1 /* 2 * Copyright (c) 2018, Red Hat, Inc. All rights reserved. 3 * 4 * This code is free software; you can redistribute it and/or modify it 5 * under the terms of the GNU General Public License version 2 only, as 6 * published by the Free Software Foundation. 7 * 8 * This code is distributed in the hope that it will be useful, but WITHOUT 9 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or 10 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License 11 * version 2 for more details (a copy is included in the LICENSE file that 12 * accompanied this code). 13 * 14 * You should have received a copy of the GNU General Public License version 15 * 2 along with this work; if not, write to the Free Software Foundation, 16 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. 17 * 18 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA 19 * or visit www.oracle.com if you need additional information or have any 20 * questions. 21 * 22 */ 23 24 #include "precompiled.hpp" 25 26 #include "gc/shenandoah/shenandoahFreeSet.hpp" 27 #include "gc/shenandoah/shenandoahHeap.hpp" 28 #include "gc/shenandoah/shenandoahHeap.inline.hpp" 29 #include "gc/shenandoah/shenandoahPacer.hpp" 30 31 /* 32 * In normal concurrent cycle, we have to pace the application to let GC finish. 33 * 34 * Here, we do not know how large would be the collection set, and what are the 35 * relative performances of the each stage in the concurrent cycle, and so we have to 36 * make some assumptions. 37 * 38 * For concurrent mark, there is no clear notion of progress. The moderately accurate 39 * and easy to get metric is the amount of live objects the mark had encountered. But, 40 * that does directly correlate with the used heap, because the heap might be fully 41 * dead or fully alive. We cannot assume either of the extremes: we would either allow 42 * application to run out of memory if we assume heap is fully dead but it is not, and, 43 * conversely, we would pacify application excessively if we assume heap is fully alive 44 * but it is not. So we need to guesstimate the particular expected value for heap liveness. 45 * The best way to do this is apparently recording the past history. 46 * 47 * For concurrent evac and update-refs, we are walking the heap per-region, and so the 48 * notion of progress is clear: we get reported the "used" size from the processed regions 49 * and use the global heap-used as the baseline. 50 * 51 * The allocatable space when GC is running is "free" at the start of cycle, but the 52 * accounted budget is based on "used". So, we need to adjust the tax knowing that. 53 * Also, since we effectively count the used space three times (mark, evac, update-refs), 54 * we need to multiply the tax by 3. Example: for 10 MB free and 90 MB used, GC would 55 * come back with 3*90 MB budget, and thus for each 1 MB of allocation, we have to pay 56 * 3*90 / 10 MBs. In the end, we would pay back the entire budget. 57 */ 58 59 void ShenandoahPacer::setup_for_mark() { 60 assert(ShenandoahPacing, "Only be here when pacing is enabled"); 61 62 size_t live = update_and_get_progress_history(); 63 size_t free = _heap->free_set()->available(); 64 65 size_t non_taxable = free * ShenandoahPacingCycleSlack / 100; 66 size_t taxable = free - non_taxable; 67 68 double tax = 1.0 * live / taxable; // base tax for available free space 69 tax *= 3; // mark is phase 1 of 3, claim 1/3 of free for it 70 tax *= ShenandoahPacingSurcharge; // additional surcharge to help unclutter heap 71 72 restart_with(non_taxable, tax); 73 74 log_info(gc, ergo)("Pacer for Mark. Expected Live: " SIZE_FORMAT "M, Free: " SIZE_FORMAT 75 "M, Non-Taxable: " SIZE_FORMAT "M, Alloc Tax Rate: %.1fx", 76 live / M, free / M, non_taxable / M, tax); 77 } 78 79 void ShenandoahPacer::setup_for_evac() { 80 assert(ShenandoahPacing, "Only be here when pacing is enabled"); 81 82 size_t used = _heap->collection_set()->used(); 83 size_t free = _heap->free_set()->available(); 84 85 size_t non_taxable = free * ShenandoahPacingCycleSlack / 100; 86 size_t taxable = free - non_taxable; 87 88 double tax = 1.0 * used / taxable; // base tax for available free space 89 tax *= 2; // evac is phase 2 of 3, claim 1/2 of remaining free 90 tax = MAX2<double>(1, tax); // never allocate more than GC processes during the phase 91 tax *= ShenandoahPacingSurcharge; // additional surcharge to help unclutter heap 92 93 restart_with(non_taxable, tax); 94 95 log_info(gc, ergo)("Pacer for Evacuation. Used CSet: " SIZE_FORMAT "M, Free: " SIZE_FORMAT 96 "M, Non-Taxable: " SIZE_FORMAT "M, Alloc Tax Rate: %.1fx", 97 used / M, free / M, non_taxable / M, tax); 98 } 99 100 void ShenandoahPacer::setup_for_updaterefs() { 101 assert(ShenandoahPacing, "Only be here when pacing is enabled"); 102 103 size_t used = _heap->used(); 104 size_t free = _heap->free_set()->available(); 105 106 size_t non_taxable = free * ShenandoahPacingCycleSlack / 100; 107 size_t taxable = free - non_taxable; 108 109 double tax = 1.0 * used / taxable; // base tax for available free space 110 tax *= 1; // update-refs is phase 3 of 3, claim the remaining free 111 tax = MAX2<double>(1, tax); // never allocate more than GC processes during the phase 112 tax *= ShenandoahPacingSurcharge; // additional surcharge to help unclutter heap 113 114 restart_with(non_taxable, tax); 115 116 log_info(gc, ergo)("Pacer for Update Refs. Used: " SIZE_FORMAT "M, Free: " SIZE_FORMAT 117 "M, Non-Taxable: " SIZE_FORMAT "M, Alloc Tax Rate: %.1fx", 118 used / M, free / M, non_taxable / M, tax); 119 } 120 121 /* 122 * Traversal walks the entire heap once, and therefore we have to make assumptions about its 123 * liveness, like concurrent mark does. 124 */ 125 126 void ShenandoahPacer::setup_for_traversal() { 127 assert(ShenandoahPacing, "Only be here when pacing is enabled"); 128 129 size_t live = update_and_get_progress_history(); 130 size_t free = _heap->free_set()->available(); 131 132 size_t non_taxable = free * ShenandoahPacingCycleSlack / 100; 133 size_t taxable = free - non_taxable; 134 135 double tax = 1.0 * live / taxable; // base tax for available free space 136 tax *= ShenandoahPacingSurcharge; // additional surcharge to help unclutter heap 137 138 restart_with(non_taxable, tax); 139 140 log_info(gc, ergo)("Pacer for Traversal. Expected Live: " SIZE_FORMAT "M, Free: " SIZE_FORMAT 141 "M, Non-Taxable: " SIZE_FORMAT "M, Alloc Tax Rate: %.1fx", 142 live / M, free / M, non_taxable / M, tax); 143 } 144 145 /* 146 * In idle phase, we have to pace the application to let control thread react with GC start. 147 * 148 * Here, we have rendezvous with concurrent thread that adds up the budget as it acknowledges 149 * it had seen recent allocations. It will naturally pace the allocations if control thread is 150 * not catching up. To bootstrap this feedback cycle, we need to start with some initial budget 151 * for applications to allocate at. 152 */ 153 154 void ShenandoahPacer::setup_for_idle() { 155 assert(ShenandoahPacing, "Only be here when pacing is enabled"); 156 157 size_t initial = _heap->capacity() * ShenandoahPacingIdleSlack / 100; 158 double tax = 1; 159 160 restart_with(initial, tax); 161 162 log_info(gc, ergo)("Pacer for Idle. Initial: " SIZE_FORMAT "M, Alloc Tax Rate: %.1fx", 163 initial / M, tax); 164 } 165 166 size_t ShenandoahPacer::update_and_get_progress_history() { 167 if (_progress == -1) { 168 // First initialization, report some prior 169 Atomic::store((intptr_t)PACING_PROGRESS_ZERO, &_progress); 170 return (size_t) (_heap->capacity() * 0.1); 171 } else { 172 // Record history, and reply historical data 173 _progress_history->add(_progress); 174 Atomic::store((intptr_t)PACING_PROGRESS_ZERO, &_progress); 175 return (size_t) (_progress_history->avg() * HeapWordSize); 176 } 177 } 178 179 void ShenandoahPacer::restart_with(size_t non_taxable_bytes, double tax_rate) { 180 size_t initial = (size_t)(non_taxable_bytes * tax_rate) >> LogHeapWordSize; 181 STATIC_ASSERT(sizeof(size_t) <= sizeof(intptr_t)); 182 Atomic::xchg((intptr_t)initial, &_budget); 183 Atomic::store(tax_rate, &_tax_rate); 184 Atomic::inc(&_epoch); 185 } 186 187 bool ShenandoahPacer::claim_for_alloc(size_t words, bool force) { 188 assert(ShenandoahPacing, "Only be here when pacing is enabled"); 189 190 intptr_t tax = MAX2<intptr_t>(1, words * Atomic::load(&_tax_rate)); 191 192 intptr_t cur = 0; 193 intptr_t new_val = 0; 194 do { 195 cur = Atomic::load(&_budget); 196 if (cur < tax && !force) { 197 // Progress depleted, alas. 198 return false; 199 } 200 new_val = cur - tax; 201 } while (Atomic::cmpxchg(new_val, &_budget, cur) != cur); 202 return true; 203 } 204 205 void ShenandoahPacer::unpace_for_alloc(intptr_t epoch, size_t words) { 206 assert(ShenandoahPacing, "Only be here when pacing is enabled"); 207 208 if (_epoch != epoch) { 209 // Stale ticket, no need to unpace. 210 return; 211 } 212 213 intptr_t tax = MAX2<intptr_t>(1, words * Atomic::load(&_tax_rate)); 214 Atomic::add(tax, &_budget); 215 } 216 217 intptr_t ShenandoahPacer::epoch() { 218 return Atomic::load(&_epoch); 219 } 220 221 void ShenandoahPacer::pace_for_alloc(size_t words) { 222 assert(ShenandoahPacing, "Only be here when pacing is enabled"); 223 224 // Fast path: try to allocate right away 225 if (claim_for_alloc(words, false)) { 226 return; 227 } 228 229 size_t max = ShenandoahPacingMaxDelay; 230 double start = os::elapsedTime(); 231 232 size_t total = 0; 233 size_t cur = 0; 234 235 while (true) { 236 // We could instead assist GC, but this would suffice for now. 237 // This code should also participate in safepointing. 238 // Perform the exponential backoff, limited by max. 239 240 cur = cur * 2; 241 if (total + cur > max) { 242 cur = (max > total) ? (max - total) : 0; 243 } 244 cur = MAX2<size_t>(1, cur); 245 246 os::sleep(Thread::current(), cur, true); 247 248 double end = os::elapsedTime(); 249 total = (size_t)((end - start) * 1000); 250 251 if (total > max) { 252 // Spent local time budget to wait for enough GC progress. 253 // Breaking out and allocating anyway, which may mean we outpace GC, 254 // and start Degenerated GC cycle. 255 _delays.add(total); 256 257 // Forcefully claim the budget: it may go negative at this point, and 258 // GC should replenish for this and subsequent allocations 259 claim_for_alloc(words, true); 260 break; 261 } 262 263 if (claim_for_alloc(words, false)) { 264 // Acquired enough permit, nice. Can allocate now. 265 _delays.add(total); 266 break; 267 } 268 } 269 } 270 271 void ShenandoahPacer::print_on(outputStream* out) const { 272 out->print_cr("ALLOCATION PACING:"); 273 out->cr(); 274 275 out->print_cr("Max pacing delay is set for " UINTX_FORMAT " ms.", ShenandoahPacingMaxDelay); 276 out->cr(); 277 278 out->print_cr("Higher delay would prevent application outpacing the GC, but it will hide the GC latencies"); 279 out->print_cr("from the STW pause times. Pacing affects the individual threads, and so it would also be"); 280 out->print_cr("invisible to the usual profiling tools, but would add up to end-to-end application latency."); 281 out->print_cr("Raise max pacing delay with care."); 282 out->cr(); 283 284 out->print_cr("Actual pacing delays histogram:"); 285 out->cr(); 286 287 out->print_cr("%10s - %10s %12s%12s", "From", "To", "Count", "Sum"); 288 289 size_t total_count = 0; 290 size_t total_sum = 0; 291 for (int c = _delays.min_level(); c <= _delays.max_level(); c++) { 292 int l = (c == 0) ? 0 : 1 << (c - 1); 293 int r = 1 << c; 294 size_t count = _delays.level(c); 295 size_t sum = count * (r - l) / 2; 296 total_count += count; 297 total_sum += sum; 298 299 out->print_cr("%7d ms - %7d ms: " SIZE_FORMAT_W(12) SIZE_FORMAT_W(12) " ms", l, r, count, sum); 300 } 301 out->print_cr("%23s: " SIZE_FORMAT_W(12) SIZE_FORMAT_W(12) " ms", "Total", total_count, total_sum); 302 out->cr(); 303 }