); 307 } 308 CollectorPolicy::initialize_flags(); 309 } 310 311 G1YoungGenSizer::G1YoungGenSizer() : _sizer_kind(SizerDefaults), _adaptive_size(true) { 312 assert(G1DefaultMinNewGenPercent <= G1DefaultMaxNewGenPercent, "Min larger than max"); 313 assert(G1DefaultMinNewGenPercent > 0 && G1DefaultMinNewGenPercent < 100, "Min out of bounds"); 314 assert(G1DefaultMaxNewGenPercent > 0 && G1DefaultMaxNewGenPercent < 100, "Max out of bounds"); 315 316 if (FLAG_IS_CMDLINE(NewRatio)) { 317 if (FLAG_IS_CMDLINE(NewSize) || FLAG_IS_CMDLINE(MaxNewSize)) { 318 warning("-XX:NewSize and -XX:MaxNewSize override -XX:NewRatio"); 319 } else { 320 _sizer_kind = SizerNewRatio; 321 _adaptive_size = false; 322 return; 323 } 324 } 325 326 if (FLAG_IS_CMDLINE(NewSize)) { 327 _min_desired_young_length = MAX2((uint) (NewSize / HeapRegion::GrainBytes), 328 1U); 329 if (FLAG_IS_CMDLINE(MaxNewSize)) { 330 _max_desired_young_length = 331 MAX2((uint) (MaxNewSize / HeapRegion::GrainBytes), 332 1U); 333 _sizer_kind = SizerMaxAndNewSize; 334 _adaptive_size = _min_desired_young_length == _max_desired_young_length; 335 } else { 336 _sizer_kind = SizerNewSizeOnly; 337 } 338 } else if (FLAG_IS_CMDLINE(MaxNewSize)) { 339 _max_desired_young_length = 340 MAX2((uint) (MaxNewSize / HeapRegion::GrainBytes), 341 1U); 342 _sizer_kind = SizerMaxNewSizeOnly; 343 } 344 } 345 346 uint G1YoungGenSizer::calculate_default_min_length(uint new_number_of_heap_regions) { 347 uint default_value = (new_number_of_heap_regions * G1DefaultMinNewGenPercent) / 100; 348 return MAX2(1U, default_value); 349 } 350 351 uint G1YoungGenSizer::calculate_default_max_length(uint new_number_of_heap_regions) { 352 uint default_value = (new_number_of_heap_regions * G1DefaultMaxNewGenPercent) / 100; 353 return MAX2(1U, default_value); 354 } 355 356 void G1YoungGenSizer::heap_size_changed(uint new_number_of_heap_regions) { 357 assert(new_number_of_heap_regions > 0, "Heap must be initialized"); 358 359 switch (_sizer_kind) { 360 case SizerDefaults: 361 _min_desired_young_length = calculate_default_min_length(new_number_of_heap_regions); 362 _max_desired_young_length = calculate_default_max_length(new_number_of_heap_regions); 363 break; 364 case SizerNewSizeOnly: 365 _max_desired_young_length = calculate_default_max_length(new_number_of_heap_regions); 366 _max_desired_young_length = MAX2(_min_desired_young_length, _max_desired_young_length); 367 break; 368 case SizerMaxNewSizeOnly: 369 _min_desired_young_length = calculate_default_min_length(new_number_of_heap_regions); 370 _min_desired_young_length = MIN2(_min_desired_young_length, _max_desired_young_length); 371 break; 372 case SizerMaxAndNewSize: 373 // Do nothing. Values set on the command line, don't update them at runtime. 374 break; 375 case SizerNewRatio: 376 _min_desired_young_length = new_number_of_heap_regions / (NewRatio + 1); 377 _max_desired_young_length = _min_desired_young_length; 378 break; 379 default: 380 ShouldNotReachHere(); 381 } 382 383 assert(_min_desired_young_length <= _max_desired_young_length, "Invalid min/max young gen size values"); 384 } 385 386 void G1CollectorPolicy::init() { 387 // Set aside an initial future to_space. 388 _g1 = G1CollectedHeap::heap(); 389 390 assert(Heap_lock->owned_by_self(), "Locking discipline."); 391 392 initialize_gc_policy_counters(); 393 394 if (adaptive_young_list_length()) { 395 _young_list_fixed_length = 0; 396 } else { 397 _young_list_fixed_length = _young_gen_sizer->min_desired_young_length(); 398 } 399 _free_regions_at_end_of_collection = _g1->free_regions(); 400 update_young_list_target_length(); 401 _prev_eden_capacity = _young_list_target_length * HeapRegion::GrainBytes; 402 403 // We may immediately start allocating regions and placing them on the 404 // collection set list. Initialize the per-collection set info 405 start_incremental_cset_building(); 406 } 407 408 // Create the jstat counters for the policy. 409 void G1CollectorPolicy::initialize_gc_policy_counters() { 410 _gc_policy_counters = new GCPolicyCounters("GarbageFirst", 1, 3); 411 } 412 413 bool G1CollectorPolicy::predict_will_fit(uint young_length, 414 double base_time_ms, 415 uint base_free_regions, 416 double target_pause_time_ms) { 417 if (young_length >= base_free_regions) { 418 // end condition 1: not enough space for the young regions 419 return false; 420 } 421 422 double accum_surv_rate = accum_yg_surv_rate_pred((int) young_length - 1); 423 size_t bytes_to_copy = 424 (size_t) (accum_surv_rate * (double) HeapRegion::GrainBytes); 425 double copy_time_ms = predict_object_copy_time_ms(bytes_to_copy); 426 double young_other_time_ms = predict_young_other_time_ms(young_length); 427 double pause_time_ms = base_time_ms + copy_time_ms + young_other_time_ms; 428 if (pause_time_ms > target_pause_time_ms) { 429 // end condition 2: prediction is over the target pause time 430 return false; 431 } 432 433 size_t free_bytes = 434 (base_free_regions - young_length) * HeapRegion::GrainBytes; 435 if ((2.0 * sigma()) * (double) bytes_to_copy > (double) free_bytes) { 436 // end condition 3: out-of-space (conservatively!) 437 return false; 438 } 439 440 // success! 441 return true; 442 } 443 444 void G1CollectorPolicy::record_new_heap_size(uint new_number_of_regions) { 445 // re-calculate the necessary reserve 446 double reserve_regions_d = (double) new_number_of_regions * _reserve_factor; 447 // We use ceiling so that if reserve_regions_d is > 0.0 (but 448 // smaller than 1.0) we'll get 1. 449 _reserve_regions = (uint) ceil(reserve_regions_d); 450 451 _young_gen_sizer->heap_size_changed(new_number_of_regions); 452 } 453 454 uint G1CollectorPolicy::calculate_young_list_desired_min_length( 455 uint base_min_length) { 456 uint desired_min_length = 0; 457 if (adaptive_young_list_length()) { 458 if (_alloc_rate_ms_seq->num() > 3) { 459 double now_sec = os::elapsedTime(); 460 double when_ms = _mmu_tracker->when_max_gc_sec(now_sec) * 1000.0; 461 double alloc_rate_ms = predict_alloc_rate_ms(); 462 desired_min_length = (uint) ceil(alloc_rate_ms * when_ms); 463 } else { 464 // otherwise we don't have enough info to make the prediction 465 } 466 } 467 desired_min_length += base_min_length; 468 // make sure we don't go below any user-defined minimum bound 469 return MAX2(_young_gen_sizer->min_desired_young_length(), desired_min_length); 470 } 471 472 uint G1CollectorPolicy::calculate_young_list_desired_max_length() { 473 // Here, we might want to also take into account any additional 474 // constraints (i.e., user-defined minimum bound). Currently, we 475 // effectively don't set this bound. 476 return _young_gen_sizer->max_desired_young_length(); 477 } 478 479 void G1CollectorPolicy::update_young_list_target_length(size_t rs_lengths) { 480 if (rs_lengths == (size_t) -1) { 481 // if it's set to the default value (-1), we should predict it; 482 // otherwise, use the given value. 483 rs_lengths = (size_t) get_new_prediction(_rs_lengths_seq); 484 } 485 486 // Calculate the absolute and desired min bounds. 487 488 // This is how many young regions we already have (currently: the survivors). 489 uint base_min_length = recorded_survivor_regions(); 490 // This is the absolute minimum young length, which ensures that we 491 // can allocate one eden region in the worst-case. 492 uint absolute_min_length = base_min_length + 1; 493 uint desired_min_length = 494 calculate_young_list_desired_min_length(base_min_length); 495 if (desired_min_length < absolute_min_length) { 496 desired_min_length = absolute_min_length; 497 } 498 499 // Calculate the absolute and desired max bounds. 500 501 // We will try our best not to "eat" into the reserve. 502 uint absolute_max_length = 0; 503 if (_free_regions_at_end_of_collection > _reserve_regions) { 504 absolute_max_length = _free_regions_at_end_of_collection - _reserve_regions; 505 } 506 uint desired_max_length = calculate_young_list_desired_max_length(); 507 if (desired_max_length > absolute_max_length) { 508 desired_max_length = absolute_max_length; 509 } 510 511 uint young_list_target_length = 0; 512 if (adaptive_young_list_length()) { 513 if (gcs_are_young()) { 514 young_list_target_length = 515 calculate_young_list_target_length(rs_lengths, 516 base_min_length, 517 desired_min_length, 518 desired_max_length); 519 _rs_lengths_prediction = rs_lengths; 520 } else { 521 // Don't calculate anything and let the code below bound it to 522 // the desired_min_length, i.e., do the next GC as soon as 523 // possible to maximize how many old regions we can add to it. 524 } 525 } else { 526 // The user asked for a fixed young gen so we'll fix the young gen 527 // whether the next GC is young or mixed. 528 young_list_target_length = _young_list_fixed_length; 529 } 530 531 // Make sure we don't go over the desired max length, nor under the 532 // desired min length. In case they clash, desired_min_length wins 533 // which is why that test is second. 534 if (young_list_target_length > desired_max_length) { 535 young_list_target_length = desired_max_length; 536 } 537 if (young_list_target_length < desired_min_length) { 538 young_list_target_length = desired_min_length; 539 } 540 541 assert(young_list_target_length > recorded_survivor_regions(), 542 "we should be able to allocate at least one eden region"); 543 assert(young_list_target_length >= absolute_min_length, "post-condition"); 544 _young_list_target_length = young_list_target_length; 545 546 update_max_gc_locker_expansion(); 547 } 548 549 uint 550 G1CollectorPolicy::calculate_young_list_target_length(size_t rs_lengths, 551 uint base_min_length, 552 uint desired_min_length, 553 uint desired_max_length) { 554 assert(adaptive_young_list_length(), "pre-condition"); 555 assert(gcs_are_young(), "only call this for young GCs"); 556 557 // In case some edge-condition makes the desired max length too small... 558 if (desired_max_length <= desired_min_length) { 559 return desired_min_length; 560 } 561 562 // We'll adjust min_young_length and max_young_length not to include 563 // the already allocated young regions (i.e., so they reflect the 564 // min and max eden regions we'll allocate). The base_min_length 565 // will be reflected in the predictions by the 566 // survivor_regions_evac_time prediction. 567 assert(desired_min_length > base_min_length, "invariant"); 568 uint min_young_length = desired_min_length - base_min_length; 569 assert(desired_max_length > base_min_length, "invariant"); 570 uint max_young_length = desired_max_length - base_min_length; 571 572 double target_pause_time_ms = _mmu_tracker->max_gc_time() * 1000.0; 573 double survivor_regions_evac_time = predict_survivor_regions_evac_time(); 574 size_t pending_cards = (size_t) get_new_prediction(_pending_cards_seq); 575 size_t adj_rs_lengths = rs_lengths + predict_rs_length_diff(); 576 size_t scanned_cards = predict_young_card_num(adj_rs_lengths); 577 double base_time_ms = 578 predict_base_elapsed_time_ms(pending_cards, scanned_cards) + 579 survivor_regions_evac_time; 580 uint available_free_regions = _free_regions_at_end_of_collection; 581 uint base_free_regions = 0; 582 if (available_free_regions > _reserve_regions) { 583 base_free_regions = available_free_regions - _reserve_regions; 584 } 585 586 // Here, we will make sure that the shortest young length that 587 // makes sense fits within the target pause time. 588 589 if (predict_will_fit(min_young_length, base_time_ms, 590 base_free_regions, target_pause_time_ms)) { 591 // The shortest young length will fit into the target pause time; 592 // we'll now check whether the absolute maximum number of young 593 // regions will fit in the target pause time. If not, we'll do 594 // a binary search between min_young_length and max_young_length. 595 if (predict_will_fit(max_young_length, base_time_ms, 596 base_free_regions, target_pause_time_ms)) { 597 // The maximum young length will fit into the target pause time. 598 // We are done so set min young length to the maximum length (as 599 // the result is assumed to be returned in min_young_length). 600 min_young_length = max_young_length; 601 } else { 602 // The maximum possible number of young regions will not fit within 603 // the target pause time so we'll search for the optimal 604 // length. The loop invariants are: 605 // 606 // min_young_length < max_young_length 607 // min_young_length is known to fit into the target pause time 608 // max_young_length is known not to fit into the target pause time 609 // 610 // Going into the loop we know the above hold as we've just 611 // checked them. Every time around the loop we check whether 612 // the middle value between min_young_length and 613 // max_young_length fits into the target pause time. If it 614 // does, it becomes the new min. If it doesn't, it becomes 615 // the new max. This way we maintain the loop invariants. 616 617 assert(min_young_length < max_young_length, "invariant"); 618 uint diff = (max_young_length - min_young_length) / 2; 619 while (diff > 0) { 620 uint young_length = min_young_length + diff; 621 if (predict_will_fit(young_length, base_time_ms, 622 base_free_regions, target_pause_time_ms)) { 623 min_young_length = young_length; 624 } else { 625 max_young_length = young_length; 626 } 627 assert(min_young_length < max_young_length, "invariant"); 628 diff = (max_young_length - min_young_length) / 2; 629 } 630 // The results is min_young_length which, according to the 631 // loop invariants, should fit within the target pause time. 632 633 // These are the post-conditions of the binary search above: 634 assert(min_young_length < max_young_length, 635 "otherwise we should have discovered that max_young_length " 636 "fits into the pause target and not done the binary search"); 637 assert(predict_will_fit(min_young_length, base_time_ms, 638 base_free_regions, target_pause_time_ms), 639 "min_young_length, the result of the binary search, should " 640 "fit into the pause target"); 641 assert(!predict_will_fit(min_young_length + 1, base_time_ms, 642 base_free_regions, target_pause_time_ms), 643 "min_young_length, the result of the binary search, should be " 644 "optimal, so no larger length should fit into the pause target"); 645 } 646 } else { 647 // Even the minimum length doesn't fit into the pause time 648 // target, return it as the result nevertheless. 649 } 650 return base_min_length + min_young_length; 651 } 652 653 double G1CollectorPolicy::predict_survivor_regions_evac_time() { 654 double survivor_regions_evac_time = 0.0; 655 for (HeapRegion * r = _recorded_survivor_head; 656 r != NULL && r != _recorded_survivor_tail->get_next_young_region(); 657 r = r->get_next_young_region()) { 658 survivor_regions_evac_time += predict_region_elapsed_time_ms(r, gcs_are_young()); 659 } 660 return survivor_regions_evac_time; 661 } 662 663 void G1CollectorPolicy::revise_young_list_target_length_if_necessary() { 664 guarantee( adaptive_young_list_length(), "should not call this otherwise" ); 665 666 size_t rs_lengths = _g1->young_list()->sampled_rs_lengths(); 667 if (rs_lengths > _rs_lengths_prediction) { 668 // add 10% to avoid having to recalculate often 669 size_t rs_lengths_prediction = rs_lengths * 1100 / 1000; 670 update_young_list_target_length(rs_lengths_prediction); 671 } 672 } 673 674 675 676 HeapWord* G1CollectorPolicy::mem_allocate_work(size_t size, 677 bool is_tlab, 678 bool* gc_overhead_limit_was_exceeded) { 679 guarantee(false, "Not using this policy feature yet."); 680 return NULL; 681 } 682 683 // This method controls how a collector handles one or more 684 // of its generations being fully allocated. 685 HeapWord* G1CollectorPolicy::satisfy_failed_allocation(size_t size, 686 bool is_tlab) { 687 guarantee(false, "Not using this policy feature yet."); 688 return NULL; 689 } 690 691 692 #ifndef PRODUCT 693 bool G1CollectorPolicy::verify_young_ages() { 694 HeapRegion* head = _g1->young_list()->first_region(); 695 return 696 verify_young_ages(head, _short_lived_surv_rate_group); 697 // also call verify_young_ages on any additional surv rate groups 698 } 699 700 bool 701 G1CollectorPolicy::verify_young_ages(HeapRegion* head, 702 SurvRateGroup *surv_rate_group) { 703 guarantee( surv_rate_group != NULL, "pre-condition" ); 704 705 const char* name = surv_rate_group->name(); 706 bool ret = true; 707 int prev_age = -1; 708 709 for (HeapRegion* curr = head; 710 curr != NULL; 711 curr = curr->get_next_young_region()) { 712 SurvRateGroup* group = curr->surv_rate_group(); 713 if (group == NULL && !curr->is_survivor()) { 714 gclog_or_tty->print_cr("## %s: encountered NULL surv_rate_group", name); 715 ret = false; 716 } 717 718 if (surv_rate_group == group) { 719 int age = curr->age_in_surv_rate_group(); 720 721 if (age < 0) { 722 gclog_or_tty->print_cr("## %s: encountered negative age", name); 723 ret = false; 724 } 725 726 if (age <= prev_age) { 727 gclog_or_tty->print_cr("## %s: region ages are not strictly increasing " 728 "(%d, %d)", name, age, prev_age); 729 ret = false; 730 } 731 prev_age = age; 732 } 733 } 734 735 return ret; 736 } 737 #endif // PRODUCT 738 739 void G1CollectorPolicy::record_full_collection_start() { 740 _full_collection_start_sec = os::elapsedTime(); 741 // Release the future to-space so that it is available for compaction into. 742 _g1->set_full_collection(); 743 } 744 745 void G1CollectorPolicy::record_full_collection_end() { 746 // Consider this like a collection pause for the purposes of allocation 747 // since last pause. 748 double end_sec = os::elapsedTime(); 749 double full_gc_time_sec = end_sec - _full_collection_start_sec; 750 double full_gc_time_ms = full_gc_time_sec * 1000.0; 751 752 _trace_gen1_time_data.record_full_collection(full_gc_time_ms); 753 754 update_recent_gc_times(end_sec, full_gc_time_ms); 755 756 _g1->clear_full_collection(); 757 758 // "Nuke" the heuristics that control the young/mixed GC 759 // transitions and make sure we start with young GCs after the Full GC. 760 set_gcs_are_young(true); 761 _last_young_gc = false; 762 clear_initiate_conc_mark_if_possible(); 763 clear_during_initial_mark_pause(); 764 _in_marking_window = false; 765 _in_marking_window_im = false; 766 767 _short_lived_surv_rate_group->start_adding_regions(); 768 // also call this on any additional surv rate groups 769 770 record_survivor_regions(0, NULL, NULL); 771 772 _free_regions_at_end_of_collection = _g1->free_regions(); 773 // Reset survivors SurvRateGroup. 774 _survivor_surv_rate_group->reset(); 775 update_young_list_target_length(); 776 _collectionSetChooser->clear(); 777 } 778 779 void G1CollectorPolicy::record_stop_world_start() { 780 _stop_world_start = os::elapsedTime(); 781 } 782 783 void G1CollectorPolicy::record_collection_pause_start(double start_time_sec, 784 size_t start_used) { 785 // We only need to do this here as the policy will only be applied 786 // to the GC we're about to start. so, no point is calculating this 787 // every time we calculate / recalculate the target young length. 788 update_survivors_policy(); 789 790 assert(_g1->used() == _g1->recalculate_used(), 791 err_msg("sanity, used: "SIZE_FORMAT" recalculate_used: "SIZE_FORMAT, 792 _g1->used(), _g1->recalculate_used())); 793 794 double s_w_t_ms = (start_time_sec - _stop_world_start) * 1000.0; 795 _trace_gen0_time_data.record_start_collection(s_w_t_ms); 796 _stop_world_start = 0.0; 797 798 phase_times()->record_cur_collection_start_sec(start_time_sec); 799 _cur_collection_pause_used_at_start_bytes = start_used; 800 _cur_collection_pause_used_regions_at_start = _g1->used_regions(); 801 _pending_cards = _g1->pending_card_num(); 802 803 _collection_set_bytes_used_before = 0; 804 _bytes_copied_during_gc = 0; 805 806 YoungList* young_list = _g1->young_list(); 807 _eden_bytes_before_gc = young_list->eden_used_bytes(); 808 _survivor_bytes_before_gc = young_list->survivor_used_bytes(); 809 _capacity_before_gc = _g1->capacity(); 810 811 _last_gc_was_young = false; 812 813 // do that for any other surv rate groups 814 _short_lived_surv_rate_group->stop_adding_regions(); 815 _survivors_age_table.clear(); 816 817 assert( verify_young_ages(), "region age verification" ); 818 } 819 820 void G1CollectorPolicy::record_concurrent_mark_init_end(double 821 mark_init_elapsed_time_ms) { 822 _during_marking = true; 823 assert(!initiate_conc_mark_if_possible(), "we should have cleared it by now"); 824 clear_during_initial_mark_pause(); 825 _cur_mark_stop_world_time_ms = mark_init_elapsed_time_ms; 826 } 827 828 void G1CollectorPolicy::record_concurrent_mark_remark_start() { 829 _mark_remark_start_sec = os::elapsedTime(); 830 _during_marking = false; 831 } 832 833 void G1CollectorPolicy::record_concurrent_mark_remark_end() { 834 double end_time_sec = os::elapsedTime(); 835 double elapsed_time_ms = (end_time_sec - _mark_remark_start_sec)*1000.0; 836 _concurrent_mark_remark_times_ms->add(elapsed_time_ms); 837 _cur_mark_stop_world_time_ms += elapsed_time_ms; 838 _prev_collection_pause_end_ms += elapsed_time_ms; 839 840 _mmu_tracker->add_pause(_mark_remark_start_sec, end_time_sec, true); 841 } 842 843 void G1CollectorPolicy::record_concurrent_mark_cleanup_start() { 844 _mark_cleanup_start_sec = os::elapsedTime(); 845 } 846 847 void G1CollectorPolicy::record_concurrent_mark_cleanup_completed() { 848 _last_young_gc = true; 849 _in_marking_window = false; 850 } 851 852 void G1CollectorPolicy::record_concurrent_pause() { 853 if (_stop_world_start > 0.0) { 854 double yield_ms = (os::elapsedTime() - _stop_world_start) * 1000.0; 855 _trace_gen0_time_data.record_yield_time(yield_ms); 856 } 857 } 858 859 bool G1CollectorPolicy::need_to_start_conc_mark(const char* source, size_t alloc_word_size) { 860 if (_g1->concurrent_mark()->cmThread()->during_cycle()) { 861 return false; 862 } 863 864 size_t marking_initiating_used_threshold = 865 (_g1->capacity() / 100) * InitiatingHeapOccupancyPercent; 866 size_t cur_used_bytes = _g1->non_young_capacity_bytes(); 867 size_t alloc_byte_size = alloc_word_size * HeapWordSize; 868 869 if ((cur_used_bytes + alloc_byte_size) > marking_initiating_used_threshold) { 870 if (gcs_are_young()) { 871 ergo_verbose5(ErgoConcCycles, 872 "request concurrent cycle initiation", 873 ergo_format_reason("occupancy higher than threshold") 874 ergo_format_byte("occupancy") 875 ergo_format_byte("allocation request") 876 ergo_format_byte_perc("threshold") 877 ergo_format_str("source"), 878 cur_used_bytes, 879 alloc_byte_size, 880 marking_initiating_used_threshold, 881 (double) InitiatingHeapOccupancyPercent, 882 source); 883 return true; 884 } else { 885 ergo_verbose5(ErgoConcCycles, 886 "do not request concurrent cycle initiation", 887 ergo_format_reason("still doing mixed collections") 888 ergo_format_byte("occupancy") 889 ergo_format_byte("allocation request") 890 ergo_format_byte_perc("threshold") 891 ergo_format_str("source"), 892 cur_used_bytes, 893 alloc_byte_size, 894 marking_initiating_used_threshold, 895 (double) InitiatingHeapOccupancyPercent, 896 source); 897 } 898 } 899 900 return false; 901 } 902 903 // Anything below that is considered to be zero 904 #define MIN_TIMER_GRANULARITY 0.0000001 905 906 void G1CollectorPolicy::record_collection_pause_end(double pause_time_ms) { 907 double end_time_sec = os::elapsedTime(); 908 assert(_cur_collection_pause_used_regions_at_start >= cset_region_length(), 909 "otherwise, the subtraction below does not make sense"); 910 size_t rs_size = 911 _cur_collection_pause_used_regions_at_start - cset_region_length(); 912 size_t cur_used_bytes = _g1->used(); 913 assert(cur_used_bytes == _g1->recalculate_used(), "It should!"); 914 bool last_pause_included_initial_mark = false; 915 bool update_stats = !_g1->evacuation_failed(); 916 917 #ifndef PRODUCT 918 if (G1YoungSurvRateVerbose) { 919 gclog_or_tty->print_cr(""); 920 _short_lived_surv_rate_group->print(); 921 // do that for any other surv rate groups too 922 } 923 #endif // PRODUCT 924 925 last_pause_included_initial_mark = during_initial_mark_pause(); 926 if (last_pause_included_initial_mark) { 927 record_concurrent_mark_init_end(0.0); 928 } else if (!_last_young_gc && need_to_start_conc_mark("end of GC")) { 929 // Note: this might have already been set, if during the last 930 // pause we decided to start a cycle but at the beginning of 931 // this pause we decided to postpone it. That's OK. 932 set_initiate_conc_mark_if_possible(); 933 } 934 935 _mmu_tracker->add_pause(end_time_sec - pause_time_ms/1000.0, 936 end_time_sec, false); 937 938 size_t freed_bytes = 939 _cur_collection_pause_used_at_start_bytes - cur_used_bytes; 940 size_t surviving_bytes = _collection_set_bytes_used_before - freed_bytes; 941 942 double survival_fraction = 943 (double)surviving_bytes/ 944 (double)_collection_set_bytes_used_before; 945 946 if (update_stats) { 947 _trace_gen0_time_data.record_end_collection(pause_time_ms, phase_times()); 948 // this is where we update the allocation rate of the application 949 double app_time_ms = 950 (phase_times()->cur_collection_start_sec() * 1000.0 - _prev_collection_pause_end_ms); 951 if (app_time_ms < MIN_TIMER_GRANULARITY) { 952 // This usually happens due to the timer not having the required 953 // granularity. Some Linuxes are the usual culprits. 954 // We'll just set it to something (arbitrarily) small. 955 app_time_ms = 1.0; 956 } 957 // We maintain the invariant that all objects allocated by mutator 958 // threads will be allocated out of eden regions. So, we can use 959 // the eden region number allocated since the previous GC to 960 // calculate the application's allocate rate. The only exception 961 // to that is humongous objects that are allocated separately. But 962 // given that humongous object allocations do not really affect 963 // either the pause's duration nor when the next pause will take 964 // place we can safely ignore them here. 965 uint regions_allocated = eden_cset_region_length(); 966 double alloc_rate_ms = (double) regions_allocated / app_time_ms; 967 _alloc_rate_ms_seq->add(alloc_rate_ms); 968 969 double interval_ms = 970 (end_time_sec - _recent_prev_end_times_for_all_gcs_sec->oldest()) * 1000.0; 971 update_recent_gc_times(end_time_sec, pause_time_ms); 972 _recent_avg_pause_time_ratio = _recent_gc_times_ms->sum()/interval_ms; 973 if (recent_avg_pause_time_ratio() < 0.0 || 974 (recent_avg_pause_time_ratio() - 1.0 > 0.0)) { 975 #ifndef PRODUCT 976 // Dump info to allow post-facto debugging 977 gclog_or_tty->print_cr("recent_avg_pause_time_ratio() out of bounds"); 978 gclog_or_tty->print_cr("-------------------------------------------"); 979 gclog_or_tty->print_cr("Recent GC Times (ms):"); 980 _recent_gc_times_ms->dump(); 981 gclog_or_tty->print_cr("(End Time=%3.3f) Recent GC End Times (s):", end_time_sec); 982 _recent_prev_end_times_for_all_gcs_sec->dump(); 983 gclog_or_tty->print_cr("GC = %3.3f, Interval = %3.3f, Ratio = %3.3f", 984 _recent_gc_times_ms->sum(), interval_ms, recent_avg_pause_time_ratio()); 985 // In debug mode, terminate the JVM if the user wants to debug at this point. 986 assert(!G1FailOnFPError, "Debugging data for CR 6898948 has been dumped above"); 987 #endif // !PRODUCT 988 // Clip ratio between 0.0 and 1.0, and continue. This will be fixed in 989 // CR 6902692 by redoing the manner in which the ratio is incrementally computed. 990 if (_recent_avg_pause_time_ratio < 0.0) { 991 _recent_avg_pause_time_ratio = 0.0; 992 } else { 993 assert(_recent_avg_pause_time_ratio - 1.0 > 0.0, "Ctl-point invariant"); 994 _recent_avg_pause_time_ratio = 1.0; 995 } 996 } 997 } 998 bool new_in_marking_window = _in_marking_window; 999 bool new_in_marking_window_im = false; 1000 if (during_initial_mark_pause()) { 1001 new_in_marking_window = true; 1002 new_in_marking_window_im = true; 1003 } 1004 1005 if (_last_young_gc) { 1006 // This is supposed to to be the "last young GC" before we start 1007 // doing mixed GCs. Here we decide whether to start mixed GCs or not. 1008 1009 if (!last_pause_included_initial_mark) { 1010 if (next_gc_should_be_mixed("start mixed GCs", 1011 "do not start mixed GCs")) { 1012 set_gcs_are_young(false); 1013 } 1014 } else { 1015 ergo_verbose0(ErgoMixedGCs, 1016 "do not start mixed GCs", 1017 ergo_format_reason("concurrent cycle is about to start")); 1018 } 1019 _last_young_gc = false; 1020 } 1021 1022 if (!_last_gc_was_young) { 1023 // This is a mixed GC. Here we decide whether to continue doing 1024 // mixed GCs or not. 1025 1026 if (!next_gc_should_be_mixed("continue mixed GCs", 1027 "do not continue mixed GCs")) { 1028 set_gcs_are_young(true); 1029 } 1030 } 1031 1032 _short_lived_surv_rate_group->start_adding_regions(); 1033 // do that for any other surv rate groupsx 1034 1035 if (update_stats) { 1036 double cost_per_card_ms = 0.0; 1037 if (_pending_cards > 0) { 1038 cost_per_card_ms = phase_times()->average_last_update_rs_time() / (double) _pending_cards; 1039 _cost_per_card_ms_seq->add(cost_per_card_ms); 1040 } 1041 1042 size_t cards_scanned = _g1->cards_scanned(); 1043 1044 double cost_per_entry_ms = 0.0; 1045 if (cards_scanned > 10) { 1046 cost_per_entry_ms = phase_times()->average_last_scan_rs_time() / (double) cards_scanned; 1047 if (_last_gc_was_young) { 1048 _cost_per_entry_ms_seq->add(cost_per_entry_ms); 1049 } else { 1050 _mixed_cost_per_entry_ms_seq->add(cost_per_entry_ms); 1051 } 1052 } 1053 1054 if (_max_rs_lengths > 0) { 1055 double cards_per_entry_ratio = 1056 (double) cards_scanned / (double) _max_rs_lengths; 1057 if (_last_gc_was_young) { 1058 _young_cards_per_entry_ratio_seq->add(cards_per_entry_ratio); 1059 } else { 1060 _mixed_cards_per_entry_ratio_seq->add(cards_per_entry_ratio); 1061 } 1062 } 1063 1064 // This is defensive. For a while _max_rs_lengths could get 1065 // smaller than _recorded_rs_lengths which was causing 1066 // rs_length_diff to get very large and mess up the RSet length 1067 // predictions. The reason was unsafe concurrent updates to the 1068 // _inc_cset_recorded_rs_lengths field which the code below guards 1069 // against (see CR 7118202). This bug has now been fixed (see CR 1070 // 7119027). However, I'm still worried that 1071 // _inc_cset_recorded_rs_lengths might still end up somewhat 1072 // inaccurate. The concurrent refinement thread calculates an 1073 // RSet's length concurrently with other CR threads updating it 1074 // which might cause it to calculate the length incorrectly (if, 1075 // say, it's in mid-coarsening). So I'll leave in the defensive 1076 // conditional below just in case. 1077 size_t rs_length_diff = 0; 1078 if (_max_rs_lengths > _recorded_rs_lengths) { 1079 rs_length_diff = _max_rs_lengths - _recorded_rs_lengths; 1080 } 1081 _rs_length_diff_seq->add((double) rs_length_diff); 1082 1083 size_t copied_bytes = surviving_bytes; 1084 double cost_per_byte_ms = 0.0; 1085 if (copied_bytes > 0) { 1086 cost_per_byte_ms = phase_times()->average_last_obj_copy_time() / (double) copied_bytes; 1087 if (_in_marking_window) { 1088 _cost_per_byte_ms_during_cm_seq->add(cost_per_byte_ms); 1089 } else { 1090 _cost_per_byte_ms_seq->add(cost_per_byte_ms); 1091 } 1092 } 1093 1094 double all_other_time_ms = pause_time_ms - 1095 (phase_times()->average_last_update_rs_time() + phase_times()->average_last_scan_rs_time() 1096 + phase_times()->average_last_obj_copy_time() + phase_times()->average_last_termination_time()); 1097 1098 double young_other_time_ms = 0.0; 1099 if (young_cset_region_length() > 0) { 1100 young_other_time_ms = 1101 phase_times()->young_cset_choice_time_ms() + 1102 phase_times()->young_free_cset_time_ms(); 1103 _young_other_cost_per_region_ms_seq->add(young_other_time_ms / 1104 (double) young_cset_region_length()); 1105 } 1106 double non_young_other_time_ms = 0.0; 1107 if (old_cset_region_length() > 0) { 1108 non_young_other_time_ms = 1109 phase_times()->non_young_cset_choice_time_ms() + 1110 phase_times()->non_young_free_cset_time_ms(); 1111 1112 _non_young_other_cost_per_region_ms_seq->add(non_young_other_time_ms / 1113 (double) old_cset_region_length()); 1114 } 1115 1116 double constant_other_time_ms = all_other_time_ms - 1117 (young_other_time_ms + non_young_other_time_ms); 1118 _constant_other_time_ms_seq->add(constant_other_time_ms); 1119 1120 double survival_ratio = 0.0; 1121 if (_collection_set_bytes_used_before > 0) { 1122 survival_ratio = (double) _bytes_copied_during_gc / 1123 (double) _collection_set_bytes_used_before; 1124 } 1125 1126 _pending_cards_seq->add((double) _pending_cards); 1127 _rs_lengths_seq->add((double) _max_rs_lengths); 1128 } 1129 1130 _in_marking_window = new_in_marking_window; 1131 _in_marking_window_im = new_in_marking_window_im; 1132 _free_regions_at_end_of_collection = _g1->free_regions(); 1133 update_young_list_target_length(); 1134 1135 // Note that _mmu_tracker->max_gc_time() returns the time in seconds. 1136 double update_rs_time_goal_ms = _mmu_tracker->max_gc_time() * MILLIUNITS * G1RSetUpdatingPauseTimePercent / 100.0; 1137 adjust_concurrent_refinement(phase_times()->average_last_update_rs_time(), 1138 phase_times()->sum_last_update_rs_processed_buffers(), update_rs_time_goal_ms); 1139 1140 _collectionSetChooser->verify(); 1141 } 1142 1143 #define EXT_SIZE_FORMAT "%.1f%s" 1144 #define EXT_SIZE_PARAMS(bytes) \ 1145 byte_size_in_proper_unit((double)(bytes)), \ 1146 proper_unit_for_byte_size((bytes)) 1147 1148 void G1CollectorPolicy::print_heap_transition() { 1149 _g1->print_size_transition(gclog_or_tty, 1150 _cur_collection_pause_used_at_start_bytes, _g1->used(), _g1->capacity()); 1151 } 1152 1153 void G1CollectorPolicy::print_detailed_heap_transition() { 1154 YoungList* young_list = _g1->young_list(); 1155 size_t eden_bytes = young_list->eden_used_bytes(); 1156 size_t survivor_bytes = young_list->survivor_used_bytes(); 1157 size_t used_before_gc = _cur_collection_pause_used_at_start_bytes; 1158 size_t used = _g1->used(); 1159 size_t capacity = _g1->capacity(); 1160 size_t eden_capacity = 1161 (_young_list_target_length * HeapRegion::GrainBytes) - survivor_bytes; 1162 1163 gclog_or_tty->print_cr( 1164 " [Eden: "EXT_SIZE_FORMAT"("EXT_SIZE_FORMAT")->"EXT_SIZE_FORMAT"("EXT_SIZE_FORMAT") " 1165 "Survivors: "EXT_SIZE_FORMAT"->"EXT_SIZE_FORMAT" " 1166 "Heap: "EXT_SIZE_FORMAT"("EXT_SIZE_FORMAT")->" 1167 EXT_SIZE_FORMAT"("EXT_SIZE_FORMAT")]", 1168 EXT_SIZE_PARAMS(_eden_bytes_before_gc), 1169 EXT_SIZE_PARAMS(_prev_eden_capacity), 1170 EXT_SIZE_PARAMS(eden_bytes), 1171 EXT_SIZE_PARAMS(eden_capacity), 1172 EXT_SIZE_PARAMS(_survivor_bytes_before_gc), 1173 EXT_SIZE_PARAMS(survivor_bytes), 1174 EXT_SIZE_PARAMS(used_before_gc), 1175 EXT_SIZE_PARAMS(_capacity_before_gc), 1176 EXT_SIZE_PARAMS(used), 1177 EXT_SIZE_PARAMS(capacity)); 1178 1179 _prev_eden_capacity = eden_capacity; 1180 } 1181 1182 void G1CollectorPolicy::adjust_concurrent_refinement(double update_rs_time, 1183 double update_rs_processed_buffers, 1184 double goal_ms) { 1185 DirtyCardQueueSet& dcqs = JavaThread::dirty_card_queue_set(); 1186 ConcurrentG1Refine *cg1r = G1CollectedHeap::heap()->concurrent_g1_refine(); 1187 1188 if (G1UseAdaptiveConcRefinement) { 1189 const int k_gy = 3, k_gr = 6; 1190 const double inc_k = 1.1, dec_k = 0.9; 1191 1192 int g = cg1r->green_zone(); 1193 if (update_rs_time > goal_ms) { 1194 g = (int)(g * dec_k); // Can become 0, that's OK. That would mean a mutator-only processing. 1195 } else { 1196 if (update_rs_time < goal_ms && update_rs_processed_buffers > g) { 1197 g = (int)MAX2(g * inc_k, g + 1.0); 1198 } 1199 } 1200 // Change the refinement threads params 1201 cg1r->set_green_zone(g); 1202 cg1r->set_yellow_zone(g * k_gy); 1203 cg1r->set_red_zone(g * k_gr); 1204 cg1r->reinitialize_threads(); 1205 1206 int processing_threshold_delta = MAX2((int)(cg1r->green_zone() * sigma()), 1); 1207 int processing_threshold = MIN2(cg1r->green_zone() + processing_threshold_delta, 1208 cg1r->yellow_zone()); 1209 // Change the barrier params 1210 dcqs.set_process_completed_threshold(processing_threshold); 1211 dcqs.set_max_completed_queue(cg1r->red_zone()); 1212 } 1213 1214 int curr_queue_size = dcqs.completed_buffers_num(); 1215 if (curr_queue_size >= cg1r->yellow_zone()) { 1216 dcqs.set_completed_queue_padding(curr_queue_size); 1217 } else { 1218 dcqs.set_completed_queue_padding(0); 1219 } 1220 dcqs.notify_if_necessary(); 1221 } 1222 1223 double 1224 G1CollectorPolicy::predict_base_elapsed_time_ms(size_t pending_cards, 1225 size_t scanned_cards) { 1226 return 1227 predict_rs_update_time_ms(pending_cards) + 1228 predict_rs_scan_time_ms(scanned_cards) + 1229 predict_constant_other_time_ms(); 1230 } 1231 1232 double 1233 G1CollectorPolicy::predict_base_elapsed_time_ms(size_t pending_cards) { 1234 size_t rs_length = predict_rs_length_diff(); 1235 size_t card_num; 1236 if (gcs_are_young()) { 1237 card_num = predict_young_card_num(rs_length); 1238 } else { 1239 card_num = predict_non_young_card_num(rs_length); 1240 } 1241 return predict_base_elapsed_time_ms(pending_cards, card_num); 1242 } 1243 1244 size_t G1CollectorPolicy::predict_bytes_to_copy(HeapRegion* hr) { 1245 size_t bytes_to_copy; 1246 if (hr->is_marked()) 1247 bytes_to_copy = hr->max_live_bytes(); 1248 else { 1249 assert(hr->is_young() && hr->age_in_surv_rate_group() != -1, "invariant"); 1250 int age = hr->age_in_surv_rate_group(); 1251 double yg_surv_rate = predict_yg_surv_rate(age, hr->surv_rate_group()); 1252 bytes_to_copy = (size_t) ((double) hr->used() * yg_surv_rate); 1253 } 1254 return bytes_to_copy; 1255 } 1256 1257 double 1258 G1CollectorPolicy::predict_region_elapsed_time_ms(HeapRegion* hr, 1259 bool for_young_gc) { 1260 size_t rs_length = hr->rem_set()->occupied(); 1261 size_t card_num; 1262 1263 // Predicting the number of cards is based on which type of GC 1264 // we're predicting for. 1265 if (for_young_gc) { 1266 card_num = predict_young_card_num(rs_length); 1267 } else { 1268 card_num = predict_non_young_card_num(rs_length); 1269 } 1270 size_t bytes_to_copy = predict_bytes_to_copy(hr); 1271 1272 double region_elapsed_time_ms = 1273 predict_rs_scan_time_ms(card_num) + 1274 predict_object_copy_time_ms(bytes_to_copy); 1275 1276 // The prediction of the "other" time for this region is based 1277 // upon the region type and NOT the GC type. 1278 if (hr->is_young()) { 1279 region_elapsed_time_ms += predict_young_other_time_ms(1); 1280 } else { 1281 region_elapsed_time_ms += predict_non_young_other_time_ms(1); 1282 } 1283 return region_elapsed_time_ms; 1284 } 1285 1286 void 1287 G1CollectorPolicy::init_cset_region_lengths(uint eden_cset_region_length, 1288 uint survivor_cset_region_length) { 1289 _eden_cset_region_length = eden_cset_region_length; 1290 _survivor_cset_region_length = survivor_cset_region_length; 1291 _old_cset_region_length = 0; 1292 } 1293 1294 void G1CollectorPolicy::set_recorded_rs_lengths(size_t rs_lengths) { 1295 _recorded_rs_lengths = rs_lengths; 1296 } 1297 1298 void G1CollectorPolicy::update_recent_gc_times(double end_time_sec, 1299 double elapsed_ms) { 1300 _recent_gc_times_ms->add(elapsed_ms); 1301 _recent_prev_end_times_for_all_gcs_sec->add(end_time_sec); 1302 _prev_collection_pause_end_ms = end_time_sec * 1000.0; 1303 } 1304 1305 size_t G1CollectorPolicy::expansion_amount() { 1306 double recent_gc_overhead = recent_avg_pause_time_ratio() * 100.0; 1307 double threshold = _gc_overhead_perc; 1308 if (recent_gc_overhead > threshold) { 1309 // We will double the existing space, or take 1310 // G1ExpandByPercentOfAvailable % of the available expansion 1311 // space, whichever is smaller, bounded below by a minimum 1312 // expansion (unless that's all that's left.) 1313 const size_t min_expand_bytes = 1*M; 1314 size_t reserved_bytes = _g1->max_capacity(); 1315 size_t committed_bytes = _g1->capacity(); 1316 size_t uncommitted_bytes = reserved_bytes - committed_bytes; 1317 size_t expand_bytes; 1318 size_t expand_bytes_via_pct = 1319 uncommitted_bytes * G1ExpandByPercentOfAvailable / 100; 1320 expand_bytes = MIN2(expand_bytes_via_pct, committed_bytes); 1321 expand_bytes = MAX2(expand_bytes, min_expand_bytes); 1322 expand_bytes = MIN2(expand_bytes, uncommitted_bytes); 1323 1324 ergo_verbose5(ErgoHeapSizing, 1325 "attempt heap expansion", 1326 ergo_format_reason("recent GC overhead higher than " 1327 "threshold after GC") 1328 ergo_format_perc("recent GC overhead") 1329 ergo_format_perc("threshold") 1330 ergo_format_byte("uncommitted") 1331 ergo_format_byte_perc("calculated expansion amount"), 1332 recent_gc_overhead, threshold, 1333 uncommitted_bytes, 1334 expand_bytes_via_pct, (double) G1ExpandByPercentOfAvailable); 1335 1336 return expand_bytes; 1337 } else { 1338 return 0; 1339 } 1340 } 1341 1342 void G1CollectorPolicy::print_tracing_info() const { 1343 _trace_gen0_time_data.print(); 1344 _trace_gen1_time_data.print(); 1345 } 1346 1347 void G1CollectorPolicy::print_yg_surv_rate_info() const { 1348 #ifndef PRODUCT 1349 _short_lived_surv_rate_group->print_surv_rate_summary(); 1350 // add this call for any other surv rate groups 1351 #endif // PRODUCT 1352 } 1353 1354 #ifndef PRODUCT 1355 // for debugging, bit of a hack... 1356 char* 1357 region_num_to_mbs(int length) { 1358 static char buffer[64]; 1359 double bytes = (double) (length * HeapRegion::GrainBytes); 1360 double mbs = bytes / (double) (1024 * 1024); 1361 sprintf(buffer, "%7.2lfMB", mbs); 1362 return buffer; 1363 } 1364 #endif // PRODUCT 1365 1366 uint G1CollectorPolicy::max_regions(int purpose) { 1367 switch (purpose) { 1368 case GCAllocForSurvived: 1369 return _max_survivor_regions; 1370 case GCAllocForTenured: 1371 return REGIONS_UNLIMITED; 1372 default: 1373 ShouldNotReachHere(); 1374 return REGIONS_UNLIMITED; 1375 }; 1376 } 1377 1378 void G1CollectorPolicy::update_max_gc_locker_expansion() { 1379 uint expansion_region_num = 0; 1380 if (GCLockerEdenExpansionPercent > 0) { 1381 double perc = (double) GCLockerEdenExpansionPercent / 100.0; 1382 double expansion_region_num_d = perc * (double) _young_list_target_length; 1383 // We use ceiling so that if expansion_region_num_d is > 0.0 (but 1384 // less than 1.0) we'll get 1. 1385 expansion_region_num = (uint) ceil(expansion_region_num_d); 1386 } else { 1387 assert(expansion_region_num == 0, "sanity"); 1388 } 1389 _young_list_max_length = _young_list_target_length + expansion_region_num; 1390 assert(_young_list_target_length <= _young_list_max_length, "post-condition"); 1391 } 1392 1393 // Calculates survivor space parameters. 1394 void G1CollectorPolicy::update_survivors_policy() { 1395 double max_survivor_regions_d = 1396 (double) _young_list_target_length / (double) SurvivorRatio; 1397 // We use ceiling so that if max_survivor_regions_d is > 0.0 (but 1398 // smaller than 1.0) we'll get 1. 1399 _max_survivor_regions = (uint) ceil(max_survivor_regions_d); 1400 1401 _tenuring_threshold = _survivors_age_table.compute_tenuring_threshold( 1402 HeapRegion::GrainWords * _max_survivor_regions); 1403 } 1404 1405 bool G1CollectorPolicy::force_initial_mark_if_outside_cycle( 1406 GCCause::Cause gc_cause) { 1407 bool during_cycle = _g1->concurrent_mark()->cmThread()->during_cycle(); 1408 if (!during_cycle) { 1409 ergo_verbose1(ErgoConcCycles, 1410 "request concurrent cycle initiation", 1411 ergo_format_reason("requested by GC cause") 1412 ergo_format_str("GC cause"), 1413 GCCause::to_string(gc_cause)); 1414 set_initiate_conc_mark_if_possible(); 1415 return true; 1416 } else { 1417 ergo_verbose1(ErgoConcCycles, 1418 "do not request concurrent cycle initiation", 1419 ergo_format_reason("concurrent cycle already in progress") 1420 ergo_format_str("GC cause"), 1421 GCCause::to_string(gc_cause)); 1422 return false; 1423 } 1424 } 1425 1426 void 1427 G1CollectorPolicy::decide_on_conc_mark_initiation() { 1428 // We are about to decide on whether this pause will be an 1429 // initial-mark pause. 1430 1431 // First, during_initial_mark_pause() should not be already set. We 1432 // will set it here if we have to. However, it should be cleared by 1433 // the end of the pause (it's only set for the duration of an 1434 // initial-mark pause). 1435 assert(!during_initial_mark_pause(), "pre-condition"); 1436 1437 if (initiate_conc_mark_if_possible()) { 1438 // We had noticed on a previous pause that the heap occupancy has 1439 // gone over the initiating threshold and we should start a 1440 // concurrent marking cycle. So we might initiate one. 1441 1442 bool during_cycle = _g1->concurrent_mark()->cmThread()->during_cycle(); 1443 if (!during_cycle) { 1444 // The concurrent marking thread is not "during a cycle", i.e., 1445 // it has completed the last one. So we can go ahead and 1446 // initiate a new cycle. 1447 1448 set_during_initial_mark_pause(); 1449 // We do not allow mixed GCs during marking. 1450 if (!gcs_are_young()) { 1451 set_gcs_are_young(true); 1452 ergo_verbose0(ErgoMixedGCs, 1453 "end mixed GCs", 1454 ergo_format_reason("concurrent cycle is about to start")); 1455 } 1456 1457 // And we can now clear initiate_conc_mark_if_possible() as 1458 // we've already acted on it. 1459 clear_initiate_conc_mark_if_possible(); 1460 1461 ergo_verbose0(ErgoConcCycles, 1462 "initiate concurrent cycle", 1463 ergo_format_reason("concurrent cycle initiation requested")); 1464 } else { 1465 // The concurrent marking thread is still finishing up the 1466 // previous cycle. If we start one right now the two cycles 1467 // overlap. In particular, the concurrent marking thread might 1468 // be in the process of clearing the next marking bitmap (which 1469 // we will use for the next cycle if we start one). Starting a 1470 // cycle now will be bad given that parts of the marking 1471 // information might get cleared by the marking thread. And we 1472 // cannot wait for the marking thread to finish the cycle as it 1473 // periodically yields while clearing the next marking bitmap 1474 // and, if it's in a yield point, it's waiting for us to 1475 // finish. So, at this point we will not start a cycle and we'll 1476 // let the concurrent marking thread complete the last one. 1477 ergo_verbose0(ErgoConcCycles, 1478 "do not initiate concurrent cycle", 1479 ergo_format_reason("concurrent cycle already in progress")); 1480 } 1481 } 1482 } 1483 1484 class KnownGarbageClosure: public HeapRegionClosure { 1485 G1CollectedHeap* _g1h; 1486 CollectionSetChooser* _hrSorted; 1487 1488 public: 1489 KnownGarbageClosure(CollectionSetChooser* hrSorted) : 1490 _g1h(G1CollectedHeap::heap()), _hrSorted(hrSorted) { } 1491 1492 bool doHeapRegion(HeapRegion* r) { 1493 // We only include humongous regions in collection 1494 // sets when concurrent mark shows that their contained object is 1495 // unreachable. 1496 1497 // Do we have any marking information for this region? 1498 if (r->is_marked()) { 1499 // We will skip any region that's currently used as an old GC 1500 // alloc region (we should not consider those for collection 1501 // before we fill them up). 1502 if (_hrSorted->should_add(r) && !_g1h->is_old_gc_alloc_region(r)) { 1503 _hrSorted->add_region(r); 1504 } 1505 } 1506 return false; 1507 } 1508 }; 1509 1510 class ParKnownGarbageHRClosure: public HeapRegionClosure { 1511 G1CollectedHeap* _g1h; 1512 CSetChooserParUpdater _cset_updater; 1513 1514 public: 1515 ParKnownGarbageHRClosure(CollectionSetChooser* hrSorted, 1516 uint chunk_size) : 1517 _g1h(G1CollectedHeap::heap()), 1518 _cset_updater(hrSorted, true /* parallel */, chunk_size) { } 1519 1520 bool doHeapRegion(HeapRegion* r) { 1521 // Do we have any marking information for this region? 1522 if (r->is_marked()) { 1523 // We will skip any region that's currently used as an old GC 1524 // alloc region (we should not consider those for collection 1525 // before we fill them up). 1526 if (_cset_updater.should_add(r) && !_g1h->is_old_gc_alloc_region(r)) { 1527 _cset_updater.add_region(r); 1528 } 1529 } 1530 return false; 1531 } 1532 }; 1533 1534 class ParKnownGarbageTask: public AbstractGangTask { 1535 CollectionSetChooser* _hrSorted; 1536 uint _chunk_size; 1537 G1CollectedHeap* _g1; 1538 public: 1539 ParKnownGarbageTask(CollectionSetChooser* hrSorted, uint chunk_size) : 1540 AbstractGangTask("ParKnownGarbageTask"), 1541 _hrSorted(hrSorted), _chunk_size(chunk_size), 1542 _g1(G1CollectedHeap::heap()) { } 1543 1544 void work(uint worker_id) { 1545 ParKnownGarbageHRClosure parKnownGarbageCl(_hrSorted, _chunk_size); 1546 1547 // Back to zero for the claim value. 1548 _g1->heap_region_par_iterate_chunked(&parKnownGarbageCl, worker_id, 1549 _g1->workers()->active_workers(), 1550 HeapRegion::InitialClaimValue); 1551 } 1552 }; 1553 1554 void 1555 G1CollectorPolicy::record_concurrent_mark_cleanup_end(int no_of_gc_threads) { 1556 _collectionSetChooser->clear(); 1557 1558 uint region_num = _g1->n_regions(); 1559 if (G1CollectedHeap::use_parallel_gc_threads()) { 1560 const uint OverpartitionFactor = 4; 1561 uint WorkUnit; 1562 // The use of MinChunkSize = 8 in the original code 1563 // causes some assertion failures when the total number of 1564 // region is less than 8. The code here tries to fix that. 1565 // Should the original code also be fixed? 1566 if (no_of_gc_threads > 0) { 1567 const uint MinWorkUnit = MAX2(region_num / no_of_gc_threads, 1U); 1568 WorkUnit = MAX2(region_num / (no_of_gc_threads * OverpartitionFactor), 1569 MinWorkUnit); 1570 } else { 1571 assert(no_of_gc_threads > 0, 1572 "The active gc workers should be greater than 0"); 1573 // In a product build do something reasonable to avoid a crash. 1574 const uint MinWorkUnit = MAX2(region_num / (uint) ParallelGCThreads, 1U); 1575 WorkUnit = 1576 MAX2(region_num / (uint) (ParallelGCThreads * OverpartitionFactor), 1577 MinWorkUnit); 1578 } 1579 _collectionSetChooser->prepare_for_par_region_addition(_g1->n_regions(), 1580 WorkUnit); 1581 ParKnownGarbageTask parKnownGarbageTask(_collectionSetChooser, 1582 (int) WorkUnit); 1583 _g1->workers()->run_task(&parKnownGarbageTask); 1584 1585 assert(_g1->check_heap_region_claim_values(HeapRegion::InitialClaimValue), 1586 "sanity check"); 1587 } else { 1588 KnownGarbageClosure knownGarbagecl(_collectionSetChooser); 1589 _g1->heap_region_iterate(&knownGarbagecl); 1590 } 1591 1592 _collectionSetChooser->sort_regions(); 1593 1594 double end_sec = os::elapsedTime(); 1595 double elapsed_time_ms = (end_sec - _mark_cleanup_start_sec) * 1000.0; 1596 _concurrent_mark_cleanup_times_ms->add(elapsed_time_ms); 1597 _cur_mark_stop_world_time_ms += elapsed_time_ms; 1598 _prev_collection_pause_end_ms += elapsed_time_ms; 1599 _mmu_tracker->add_pause(_mark_cleanup_start_sec, end_sec, true); 1600 } 1601 1602 // Add the heap region at the head of the non-incremental collection set 1603 void G1CollectorPolicy::add_old_region_to_cset(HeapRegion* hr) { 1604 assert(_inc_cset_build_state == Active, "Precondition"); 1605 assert(!hr->is_young(), "non-incremental add of young region"); 1606 1607 assert(!hr->in_collection_set(), "should not already be in the CSet"); 1608 hr->set_in_collection_set(true); 1609 hr->set_next_in_collection_set(_collection_set); 1610 _collection_set = hr; 1611 _collection_set_bytes_used_before += hr->used(); 1612 _g1->register_region_with_in_cset_fast_test(hr); 1613 size_t rs_length = hr->rem_set()->occupied(); 1614 _recorded_rs_lengths += rs_length; 1615 _old_cset_region_length += 1; 1616 } 1617 1618 // Initialize the per-collection-set information 1619 void G1CollectorPolicy::start_incremental_cset_building() { 1620 assert(_inc_cset_build_state == Inactive, "Precondition"); 1621 1622 _inc_cset_head = NULL; 1623 _inc_cset_tail = NULL; 1624 _inc_cset_bytes_used_before = 0; 1625 1626 _inc_cset_max_finger = 0; 1627 _inc_cset_recorded_rs_lengths = 0; 1628 _inc_cset_recorded_rs_lengths_diffs = 0; 1629 _inc_cset_predicted_elapsed_time_ms = 0.0; 1630 _inc_cset_predicted_elapsed_time_ms_diffs = 0.0; 1631 _inc_cset_build_state = Active; 1632 } 1633 1634 void G1CollectorPolicy::finalize_incremental_cset_building() { 1635 assert(_inc_cset_build_state == Active, "Precondition"); 1636 assert(SafepointSynchronize::is_at_safepoint(), "should be at a safepoint"); 1637 1638 // The two "main" fields, _inc_cset_recorded_rs_lengths and 1639 // _inc_cset_predicted_elapsed_time_ms, are updated by the thread 1640 // that adds a new region to the CSet. Further updates by the 1641 // concurrent refinement thread that samples the young RSet lengths 1642 // are accumulated in the *_diffs fields. Here we add the diffs to 1643 // the "main" fields. 1644 1645 if (_inc_cset_recorded_rs_lengths_diffs >= 0) { 1646 _inc_cset_recorded_rs_lengths += _inc_cset_recorded_rs_lengths_diffs; 1647 } else { 1648 // This is defensive. The diff should in theory be always positive 1649 // as RSets can only grow between GCs. However, given that we 1650 // sample their size concurrently with other threads updating them 1651 // it's possible that we might get the wrong size back, which 1652 // could make the calculations somewhat inaccurate. 1653 size_t diffs = (size_t) (-_inc_cset_recorded_rs_lengths_diffs); 1654 if (_inc_cset_recorded_rs_lengths >= diffs) { 1655 _inc_cset_recorded_rs_lengths -= diffs; 1656 } else { 1657 _inc_cset_recorded_rs_lengths = 0; 1658 } 1659 } 1660 _inc_cset_predicted_elapsed_time_ms += 1661 _inc_cset_predicted_elapsed_time_ms_diffs; 1662 1663 _inc_cset_recorded_rs_lengths_diffs = 0; 1664 _inc_cset_predicted_elapsed_time_ms_diffs = 0.0; 1665 } 1666 1667 void G1CollectorPolicy::add_to_incremental_cset_info(HeapRegion* hr, size_t rs_length) { 1668 // This routine is used when: 1669 // * adding survivor regions to the incremental cset at the end of an 1670 // evacuation pause, 1671 // * adding the current allocation region to the incremental cset 1672 // when it is retired, and 1673 // * updating existing policy information for a region in the 1674 // incremental cset via young list RSet sampling. 1675 // Therefore this routine may be called at a safepoint by the 1676 // VM thread, or in-between safepoints by mutator threads (when 1677 // retiring the current allocation region) or a concurrent 1678 // refine thread (RSet sampling). 1679 1680 double region_elapsed_time_ms = predict_region_elapsed_time_ms(hr, gcs_are_young()); 1681 size_t used_bytes = hr->used(); 1682 _inc_cset_recorded_rs_lengths += rs_length; 1683 _inc_cset_predicted_elapsed_time_ms += region_elapsed_time_ms; 1684 _inc_cset_bytes_used_before += used_bytes; 1685 1686 // Cache the values we have added to the aggregated informtion 1687 // in the heap region in case we have to remove this region from 1688 // the incremental collection set, or it is updated by the 1689 // rset sampling code 1690 hr->set_recorded_rs_length(rs_length); 1691 hr->set_predicted_elapsed_time_ms(region_elapsed_time_ms); 1692 } 1693 1694 void G1CollectorPolicy::update_incremental_cset_info(HeapRegion* hr, 1695 size_t new_rs_length) { 1696 // Update the CSet information that is dependent on the new RS length 1697 assert(hr->is_young(), "Precondition"); 1698 assert(!SafepointSynchronize::is_at_safepoint(), 1699 "should not be at a safepoint"); 1700 1701 // We could have updated _inc_cset_recorded_rs_lengths and 1702 // _inc_cset_predicted_elapsed_time_ms directly but we'd need to do 1703 // that atomically, as this code is executed by a concurrent 1704 // refinement thread, potentially concurrently with a mutator thread 1705 // allocating a new region and also updating the same fields. To 1706 // avoid the atomic operations we accumulate these updates on two 1707 // separate fields (*_diffs) and we'll just add them to the "main" 1708 // fields at the start of a GC. 1709 1710 ssize_t old_rs_length = (ssize_t) hr->recorded_rs_length(); 1711 ssize_t rs_lengths_diff = (ssize_t) new_rs_length - old_rs_length; 1712 _inc_cset_recorded_rs_lengths_diffs += rs_lengths_diff; 1713 1714 double old_elapsed_time_ms = hr->predicted_elapsed_time_ms(); 1715 double new_region_elapsed_time_ms = predict_region_elapsed_time_ms(hr, gcs_are_young()); 1716 double elapsed_ms_diff = new_region_elapsed_time_ms - old_elapsed_time_ms; 1717 _inc_cset_predicted_elapsed_time_ms_diffs += elapsed_ms_diff; 1718 1719 hr->set_recorded_rs_length(new_rs_length); 1720 hr->set_predicted_elapsed_time_ms(new_region_elapsed_time_ms); 1721 } 1722 1723 void G1CollectorPolicy::add_region_to_incremental_cset_common(HeapRegion* hr) { 1724 assert(hr->is_young(), "invariant"); 1725 assert(hr->young_index_in_cset() > -1, "should have already been set"); 1726 assert(_inc_cset_build_state == Active, "Precondition"); 1727 1728 // We need to clear and set the cached recorded/cached collection set 1729 // information in the heap region here (before the region gets added 1730 // to the collection set). An individual heap region's cached values 1731 // are calculated, aggregated with the policy collection set info, 1732 // and cached in the heap region here (initially) and (subsequently) 1733 // by the Young List sampling code. 1734 1735 size_t rs_length = hr->rem_set()->occupied(); 1736 add_to_incremental_cset_info(hr, rs_length); 1737 1738 HeapWord* hr_end = hr->end(); 1739 _inc_cset_max_finger = MAX2(_inc_cset_max_finger, hr_end); 1740 1741 assert(!hr->in_collection_set(), "invariant"); 1742 hr->set_in_collection_set(true); 1743 assert( hr->next_in_collection_set() == NULL, "invariant"); 1744 1745 _g1->register_region_with_in_cset_fast_test(hr); 1746 } 1747 1748 // Add the region at the RHS of the incremental cset 1749 void G1CollectorPolicy::add_region_to_incremental_cset_rhs(HeapRegion* hr) { 1750 // We should only ever be appending survivors at the end of a pause 1751 assert( hr->is_survivor(), "Logic"); 1752 1753 // Do the 'common' stuff 1754 add_region_to_incremental_cset_common(hr); 1755 1756 // Now add the region at the right hand side 1757 if (_inc_cset_tail == NULL) { 1758 assert(_inc_cset_head == NULL, "invariant"); 1759 _inc_cset_head = hr; 1760 } else { 1761 _inc_cset_tail->set_next_in_collection_set(hr); 1762 } 1763 _inc_cset_tail = hr; 1764 } 1765 1766 // Add the region to the LHS of the incremental cset 1767 void G1CollectorPolicy::add_region_to_incremental_cset_lhs(HeapRegion* hr) { 1768 // Survivors should be added to the RHS at the end of a pause 1769 assert(!hr->is_survivor(), "Logic"); 1770 1771 // Do the 'common' stuff 1772 add_region_to_incremental_cset_common(hr); 1773 1774 // Add the region at the left hand side 1775 hr->set_next_in_collection_set(_inc_cset_head); 1776 if (_inc_cset_head == NULL) { 1777 assert(_inc_cset_tail == NULL, "Invariant"); 1778 _inc_cset_tail = hr; 1779 } 1780 _inc_cset_head = hr; 1781 } 1782 1783 #ifndef PRODUCT 1784 void G1CollectorPolicy::print_collection_set(HeapRegion* list_head, outputStream* st) { 1785 assert(list_head == inc_cset_head() || list_head == collection_set(), "must be"); 1786 1787 st->print_cr("\nCollection_set:"); 1788 HeapRegion* csr = list_head; 1789 while (csr != NULL) { 1790 HeapRegion* next = csr->next_in_collection_set(); 1791 assert(csr->in_collection_set(), "bad CS"); 1792 st->print_cr(" "HR_FORMAT", P: "PTR_FORMAT "N: "PTR_FORMAT", age: %4d", 1793 HR_FORMAT_PARAMS(csr), 1794 csr->prev_top_at_mark_start(), csr->next_top_at_mark_start(), 1795 csr->age_in_surv_rate_group_cond()); 1796 csr = next; 1797 } 1798 } 1799 #endif // !PRODUCT 1800 1801 bool G1CollectorPolicy::next_gc_should_be_mixed(const char* true_action_str, 1802 const char* false_action_str) { 1803 CollectionSetChooser* cset_chooser = _collectionSetChooser; 1804 if (cset_chooser->is_empty()) { 1805 ergo_verbose0(ErgoMixedGCs, 1806 false_action_str, 1807 ergo_format_reason("candidate old regions not available")); 1808 return false; 1809 } 1810 size_t reclaimable_bytes = cset_chooser->remaining_reclaimable_bytes(); 1811 size_t capacity_bytes = _g1->capacity(); 1812 double perc = (double) reclaimable_bytes * 100.0 / (double) capacity_bytes; 1813 double threshold = (double) G1HeapWastePercent; 1814 if (perc < threshold) { 1815 ergo_verbose4(ErgoMixedGCs, 1816 false_action_str, 1817 ergo_format_reason("reclaimable percentage lower than threshold") 1818 ergo_format_region("candidate old regions") 1819 ergo_format_byte_perc("reclaimable") 1820 ergo_format_perc("threshold"), 1821 cset_chooser->remaining_regions(), 1822 reclaimable_bytes, perc, threshold); 1823 return false; 1824 } 1825 1826 ergo_verbose4(ErgoMixedGCs, 1827 true_action_str, 1828 ergo_format_reason("candidate old regions available") 1829 ergo_format_region("candidate old regions") 1830 ergo_format_byte_perc("reclaimable") 1831 ergo_format_perc("threshold"), 1832 cset_chooser->remaining_regions(), 1833 reclaimable_bytes, perc, threshold); 1834 return true; 1835 } 1836 1837 void G1CollectorPolicy::finalize_cset(double target_pause_time_ms) { 1838 double young_start_time_sec = os::elapsedTime(); 1839 1840 YoungList* young_list = _g1->young_list(); 1841 finalize_incremental_cset_building(); 1842 1843 guarantee(target_pause_time_ms > 0.0, 1844 err_msg("target_pause_time_ms = %1.6lf should be positive", 1845 target_pause_time_ms)); 1846 guarantee(_collection_set == NULL, "Precondition"); 1847 1848 double base_time_ms = predict_base_elapsed_time_ms(_pending_cards); 1849 double predicted_pause_time_ms = base_time_ms; 1850 double time_remaining_ms = target_pause_time_ms - base_time_ms; 1851 1852 ergo_verbose4(ErgoCSetConstruction | ErgoHigh, 1853 "start choosing CSet", 1854 ergo_format_size("_pending_cards") 1855 ergo_format_ms("predicted base time") 1856 ergo_format_ms("remaining time") 1857 ergo_format_ms("target pause time"), 1858 _pending_cards, base_time_ms, time_remaining_ms, target_pause_time_ms); 1859 1860 _last_gc_was_young = gcs_are_young() ? true : false; 1861 1862 if (_last_gc_was_young) { 1863 _trace_gen0_time_data.increment_young_collection_count(); 1864 } else { 1865 _trace_gen0_time_data.increment_mixed_collection_count(); 1866 } 1867 1868 // The young list is laid with the survivor regions from the previous 1869 // pause are appended to the RHS of the young list, i.e. 1870 // [Newly Young Regions ++ Survivors from last pause]. 1871 1872 uint survivor_region_length = young_list->survivor_length(); 1873 uint eden_region_length = young_list->length() - survivor_region_length; 1874 init_cset_region_lengths(eden_region_length, survivor_region_length); 1875 1876 HeapRegion* hr = young_list->first_survivor_region(); 1877 while (hr != NULL) { 1878 assert(hr->is_survivor(), "badly formed young list"); 1879 hr->set_young(); 1880 hr = hr->get_next_young_region(); 1881 } 1882 1883 // Clear the fields that point to the survivor list - they are all young now. 1884 young_list->clear_survivors(); 1885 1886 _collection_set = _inc_cset_head; 1887 _collection_set_bytes_used_before = _inc_cset_bytes_used_before; 1888 time_remaining_ms -= _inc_cset_predicted_elapsed_time_ms; 1889 predicted_pause_time_ms += _inc_cset_predicted_elapsed_time_ms; 1890 1891 ergo_verbose3(ErgoCSetConstruction | ErgoHigh, 1892 "add young regions to CSet", 1893 ergo_format_region("eden") 1894 ergo_format_region("survivors") 1895 ergo_format_ms("predicted young region time"), 1896 eden_region_length, survivor_region_length, 1897 _inc_cset_predicted_elapsed_time_ms); 1898 1899 // The number of recorded young regions is the incremental 1900 // collection set's current size 1901 set_recorded_rs_lengths(_inc_cset_recorded_rs_lengths); 1902 1903 double young_end_time_sec = os::elapsedTime(); 1904 phase_times()->record_young_cset_choice_time_ms((young_end_time_sec - young_start_time_sec) * 1000.0); 1905 1906 // Set the start of the non-young choice time. 1907 double non_young_start_time_sec = young_end_time_sec; 1908 1909 if (!gcs_are_young()) { 1910 CollectionSetChooser* cset_chooser = _collectionSetChooser; 1911 cset_chooser->verify(); 1912 const uint min_old_cset_length = cset_chooser->calc_min_old_cset_length(); 1913 const uint max_old_cset_length = cset_chooser->calc_max_old_cset_length(); 1914 1915 uint expensive_region_num = 0; 1916 bool check_time_remaining = adaptive_young_list_length(); 1917 1918 HeapRegion* hr = cset_chooser->peek(); 1919 while (hr != NULL) { 1920 if (old_cset_region_length() >= max_old_cset_length) { 1921 // Added maximum number of old regions to the CSet. 1922 ergo_verbose2(ErgoCSetConstruction, 1923 "finish adding old regions to CSet", 1924 ergo_format_reason("old CSet region num reached max") 1925 ergo_format_region("old") 1926 ergo_format_region("max"), 1927 old_cset_region_length(), max_old_cset_length); 1928 break; 1929 } 1930 1931 double predicted_time_ms = predict_region_elapsed_time_ms(hr, gcs_are_young()); 1932 if (check_time_remaining) { 1933 if (predicted_time_ms > time_remaining_ms) { 1934 // Too expensive for the current CSet. 1935 1936 if (old_cset_region_length() >= min_old_cset_length) { 1937 // We have added the minimum number of old regions to the CSet, 1938 // we are done with this CSet. 1939 ergo_verbose4(ErgoCSetConstruction, 1940 "finish adding old regions to CSet", 1941 ergo_format_reason("predicted time is too high") 1942 ergo_format_ms("predicted time") 1943 ergo_format_ms("remaining time") 1944 ergo_format_region("old") 1945 ergo_format_region("min"), 1946 predicted_time_ms, time_remaining_ms, 1947 old_cset_region_length(), min_old_cset_length); 1948 break; 1949 } 1950 1951 // We'll add it anyway given that we haven't reached the 1952 // minimum number of old regions. 1953 expensive_region_num += 1; 1954 } 1955 } else { 1956 if (old_cset_region_length() >= min_old_cset_length) { 1957 // In the non-auto-tuning case, we'll finish adding regions 1958 // to the CSet if we reach the minimum. 1959 ergo_verbose2(ErgoCSetConstruction, 1960 "finish adding old regions to CSet", 1961 ergo_format_reason("old CSet region num reached min") 1962 ergo_format_region("old") 1963 ergo_format_region("min"), 1964 old_cset_region_length(), min_old_cset_length); 1965 break; 1966 } 1967 } 1968 1969 // We will add this region to the CSet. 1970 time_remaining_ms -= predicted_time_ms; 1971 predicted_pause_time_ms += predicted_time_ms; 1972 cset_chooser->remove_and_move_to_next(hr); 1973 _g1->old_set_remove(hr); 1974 add_old_region_to_cset(hr); 1975 1976 hr = cset_chooser->peek(); 1977 } 1978 if (hr == NULL) { 1979 ergo_verbose0(ErgoCSetConstruction, 1980 "finish adding old regions to CSet", 1981 ergo_format_reason("candidate old regions not available")); 1982 } 1983 1984 if (expensive_region_num > 0) { 1985 // We print the information once here at the end, predicated on 1986 // whether we added any apparently expensive regions or not, to 1987 // avoid generating output per region. 1988 ergo_verbose4(ErgoCSetConstruction, 1989 "added expensive regions to CSet", 1990 ergo_format_reason("old CSet region num not reached min") 1991 ergo_format_region("old") 1992 ergo_format_region("expensive") 1993 ergo_format_region("min") 1994 ergo_format_ms("remaining time"), 1995 old_cset_region_length(), 1996 expensive_region_num, 1997 min_old_cset_length, 1998 time_remaining_ms); 1999 } 2000 2001 cset_chooser->verify(); 2002 } 2003 2004 stop_incremental_cset_building(); 2005 2006 ergo_verbose5(ErgoCSetConstruction, 2007 "finish choosing CSet", 2008 ergo_format_region("eden") 2009 ergo_format_region("survivors") 2010 ergo_format_region("old") 2011 ergo_format_ms("predicted pause time") 2012 ergo_format_ms("target pause time"), 2013 eden_region_length, survivor_region_length, 2014 old_cset_region_length(), 2015 predicted_pause_time_ms, target_pause_time_ms); 2016 2017 double non_young_end_time_sec = os::elapsedTime(); 2018 phase_times()->record_non_young_cset_choice_time_ms((non_young_end_time_sec - non_young_start_time_sec) * 1000.0); 2019 } 2020 2021 void TraceGen0TimeData::record_start_collection(double time_to_stop_the_world_ms) { 2022 if(TraceGen0Time) { 2023 _all_stop_world_times_ms.add(time_to_stop_the_world_ms); 2024 } 2025 } 2026 2027 void TraceGen0TimeData::record_yield_time(double yield_time_ms) { 2028 if(TraceGen0Time) { 2029 _all_yield_times_ms.add(yield_time_ms); 2030 } 2031 } 2032 2033 void TraceGen0TimeData::record_end_collection(double pause_time_ms, G1GCPhaseTimes* phase_times) { 2034 if(TraceGen0Time) { 2035 _total.add(pause_time_ms); 2036 _other.add(pause_time_ms - phase_times->accounted_time_ms()); 2037 _root_region_scan_wait.add(phase_times->root_region_scan_wait_time_ms()); 2038 _parallel.add(phase_times->cur_collection_par_time_ms()); 2039 _ext_root_scan.add(phase_times->average_last_ext_root_scan_time()); 2040 _satb_filtering.add(phase_times->average_last_satb_filtering_times_ms()); 2041 _update_rs.add(phase_times->average_last_update_rs_time()); 2042 _scan_rs.add(phase_times->average_last_scan_rs_time()); 2043 _obj_copy.add(phase_times->average_last_obj_copy_time()); 2044 _termination.add(phase_times->average_last_termination_time()); 2045 2046 double parallel_known_time = phase_times->average_last_ext_root_scan_time() + 2047 phase_times->average_last_satb_filtering_times_ms() + 2048 phase_times->average_last_update_rs_time() + 2049 phase_times->average_last_scan_rs_time() + 2050 phase_times->average_last_obj_copy_time() + 2051 + phase_times->average_last_termination_time(); 2052 2053 double parallel_other_time = phase_times->cur_collection_par_time_ms() - parallel_known_time; 2054 _parallel_other.add(parallel_other_time); 2055 _clear_ct.add(phase_times->cur_clear_ct_time_ms()); 2056 } 2057 } 2058 2059 void TraceGen0TimeData::increment_young_collection_count() { 2060 if(TraceGen0Time) { 2061 ++_young_pause_num; 2062 } 2063 } 2064 2065 void TraceGen0TimeData::increment_mixed_collection_count() { 2066 if(TraceGen0Time) { 2067 ++_mixed_pause_num; 2068 } 2069 } 2070 2071 void TraceGen0TimeData::print_summary(const char* str, 2072 const NumberSeq* seq) const { 2073 double sum = seq->sum(); 2074 gclog_or_tty->print_cr("%-27s = %8.2lf s (avg = %8.2lf ms)", 2075 str, sum / 1000.0, seq->avg()); 2076 } 2077 2078 void TraceGen0TimeData::print_summary_sd(const char* str, 2079 const NumberSeq* seq) const { 2080 print_summary(str, seq); 2081 gclog_or_tty->print_cr("%+45s = %5d, std dev = %8.2lf ms, max = %8.2lf ms)", 2082 "(num", seq->num(), seq->sd(), seq->maximum()); 2083 } 2084 2085 void TraceGen0TimeData::print() const { 2086 if (!TraceGen0Time) { 2087 return; 2088 } 2089 2090 gclog_or_tty->print_cr("ALL PAUSES"); 2091 print_summary_sd(" Total", &_total); 2092 gclog_or_tty->print_cr(""); 2093 gclog_or_tty->print_cr(""); 2094 gclog_or_tty->print_cr(" Young GC Pauses: %8d", _young_pause_num); 2095 gclog_or_tty->print_cr(" Mixed GC Pauses: %8d", _mixed_pause_num); 2096 gclog_or_tty->print_cr(""); 2097 2098 gclog_or_tty->print_cr("EVACUATION PAUSES"); 2099 2100 if (_young_pause_num == 0 && _mixed_pause_num == 0) { 2101 gclog_or_tty->print_cr("none"); 2102 } else { 2103 print_summary_sd(" Evacuation Pauses", &_total); 2104 print_summary(" Root Region Scan Wait", &_root_region_scan_wait); 2105 print_summary(" Parallel Time", &_parallel); 2106 print_summary(" Ext Root Scanning", &_ext_root_scan); 2107 print_summary(" SATB Filtering", &_satb_filtering); 2108 print_summary(" Update RS", &_update_rs); 2109 print_summary(" Scan RS", &_scan_rs); 2110 print_summary(" Object Copy", &_obj_copy); 2111 print_summary(" Termination", &_termination); 2112 print_summary(" Parallel Other", &_parallel_other); 2113 print_summary(" Clear CT", &_clear_ct); 2114 print_summary(" Other", &_other); 2115 } 2116 gclog_or_tty->print_cr(""); 2117 2118 gclog_or_tty->print_cr("MISC"); 2119 print_summary_sd(" Stop World", &_all_stop_world_times_ms); 2120 print_summary_sd(" Yields", &_all_yield_times_ms); 2121 } 2122 2123 void TraceGen1TimeData::record_full_collection(double full_gc_time_ms) { 2124 if (TraceGen1Time) { 2125 _all_full_gc_times.add(full_gc_time_ms); 2126 } 2127 } 2128 2129 void TraceGen1TimeData::print() const { 2130 if (!TraceGen1Time) { 2131 return; 2132 } 2133 2134 if (_all_full_gc_times.num() > 0) { 2135 gclog_or_tty->print("\n%4d full_gcs: total time = %8.2f s", 2136 _all_full_gc_times.num(), 2137 _all_full_gc_times.sum() / 1000.0); 2138 gclog_or_tty->print_cr(" (avg = %8.2fms).", _all_full_gc_times.avg()); 2139 gclog_or_tty->print_cr(" [std. dev = %8.2f ms, max = %8.2f ms]", 2140 _all_full_gc_times.sd(), 2141 _all_full_gc_times.maximum()); 2142 } 2143 }