comparison src/share/vm/gc_implementation/g1/g1RemSetSummary.cpp @ 10372:e72f7eecc96d

8013895: G1: G1SummarizeRSetStats output on Linux needs improvemen Summary: Fixed the output of G1SummarizeRSetStats: too small datatype for the number of concurrently processed cards, added concurrent remembered set thread time retrieval for Linux and Windows (BSD uses os::elapsedTime() now), and other cleanup. The information presented during VM operation is now relative to the previous output, not always cumulative if G1SummarizeRSetStatsPeriod > 0. At VM exit, the code prints a cumulative summary. Reviewed-by: johnc, jwilhelm
author tschatzl
date Tue, 28 May 2013 09:32:06 +0200
parents
children 5888334c9c24
comparison
equal deleted inserted replaced
10371:c20186fa611b 10372:e72f7eecc96d
1 /*
2 * Copyright (c) 2013, 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/concurrentG1RefineThread.hpp"
28 #include "gc_implementation/g1/heapRegion.hpp"
29 #include "gc_implementation/g1/g1CollectedHeap.inline.hpp"
30 #include "gc_implementation/g1/g1RemSet.inline.hpp"
31 #include "gc_implementation/g1/g1RemSetSummary.hpp"
32 #include "gc_implementation/g1/heapRegionRemSet.hpp"
33 #include "runtime/thread.inline.hpp"
34
35 class GetRSThreadVTimeClosure : public ThreadClosure {
36 private:
37 G1RemSetSummary* _summary;
38 uint _counter;
39
40 public:
41 GetRSThreadVTimeClosure(G1RemSetSummary * summary) : ThreadClosure(), _summary(summary), _counter(0) {
42 assert(_summary != NULL, "just checking");
43 }
44
45 virtual void do_thread(Thread* t) {
46 ConcurrentG1RefineThread* crt = (ConcurrentG1RefineThread*) t;
47 _summary->set_rs_thread_vtime(_counter, crt->vtime_accum());
48 _counter++;
49 }
50 };
51
52 void G1RemSetSummary::update() {
53 _num_refined_cards = remset()->conc_refine_cards();
54 DirtyCardQueueSet& dcqs = JavaThread::dirty_card_queue_set();
55 _num_processed_buf_mutator = dcqs.processed_buffers_mut();
56 _num_processed_buf_rs_threads = dcqs.processed_buffers_rs_thread();
57
58 _num_coarsenings = HeapRegionRemSet::n_coarsenings();
59
60 ConcurrentG1Refine * cg1r = G1CollectedHeap::heap()->concurrent_g1_refine();
61 if (_rs_threads_vtimes != NULL) {
62 GetRSThreadVTimeClosure p(this);
63 cg1r->worker_threads_do(&p);
64 }
65 set_sampling_thread_vtime(cg1r->sampling_thread()->vtime_accum());
66 }
67
68 void G1RemSetSummary::set_rs_thread_vtime(uint thread, double value) {
69 assert(_rs_threads_vtimes != NULL, "just checking");
70 assert(thread < _num_vtimes, "just checking");
71 _rs_threads_vtimes[thread] = value;
72 }
73
74 double G1RemSetSummary::rs_thread_vtime(uint thread) const {
75 assert(_rs_threads_vtimes != NULL, "just checking");
76 assert(thread < _num_vtimes, "just checking");
77 return _rs_threads_vtimes[thread];
78 }
79
80 void G1RemSetSummary::initialize(G1RemSet* remset, uint num_workers) {
81 assert(_rs_threads_vtimes == NULL, "just checking");
82 assert(remset != NULL, "just checking");
83
84 _remset = remset;
85 _num_vtimes = num_workers;
86 _rs_threads_vtimes = NEW_C_HEAP_ARRAY(double, _num_vtimes, mtGC);
87 memset(_rs_threads_vtimes, 0, sizeof(double) * _num_vtimes);
88
89 update();
90 }
91
92 void G1RemSetSummary::set(G1RemSetSummary* other) {
93 assert(other != NULL, "just checking");
94 assert(remset() == other->remset(), "just checking");
95 assert(_num_vtimes == other->_num_vtimes, "just checking");
96
97 _num_refined_cards = other->num_concurrent_refined_cards();
98
99 _num_processed_buf_mutator = other->num_processed_buf_mutator();
100 _num_processed_buf_rs_threads = other->num_processed_buf_rs_threads();
101
102 _num_coarsenings = other->_num_coarsenings;
103
104 memcpy(_rs_threads_vtimes, other->_rs_threads_vtimes, sizeof(double) * _num_vtimes);
105
106 set_sampling_thread_vtime(other->sampling_thread_vtime());
107 }
108
109 void G1RemSetSummary::subtract_from(G1RemSetSummary* other) {
110 assert(other != NULL, "just checking");
111 assert(remset() == other->remset(), "just checking");
112 assert(_num_vtimes == other->_num_vtimes, "just checking");
113
114 _num_refined_cards = other->num_concurrent_refined_cards() - _num_refined_cards;
115
116 _num_processed_buf_mutator = other->num_processed_buf_mutator() - _num_processed_buf_mutator;
117 _num_processed_buf_rs_threads = other->num_processed_buf_rs_threads() - _num_processed_buf_rs_threads;
118
119 _num_coarsenings = other->num_coarsenings() - _num_coarsenings;
120
121 for (uint i = 0; i < _num_vtimes; i++) {
122 set_rs_thread_vtime(i, other->rs_thread_vtime(i) - rs_thread_vtime(i));
123 }
124
125 _sampling_thread_vtime = other->sampling_thread_vtime() - _sampling_thread_vtime;
126 }
127
128 class HRRSStatsIter: public HeapRegionClosure {
129 size_t _occupied;
130 size_t _total_mem_sz;
131 size_t _max_mem_sz;
132 HeapRegion* _max_mem_sz_region;
133 public:
134 HRRSStatsIter() :
135 _occupied(0),
136 _total_mem_sz(0),
137 _max_mem_sz(0),
138 _max_mem_sz_region(NULL)
139 {}
140
141 bool doHeapRegion(HeapRegion* r) {
142 size_t mem_sz = r->rem_set()->mem_size();
143 if (mem_sz > _max_mem_sz) {
144 _max_mem_sz = mem_sz;
145 _max_mem_sz_region = r;
146 }
147 _total_mem_sz += mem_sz;
148 size_t occ = r->rem_set()->occupied();
149 _occupied += occ;
150 return false;
151 }
152 size_t total_mem_sz() { return _total_mem_sz; }
153 size_t max_mem_sz() { return _max_mem_sz; }
154 size_t occupied() { return _occupied; }
155 HeapRegion* max_mem_sz_region() { return _max_mem_sz_region; }
156 };
157
158 double calc_percentage(size_t numerator, size_t denominator) {
159 if (denominator != 0) {
160 return (double)numerator / denominator * 100.0;
161 } else {
162 return 0.0f;
163 }
164 }
165
166 void G1RemSetSummary::print_on(outputStream* out) {
167 out->print_cr("\n Concurrent RS processed "SIZE_FORMAT" cards",
168 num_concurrent_refined_cards());
169 out->print_cr(" Of %d completed buffers:", num_processed_buf_total());
170 out->print_cr(" %8d (%5.1f%%) by concurrent RS threads.",
171 num_processed_buf_total(),
172 calc_percentage(num_processed_buf_rs_threads(), num_processed_buf_total()));
173 out->print_cr(" %8d (%5.1f%%) by mutator threads.",
174 num_processed_buf_mutator(),
175 calc_percentage(num_processed_buf_mutator(), num_processed_buf_total()));
176 out->print_cr(" Concurrent RS threads times (s)");
177 out->print(" ");
178 for (uint i = 0; i < _num_vtimes; i++) {
179 out->print(" %5.2f", rs_thread_vtime(i));
180 }
181 out->cr();
182 out->print_cr(" Concurrent sampling threads times (s)");
183 out->print_cr(" %5.2f", sampling_thread_vtime());
184
185 HRRSStatsIter blk;
186 G1CollectedHeap::heap()->heap_region_iterate(&blk);
187 out->print_cr(" Total heap region rem set sizes = "SIZE_FORMAT"K."
188 " Max = "SIZE_FORMAT"K.",
189 blk.total_mem_sz()/K, blk.max_mem_sz()/K);
190 out->print_cr(" Static structures = "SIZE_FORMAT"K,"
191 " free_lists = "SIZE_FORMAT"K.",
192 HeapRegionRemSet::static_mem_size() / K,
193 HeapRegionRemSet::fl_mem_size() / K);
194 out->print_cr(" "SIZE_FORMAT" occupied cards represented.",
195 blk.occupied());
196 HeapRegion* max_mem_sz_region = blk.max_mem_sz_region();
197 HeapRegionRemSet* rem_set = max_mem_sz_region->rem_set();
198 out->print_cr(" Max size region = "HR_FORMAT", "
199 "size = "SIZE_FORMAT "K, occupied = "SIZE_FORMAT"K.",
200 HR_FORMAT_PARAMS(max_mem_sz_region),
201 (rem_set->mem_size() + K - 1)/K,
202 (rem_set->occupied() + K - 1)/K);
203
204 out->print_cr(" Did %d coarsenings.", num_coarsenings());
205 }