요약·해설과 원문, 전문 번역을 서로 분리했습니다. API 이름, symbol, source path는 원문 표기를 사용합니다.
1. 요약·해설
원문의 핵심 논리와 kernel programming 관점의 보충 설명입니다. 아래의 전문 번역과는 별도로 작성했습니다.
2. 영어 원문 전체
번역 기준이 된 Linux v6.18.37 원문입니다. 줄 번호는 이 버전의 파일 좌표입니다.
원문 전체 펼치기
1
.. SPDX-License-Identifier: GPL-2.0
3
=================================================
4
DAMON Moniting Interval Parameters Tuning Example
5
=================================================
7
DAMON's monitoring parameters need tuning based on given workload and the
8
monitoring purpose. There is a :ref:`tuning guide
9
<damon_design_monitoring_params_tuning_guide>` for that. This document
10
provides an example tuning based on the guide.
12
Setup
13
=====
15
For below example, DAMON of Linux kernel v6.11 and `damo
16
<https://github.com/damonitor/damo>`_ (DAMON user-space tool) v2.5.9 was used to
17
monitor and visualize access patterns on the physical address space of a system
18
running a real-world server workload.
20
5ms/100ms intervals: Too Short Interval
21
=======================================
23
Let's start by capturing the access pattern snapshot on the physical address
24
space of the system using DAMON, with the default interval parameters (5
25
milliseconds and 100 milliseconds for the sampling and the aggregation
26
intervals, respectively). Wait ten minutes between the start of DAMON and
27
the capturing of the snapshot, to show a meaningful time-wise access patterns.
28
::
30
# damo start
31
# sleep 600
32
# damo record --snapshot 0 1
33
# damo stop
35
Then, list the DAMON-found regions of different access patterns, sorted by the
36
"access temperature". "Access temperature" is a metric representing the
37
access-hotness of a region. It is calculated as a weighted sum of the access
38
frequency and the age of the region. If the access frequency is 0 %, the
39
temperature is multiplied by minus one. That is, if a region is not accessed,
40
it gets minus temperature and it gets lower as not accessed for longer time.
41
The sorting is in temperature-ascendint order, so the region at the top of the
42
list is the coldest, and the one at the bottom is the hottest one. ::
44
# damo report access --sort_regions_by temperature
45
0 addr 16.052 GiB size 5.985 GiB access 0 % age 5.900 s # coldest
46
1 addr 22.037 GiB size 6.029 GiB access 0 % age 5.300 s
47
2 addr 28.065 GiB size 6.045 GiB access 0 % age 5.200 s
48
3 addr 10.069 GiB size 5.983 GiB access 0 % age 4.500 s
49
4 addr 4.000 GiB size 6.069 GiB access 0 % age 4.400 s
50
5 addr 62.008 GiB size 3.992 GiB access 0 % age 3.700 s
51
6 addr 56.795 GiB size 5.213 GiB access 0 % age 3.300 s
52
7 addr 39.393 GiB size 6.096 GiB access 0 % age 2.800 s
53
8 addr 50.782 GiB size 6.012 GiB access 0 % age 2.800 s
54
9 addr 34.111 GiB size 5.282 GiB access 0 % age 2.300 s
55
10 addr 45.489 GiB size 5.293 GiB access 0 % age 1.800 s # hottest
56
total size: 62.000 GiB
58
The list shows not seemingly hot regions, and only minimum access pattern
59
diversity. Every region has zero access frequency. The number of region is
60
10, which is the default ``min_nr_regions value``. Size of each region is also
61
nearly identical. We can suspect this is because “adaptive regions adjustment”
62
mechanism was not well working. As the guide suggested, we can get relative
63
hotness of regions using ``age`` as the recency information. That would be
64
better than nothing, but given the fact that the longest age is only about 6
65
seconds while we waited about ten minutes, it is unclear how useful this will
66
be.
68
The temperature ranges to total size of regions of each range histogram
69
visualization of the results also shows no interesting distribution pattern. ::
71
# damo report access --style temperature-sz-hist
72
<temperature> <total size>
73
[-,590,000,000, -,549,000,000) 5.985 GiB |********** |
74
[-,549,000,000, -,508,000,000) 12.074 GiB |********************|
75
[-,508,000,000, -,467,000,000) 0 B | |
76
[-,467,000,000, -,426,000,000) 12.052 GiB |********************|
77
[-,426,000,000, -,385,000,000) 0 B | |
78
[-,385,000,000, -,344,000,000) 3.992 GiB |******* |
79
[-,344,000,000, -,303,000,000) 5.213 GiB |********* |
80
[-,303,000,000, -,262,000,000) 12.109 GiB |********************|
81
[-,262,000,000, -,221,000,000) 5.282 GiB |********* |
82
[-,221,000,000, -,180,000,000) 0 B | |
83
[-,180,000,000, -,139,000,000) 5.293 GiB |********* |
84
total size: 62.000 GiB
86
In short, the parameters provide poor quality monitoring results for hot
87
regions detection. According to the :ref:`guide
88
<damon_design_monitoring_params_tuning_guide>`, this is due to the too short
89
aggregation interval.
91
100ms/2s intervals: Starts Showing Small Hot Regions
92
====================================================
94
Following the guide, increase the interval 20 times (100 milliseocnds and 2
95
seconds for sampling and aggregation intervals, respectively). ::
97
# damo start -s 100ms -a 2s
98
# sleep 600
99
# damo record --snapshot 0 1
100
# damo stop
101
# damo report access --sort_regions_by temperature
102
0 addr 10.180 GiB size 6.117 GiB access 0 % age 7 m 8 s # coldest
103
1 addr 49.275 GiB size 6.195 GiB access 0 % age 6 m 14 s
104
2 addr 62.421 GiB size 3.579 GiB access 0 % age 6 m 4 s
105
3 addr 40.154 GiB size 6.127 GiB access 0 % age 5 m 40 s
106
4 addr 16.296 GiB size 6.182 GiB access 0 % age 5 m 32 s
107
5 addr 34.254 GiB size 5.899 GiB access 0 % age 5 m 24 s
108
6 addr 46.281 GiB size 2.995 GiB access 0 % age 5 m 20 s
109
7 addr 28.420 GiB size 5.835 GiB access 0 % age 5 m 6 s
110
8 addr 4.000 GiB size 6.180 GiB access 0 % age 4 m 16 s
111
9 addr 22.478 GiB size 5.942 GiB access 0 % age 3 m 58 s
112
10 addr 55.470 GiB size 915.645 MiB access 0 % age 3 m 6 s
113
11 addr 56.364 GiB size 6.056 GiB access 0 % age 2 m 8 s
114
12 addr 56.364 GiB size 4.000 KiB access 95 % age 16 s
115
13 addr 49.275 GiB size 4.000 KiB access 100 % age 8 m 24 s # hottest
116
total size: 62.000 GiB
117
# damo report access --style temperature-sz-hist
118
<temperature> <total size>
119
[-42,800,000,000, -33,479,999,000) 22.018 GiB |***************** |
120
[-33,479,999,000, -24,159,998,000) 27.090 GiB |********************|
121
[-24,159,998,000, -14,839,997,000) 6.836 GiB |****** |
122
[-14,839,997,000, -5,519,996,000) 6.056 GiB |***** |
123
[-5,519,996,000, 3,800,005,000) 4.000 KiB |* |
124
[3,800,005,000, 13,120,006,000) 0 B | |
125
[13,120,006,000, 22,440,007,000) 0 B | |
126
[22,440,007,000, 31,760,008,000) 0 B | |
127
[31,760,008,000, 41,080,009,000) 0 B | |
128
[41,080,009,000, 50,400,010,000) 0 B | |
129
[50,400,010,000, 59,720,011,000) 4.000 KiB |* |
130
total size: 62.000 GiB
132
DAMON found two distinct 4 KiB regions that pretty hot. The regions are also
133
well aged. The hottest 4 KiB region was keeping the access frequency for about
134
8 minutes, and the coldest region was keeping no access for about 7 minutes.
135
The distribution on the histogram also looks like having a pattern.
137
Especially, the finding of the 4 KiB regions among the 62 GiB total memory
138
shows DAMON’s adaptive regions adjustment is working as designed.
140
Still the number of regions is close to the ``min_nr_regions``, and sizes of
141
cold regions are similar, though. Apparently it is improved, but it still has
142
rooms to improve.
144
400ms/8s intervals: Pretty Improved Results
145
===========================================
147
Increase the intervals four times (400 milliseconds and 8 seconds
148
for sampling and aggregation intervals, respectively). ::
150
# damo start -s 400ms -a 8s
151
# sleep 600
152
# damo record --snapshot 0 1
153
# damo stop
154
# damo report access --sort_regions_by temperature
155
0 addr 64.492 GiB size 1.508 GiB access 0 % age 6 m 48 s # coldest
156
1 addr 21.749 GiB size 5.674 GiB access 0 % age 6 m 8 s
157
2 addr 27.422 GiB size 5.801 GiB access 0 % age 6 m
158
3 addr 49.431 GiB size 8.675 GiB access 0 % age 5 m 28 s
159
4 addr 33.223 GiB size 5.645 GiB access 0 % age 5 m 12 s
160
5 addr 58.321 GiB size 6.170 GiB access 0 % age 5 m 4 s
161
[...]
162
25 addr 6.615 GiB size 297.531 MiB access 15 % age 0 ns
163
26 addr 9.513 GiB size 12.000 KiB access 20 % age 0 ns
164
27 addr 9.511 GiB size 108.000 KiB access 25 % age 0 ns
165
28 addr 9.513 GiB size 20.000 KiB access 25 % age 0 ns
166
29 addr 9.511 GiB size 12.000 KiB access 30 % age 0 ns
167
30 addr 9.520 GiB size 4.000 KiB access 40 % age 0 ns
168
[...]
169
41 addr 9.520 GiB size 4.000 KiB access 80 % age 56 s
170
42 addr 9.511 GiB size 12.000 KiB access 100 % age 6 m 16 s
171
43 addr 58.321 GiB size 4.000 KiB access 100 % age 6 m 24 s
172
44 addr 9.512 GiB size 4.000 KiB access 100 % age 6 m 48 s
173
45 addr 58.106 GiB size 4.000 KiB access 100 % age 6 m 48 s # hottest
174
total size: 62.000 GiB
175
# damo report access --style temperature-sz-hist
176
<temperature> <total size>
177
[-40,800,000,000, -32,639,999,000) 21.657 GiB |********************|
178
[-32,639,999,000, -24,479,998,000) 17.938 GiB |***************** |
179
[-24,479,998,000, -16,319,997,000) 16.885 GiB |**************** |
180
[-16,319,997,000, -8,159,996,000) 586.879 MiB |* |
181
[-8,159,996,000, 5,000) 4.946 GiB |***** |
182
[5,000, 8,160,006,000) 260.000 KiB |* |
183
[8,160,006,000, 16,320,007,000) 0 B | |
184
[16,320,007,000, 24,480,008,000) 0 B | |
185
[24,480,008,000, 32,640,009,000) 0 B | |
186
[32,640,009,000, 40,800,010,000) 16.000 KiB |* |
187
[40,800,010,000, 48,960,011,000) 8.000 KiB |* |
188
total size: 62.000 GiB
190
The number of regions having different access patterns has significantly
191
increased. Size of each region is also more varied. Total size of non-zero
192
access frequency regions is also significantly increased. Maybe this is already
193
good enough to make some meaningful memory management efficiency changes.
195
800ms/16s intervals: Another bias
196
=================================
198
Further double the intervals (800 milliseconds and 16 seconds for sampling
199
and aggregation intervals, respectively). The results is more improved for the
200
hot regions detection, but starts looking degrading cold regions detection. ::
202
# damo start -s 800ms -a 16s
203
# sleep 600
204
# damo record --snapshot 0 1
205
# damo stop
206
# damo report access --sort_regions_by temperature
207
0 addr 64.781 GiB size 1.219 GiB access 0 % age 4 m 48 s
208
1 addr 24.505 GiB size 2.475 GiB access 0 % age 4 m 16 s
209
2 addr 26.980 GiB size 504.273 MiB access 0 % age 4 m
210
3 addr 29.443 GiB size 2.462 GiB access 0 % age 4 m
211
4 addr 37.264 GiB size 5.645 GiB access 0 % age 4 m
212
5 addr 31.905 GiB size 5.359 GiB access 0 % age 3 m 44 s
213
[...]
214
20 addr 8.711 GiB size 40.000 KiB access 5 % age 2 m 40 s
215
21 addr 27.473 GiB size 1.970 GiB access 5 % age 4 m
216
22 addr 48.185 GiB size 4.625 GiB access 5 % age 4 m
217
23 addr 47.304 GiB size 902.117 MiB access 10 % age 4 m
218
24 addr 8.711 GiB size 4.000 KiB access 100 % age 4 m
219
25 addr 20.793 GiB size 3.713 GiB access 5 % age 4 m 16 s
220
26 addr 8.773 GiB size 4.000 KiB access 100 % age 4 m 16 s
221
total size: 62.000 GiB
222
# damo report access --style temperature-sz-hist
223
<temperature> <total size>
224
[-28,800,000,000, -23,359,999,000) 12.294 GiB |***************** |
225
[-23,359,999,000, -17,919,998,000) 9.753 GiB |************* |
226
[-17,919,998,000, -12,479,997,000) 15.131 GiB |********************|
227
[-12,479,997,000, -7,039,996,000) 0 B | |
228
[-7,039,996,000, -1,599,995,000) 7.506 GiB |********** |
229
[-1,599,995,000, 3,840,006,000) 6.127 GiB |********* |
230
[3,840,006,000, 9,280,007,000) 0 B | |
231
[9,280,007,000, 14,720,008,000) 136.000 KiB |* |
232
[14,720,008,000, 20,160,009,000) 40.000 KiB |* |
233
[20,160,009,000, 25,600,010,000) 11.188 GiB |*************** |
234
[25,600,010,000, 31,040,011,000) 4.000 KiB |* |
235
total size: 62.000 GiB
237
It found more non-zero access frequency regions. The number of regions is still
238
much higher than the ``min_nr_regions``, but it is reduced from that of the
239
previous setup. And apparently the distribution seems bit biased to hot
240
regions.
242
Conclusion
243
==========
245
With the above experimental tuning results, we can conclude the theory and the
246
guide makes sense to at least this workload, and could be applied to similar
247
cases.
3. 한국어 전문 번역
영어 원문의 문단 순서와 의미를 유지한 전체 번역입니다. 코드, 함수명, symbol과 URL은 원문 표기를 유지합니다.
목적과 시험 환경
1-18DAMON 모니터링 매개변수는 workload와 모니터링 목적에 맞게 조정해야 합니다. 이 문서는 monitoring parameter tuning guide를 실제로 적용한 예를 제공합니다.
예제는 Linux kernel v6.11의 DAMON과 DAMON 사용자 공간 도구 `damo` v2.5.9를 사용해 실제 server workload가 실행 중인 시스템의 물리 주소 공간 접근 패턴을 모니터링하고 시각화했습니다.
.. SPDX-License-Identifier: GPL-2.0
=================================================
DAMON Moniting Interval Parameters Tuning Example
=================================================
DAMON's monitoring parameters need tuning based on given workload and the
monitoring purpose. There is a :ref:`tuning guide
<damon_design_monitoring_params_tuning_guide>` for that. This document
provides an example tuning based on the guide.
Setup
=====
For below example, DAMON of Linux kernel v6.11 and `damo
<https://github.com/damonitor/damo>`_ (DAMON user-space tool) v2.5.9 was used to
monitor and visualize access patterns on the physical address space of a system
running a real-world server workload.5ms/100ms: 너무 짧은 interval
19-90먼저 기본 매개변수인 sampling 5ms와 aggregation 100ms로 시스템 물리 주소 공간의 접근 패턴 snapshot을 수집합니다. 시간에 따른 의미 있는 패턴이 나타나도록 DAMON을 시작한 뒤 snapshot을 얻기 전까지 10분 기다립니다.
# damo start
# sleep 600
# damo record --snapshot 0 1
# damo stop그다음 DAMON이 찾은 서로 다른 접근 패턴의 region을 `access temperature` 순서로 나열합니다. Access temperature는 region의 접근 빈도와 `age`를 가중합해 접근 열도를 나타냅니다. 접근 빈도가 0%이면 temperature에 -1을 곱하므로 접근되지 않은 region은 음수가 되고, 접근되지 않은 기간이 길수록 더 낮아집니다. 오름차순 정렬에서 맨 위가 가장 cold하고 맨 아래가 가장 hot합니다.
# damo report access --sort_regions_by temperature
0 addr 16.052 GiB size 5.985 GiB access 0 % age 5.900 s # coldest
1 addr 22.037 GiB size 6.029 GiB access 0 % age 5.300 s
2 addr 28.065 GiB size 6.045 GiB access 0 % age 5.200 s
3 addr 10.069 GiB size 5.983 GiB access 0 % age 4.500 s
4 addr 4.000 GiB size 6.069 GiB access 0 % age 4.400 s
5 addr 62.008 GiB size 3.992 GiB access 0 % age 3.700 s
6 addr 56.795 GiB size 5.213 GiB access 0 % age 3.300 s
7 addr 39.393 GiB size 6.096 GiB access 0 % age 2.800 s
8 addr 50.782 GiB size 6.012 GiB access 0 % age 2.800 s
9 addr 34.111 GiB size 5.282 GiB access 0 % age 2.300 s
10 addr 45.489 GiB size 5.293 GiB access 0 % age 1.800 s # hottest
total size: 62.000 GiB목록에는 뚜렷하게 hot한 region이 없고 접근 패턴 다양성도 최소 수준입니다. 모든 region의 접근 빈도가 0이며 region 수는 기본 `min_nr_regions` 값인 10이고 크기도 거의 같습니다. Adaptive regions adjustment가 제대로 동작하지 못했다고 의심할 수 있습니다.
지침이 말하듯 `age`를 최근성 정보로 사용하면 region의 상대적 hotness를 얻을 수 있어 아무 정보가 없는 것보다는 낫습니다. 그러나 10분가량 기다렸는데 가장 긴 `age`가 약 6초뿐이므로 얼마나 유용할지는 분명하지 않습니다.
Temperature 구간별 region 총크기 histogram도 흥미로운 분포 패턴을 보여 주지 않습니다.
# damo report access --style temperature-sz-hist
<temperature> <total size>
[-,590,000,000, -,549,000,000) 5.985 GiB |********** |
[-,549,000,000, -,508,000,000) 12.074 GiB |********************|
[-,508,000,000, -,467,000,000) 0 B | |
[-,467,000,000, -,426,000,000) 12.052 GiB |********************|
[-,426,000,000, -,385,000,000) 0 B | |
[-,385,000,000, -,344,000,000) 3.992 GiB |******* |
[-,344,000,000, -,303,000,000) 5.213 GiB |********* |
[-,303,000,000, -,262,000,000) 12.109 GiB |********************|
[-,262,000,000, -,221,000,000) 5.282 GiB |********* |
[-,221,000,000, -,180,000,000) 0 B | |
[-,180,000,000, -,139,000,000) 5.293 GiB |********* |
total size: 62.000 GiB요약하면 이 매개변수는 hot region 탐지 품질이 낮습니다. 조정 지침에 따르면 원인은 지나치게 짧은 `aggregation interval`입니다.
5ms/100ms intervals: Too Short Interval
=======================================
Let's start by capturing the access pattern snapshot on the physical address
space of the system using DAMON, with the default interval parameters (5
milliseconds and 100 milliseconds for the sampling and the aggregation
intervals, respectively). Wait ten minutes between the start of DAMON and
the capturing of the snapshot, to show a meaningful time-wise access patterns.
::
# damo start
# sleep 600
# damo record --snapshot 0 1
# damo stop
Then, list the DAMON-found regions of different access patterns, sorted by the
"access temperature". "Access temperature" is a metric representing the
access-hotness of a region. It is calculated as a weighted sum of the access
frequency and the age of the region. If the access frequency is 0 %, the
temperature is multiplied by minus one. That is, if a region is not accessed,
it gets minus temperature and it gets lower as not accessed for longer time.
The sorting is in temperature-ascendint order, so the region at the top of the
list is the coldest, and the one at the bottom is the hottest one. ::
# damo report access --sort_regions_by temperature
0 addr 16.052 GiB size 5.985 GiB access 0 % age 5.900 s # coldest
1 addr 22.037 GiB size 6.029 GiB access 0 % age 5.300 s
2 addr 28.065 GiB size 6.045 GiB access 0 % age 5.200 s
3 addr 10.069 GiB size 5.983 GiB access 0 % age 4.500 s
4 addr 4.000 GiB size 6.069 GiB access 0 % age 4.400 s
5 addr 62.008 GiB size 3.992 GiB access 0 % age 3.700 s
6 addr 56.795 GiB size 5.213 GiB access 0 % age 3.300 s
7 addr 39.393 GiB size 6.096 GiB access 0 % age 2.800 s
8 addr 50.782 GiB size 6.012 GiB access 0 % age 2.800 s
9 addr 34.111 GiB size 5.282 GiB access 0 % age 2.300 s
10 addr 45.489 GiB size 5.293 GiB access 0 % age 1.800 s # hottest
total size: 62.000 GiB
The list shows not seemingly hot regions, and only minimum access pattern
diversity. Every region has zero access frequency. The number of region is
10, which is the default ``min_nr_regions value``. Size of each region is also
nearly identical. We can suspect this is because “adaptive regions adjustment”
mechanism was not well working. As the guide suggested, we can get relative
hotness of regions using ``age`` as the recency information. That would be
better than nothing, but given the fact that the longest age is only about 6
seconds while we waited about ten minutes, it is unclear how useful this will
be.
The temperature ranges to total size of regions of each range histogram
visualization of the results also shows no interesting distribution pattern. ::
# damo report access --style temperature-sz-hist
<temperature> <total size>
[-,590,000,000, -,549,000,000) 5.985 GiB |********** |
[-,549,000,000, -,508,000,000) 12.074 GiB |********************|
[-,508,000,000, -,467,000,000) 0 B | |
[-,467,000,000, -,426,000,000) 12.052 GiB |********************|
[-,426,000,000, -,385,000,000) 0 B | |
[-,385,000,000, -,344,000,000) 3.992 GiB |******* |
[-,344,000,000, -,303,000,000) 5.213 GiB |********* |
[-,303,000,000, -,262,000,000) 12.109 GiB |********************|
[-,262,000,000, -,221,000,000) 5.282 GiB |********* |
[-,221,000,000, -,180,000,000) 0 B | |
[-,180,000,000, -,139,000,000) 5.293 GiB |********* |
total size: 62.000 GiB
In short, the parameters provide poor quality monitoring results for hot
regions detection. According to the :ref:`guide
<damon_design_monitoring_params_tuning_guide>`, this is due to the too short
aggregation interval.
100ms/2s: 작은 hot region이 보이기 시작함
91-143지침에 따라 sampling interval과 aggregation interval을 각각 100ms와 2s로 20배 늘립니다.
# damo start -s 100ms -a 2s
# sleep 600
# damo record --snapshot 0 1
# damo stop
# damo report access --sort_regions_by temperature
0 addr 10.180 GiB size 6.117 GiB access 0 % age 7 m 8 s # coldest
1 addr 49.275 GiB size 6.195 GiB access 0 % age 6 m 14 s
2 addr 62.421 GiB size 3.579 GiB access 0 % age 6 m 4 s
3 addr 40.154 GiB size 6.127 GiB access 0 % age 5 m 40 s
4 addr 16.296 GiB size 6.182 GiB access 0 % age 5 m 32 s
5 addr 34.254 GiB size 5.899 GiB access 0 % age 5 m 24 s
6 addr 46.281 GiB size 2.995 GiB access 0 % age 5 m 20 s
7 addr 28.420 GiB size 5.835 GiB access 0 % age 5 m 6 s
8 addr 4.000 GiB size 6.180 GiB access 0 % age 4 m 16 s
9 addr 22.478 GiB size 5.942 GiB access 0 % age 3 m 58 s
10 addr 55.470 GiB size 915.645 MiB access 0 % age 3 m 6 s
11 addr 56.364 GiB size 6.056 GiB access 0 % age 2 m 8 s
12 addr 56.364 GiB size 4.000 KiB access 95 % age 16 s
13 addr 49.275 GiB size 4.000 KiB access 100 % age 8 m 24 s # hottest
total size: 62.000 GiB
# damo report access --style temperature-sz-hist
<temperature> <total size>
[-42,800,000,000, -33,479,999,000) 22.018 GiB |***************** |
[-33,479,999,000, -24,159,998,000) 27.090 GiB |********************|
[-24,159,998,000, -14,839,997,000) 6.836 GiB |****** |
[-14,839,997,000, -5,519,996,000) 6.056 GiB |***** |
[-5,519,996,000, 3,800,005,000) 4.000 KiB |* |
[3,800,005,000, 13,120,006,000) 0 B | |
[13,120,006,000, 22,440,007,000) 0 B | |
[22,440,007,000, 31,760,008,000) 0 B | |
[31,760,008,000, 41,080,009,000) 0 B | |
[41,080,009,000, 50,400,010,000) 0 B | |
[50,400,010,000, 59,720,011,000) 4.000 KiB |* |
total size: 62.000 GiBDAMON은 뚜렷하게 hot한 4 KiB region 두 개를 찾았습니다. 이 region들의 `age`도 충분히 큽니다. 가장 hot한 4 KiB region은 약 8분 동안 해당 접근 빈도를 유지했고, 가장 cold한 region은 약 7분 동안 접근되지 않았습니다. Histogram 분포도 패턴을 보이기 시작합니다.
전체 62 GiB 메모리에서 4 KiB region을 찾아냈다는 사실은 DAMON adaptive regions adjustment가 설계대로 동작한다는 것을 보여 줍니다.
다만 region 수는 여전히 `min_nr_regions`에 가깝고 cold region 크기도 비슷합니다. 분명 개선됐지만 더 개선할 여지가 있습니다.
100ms/2s intervals: Starts Showing Small Hot Regions
====================================================
Following the guide, increase the interval 20 times (100 milliseocnds and 2
seconds for sampling and aggregation intervals, respectively). ::
# damo start -s 100ms -a 2s
# sleep 600
# damo record --snapshot 0 1
# damo stop
# damo report access --sort_regions_by temperature
0 addr 10.180 GiB size 6.117 GiB access 0 % age 7 m 8 s # coldest
1 addr 49.275 GiB size 6.195 GiB access 0 % age 6 m 14 s
2 addr 62.421 GiB size 3.579 GiB access 0 % age 6 m 4 s
3 addr 40.154 GiB size 6.127 GiB access 0 % age 5 m 40 s
4 addr 16.296 GiB size 6.182 GiB access 0 % age 5 m 32 s
5 addr 34.254 GiB size 5.899 GiB access 0 % age 5 m 24 s
6 addr 46.281 GiB size 2.995 GiB access 0 % age 5 m 20 s
7 addr 28.420 GiB size 5.835 GiB access 0 % age 5 m 6 s
8 addr 4.000 GiB size 6.180 GiB access 0 % age 4 m 16 s
9 addr 22.478 GiB size 5.942 GiB access 0 % age 3 m 58 s
10 addr 55.470 GiB size 915.645 MiB access 0 % age 3 m 6 s
11 addr 56.364 GiB size 6.056 GiB access 0 % age 2 m 8 s
12 addr 56.364 GiB size 4.000 KiB access 95 % age 16 s
13 addr 49.275 GiB size 4.000 KiB access 100 % age 8 m 24 s # hottest
total size: 62.000 GiB
# damo report access --style temperature-sz-hist
<temperature> <total size>
[-42,800,000,000, -33,479,999,000) 22.018 GiB |***************** |
[-33,479,999,000, -24,159,998,000) 27.090 GiB |********************|
[-24,159,998,000, -14,839,997,000) 6.836 GiB |****** |
[-14,839,997,000, -5,519,996,000) 6.056 GiB |***** |
[-5,519,996,000, 3,800,005,000) 4.000 KiB |* |
[3,800,005,000, 13,120,006,000) 0 B | |
[13,120,006,000, 22,440,007,000) 0 B | |
[22,440,007,000, 31,760,008,000) 0 B | |
[31,760,008,000, 41,080,009,000) 0 B | |
[41,080,009,000, 50,400,010,000) 0 B | |
[50,400,010,000, 59,720,011,000) 4.000 KiB |* |
total size: 62.000 GiB
DAMON found two distinct 4 KiB regions that pretty hot. The regions are also
well aged. The hottest 4 KiB region was keeping the access frequency for about
8 minutes, and the coldest region was keeping no access for about 7 minutes.
The distribution on the histogram also looks like having a pattern.
Especially, the finding of the 4 KiB regions among the 62 GiB total memory
shows DAMON’s adaptive regions adjustment is working as designed.
Still the number of regions is close to the ``min_nr_regions``, and sizes of
cold regions are similar, though. Apparently it is improved, but it still has
rooms to improve.
400ms/8s: 크게 개선된 결과
144-194Interval을 다시 네 배 늘려 sampling 400ms와 aggregation 8s로 설정합니다.
# damo start -s 400ms -a 8s
# sleep 600
# damo record --snapshot 0 1
# damo stop
# damo report access --sort_regions_by temperature
0 addr 64.492 GiB size 1.508 GiB access 0 % age 6 m 48 s # coldest
1 addr 21.749 GiB size 5.674 GiB access 0 % age 6 m 8 s
2 addr 27.422 GiB size 5.801 GiB access 0 % age 6 m
3 addr 49.431 GiB size 8.675 GiB access 0 % age 5 m 28 s
4 addr 33.223 GiB size 5.645 GiB access 0 % age 5 m 12 s
5 addr 58.321 GiB size 6.170 GiB access 0 % age 5 m 4 s
[...]
25 addr 6.615 GiB size 297.531 MiB access 15 % age 0 ns
26 addr 9.513 GiB size 12.000 KiB access 20 % age 0 ns
27 addr 9.511 GiB size 108.000 KiB access 25 % age 0 ns
28 addr 9.513 GiB size 20.000 KiB access 25 % age 0 ns
29 addr 9.511 GiB size 12.000 KiB access 30 % age 0 ns
30 addr 9.520 GiB size 4.000 KiB access 40 % age 0 ns
[...]
41 addr 9.520 GiB size 4.000 KiB access 80 % age 56 s
42 addr 9.511 GiB size 12.000 KiB access 100 % age 6 m 16 s
43 addr 58.321 GiB size 4.000 KiB access 100 % age 6 m 24 s
44 addr 9.512 GiB size 4.000 KiB access 100 % age 6 m 48 s
45 addr 58.106 GiB size 4.000 KiB access 100 % age 6 m 48 s # hottest
total size: 62.000 GiB
# damo report access --style temperature-sz-hist
<temperature> <total size>
[-40,800,000,000, -32,639,999,000) 21.657 GiB |********************|
[-32,639,999,000, -24,479,998,000) 17.938 GiB |***************** |
[-24,479,998,000, -16,319,997,000) 16.885 GiB |**************** |
[-16,319,997,000, -8,159,996,000) 586.879 MiB |* |
[-8,159,996,000, 5,000) 4.946 GiB |***** |
[5,000, 8,160,006,000) 260.000 KiB |* |
[8,160,006,000, 16,320,007,000) 0 B | |
[16,320,007,000, 24,480,008,000) 0 B | |
[24,480,008,000, 32,640,009,000) 0 B | |
[32,640,009,000, 40,800,010,000) 16.000 KiB |* |
[40,800,010,000, 48,960,011,000) 8.000 KiB |* |
total size: 62.000 GiB서로 다른 접근 패턴을 가진 region 수가 크게 증가했고 각 region 크기도 더 다양해졌습니다. 접근 빈도가 0이 아닌 region의 총크기도 크게 늘었습니다. 이 정도면 의미 있는 메모리 관리 효율 개선에 사용할 수 있을 가능성이 있습니다.
400ms/8s intervals: Pretty Improved Results
===========================================
Increase the intervals four times (400 milliseconds and 8 seconds
for sampling and aggregation intervals, respectively). ::
# damo start -s 400ms -a 8s
# sleep 600
# damo record --snapshot 0 1
# damo stop
# damo report access --sort_regions_by temperature
0 addr 64.492 GiB size 1.508 GiB access 0 % age 6 m 48 s # coldest
1 addr 21.749 GiB size 5.674 GiB access 0 % age 6 m 8 s
2 addr 27.422 GiB size 5.801 GiB access 0 % age 6 m
3 addr 49.431 GiB size 8.675 GiB access 0 % age 5 m 28 s
4 addr 33.223 GiB size 5.645 GiB access 0 % age 5 m 12 s
5 addr 58.321 GiB size 6.170 GiB access 0 % age 5 m 4 s
[...]
25 addr 6.615 GiB size 297.531 MiB access 15 % age 0 ns
26 addr 9.513 GiB size 12.000 KiB access 20 % age 0 ns
27 addr 9.511 GiB size 108.000 KiB access 25 % age 0 ns
28 addr 9.513 GiB size 20.000 KiB access 25 % age 0 ns
29 addr 9.511 GiB size 12.000 KiB access 30 % age 0 ns
30 addr 9.520 GiB size 4.000 KiB access 40 % age 0 ns
[...]
41 addr 9.520 GiB size 4.000 KiB access 80 % age 56 s
42 addr 9.511 GiB size 12.000 KiB access 100 % age 6 m 16 s
43 addr 58.321 GiB size 4.000 KiB access 100 % age 6 m 24 s
44 addr 9.512 GiB size 4.000 KiB access 100 % age 6 m 48 s
45 addr 58.106 GiB size 4.000 KiB access 100 % age 6 m 48 s # hottest
total size: 62.000 GiB
# damo report access --style temperature-sz-hist
<temperature> <total size>
[-40,800,000,000, -32,639,999,000) 21.657 GiB |********************|
[-32,639,999,000, -24,479,998,000) 17.938 GiB |***************** |
[-24,479,998,000, -16,319,997,000) 16.885 GiB |**************** |
[-16,319,997,000, -8,159,996,000) 586.879 MiB |* |
[-8,159,996,000, 5,000) 4.946 GiB |***** |
[5,000, 8,160,006,000) 260.000 KiB |* |
[8,160,006,000, 16,320,007,000) 0 B | |
[16,320,007,000, 24,480,008,000) 0 B | |
[24,480,008,000, 32,640,009,000) 0 B | |
[32,640,009,000, 40,800,010,000) 16.000 KiB |* |
[40,800,010,000, 48,960,011,000) 8.000 KiB |* |
total size: 62.000 GiB
The number of regions having different access patterns has significantly
increased. Size of each region is also more varied. Total size of non-zero
access frequency regions is also significantly increased. Maybe this is already
good enough to make some meaningful memory management efficiency changes.
800ms/16s: 다른 방향의 편향
195-241Interval을 다시 두 배로 늘려 sampling 800ms와 aggregation 16s로 설정합니다. Hot region 탐지는 더 개선되지만 cold region 탐지는 나빠지기 시작하는 모습입니다.
# damo start -s 800ms -a 16s
# sleep 600
# damo record --snapshot 0 1
# damo stop
# damo report access --sort_regions_by temperature
0 addr 64.781 GiB size 1.219 GiB access 0 % age 4 m 48 s
1 addr 24.505 GiB size 2.475 GiB access 0 % age 4 m 16 s
2 addr 26.980 GiB size 504.273 MiB access 0 % age 4 m
3 addr 29.443 GiB size 2.462 GiB access 0 % age 4 m
4 addr 37.264 GiB size 5.645 GiB access 0 % age 4 m
5 addr 31.905 GiB size 5.359 GiB access 0 % age 3 m 44 s
[...]
20 addr 8.711 GiB size 40.000 KiB access 5 % age 2 m 40 s
21 addr 27.473 GiB size 1.970 GiB access 5 % age 4 m
22 addr 48.185 GiB size 4.625 GiB access 5 % age 4 m
23 addr 47.304 GiB size 902.117 MiB access 10 % age 4 m
24 addr 8.711 GiB size 4.000 KiB access 100 % age 4 m
25 addr 20.793 GiB size 3.713 GiB access 5 % age 4 m 16 s
26 addr 8.773 GiB size 4.000 KiB access 100 % age 4 m 16 s
total size: 62.000 GiB
# damo report access --style temperature-sz-hist
<temperature> <total size>
[-28,800,000,000, -23,359,999,000) 12.294 GiB |***************** |
[-23,359,999,000, -17,919,998,000) 9.753 GiB |************* |
[-17,919,998,000, -12,479,997,000) 15.131 GiB |********************|
[-12,479,997,000, -7,039,996,000) 0 B | |
[-7,039,996,000, -1,599,995,000) 7.506 GiB |********** |
[-1,599,995,000, 3,840,006,000) 6.127 GiB |********* |
[3,840,006,000, 9,280,007,000) 0 B | |
[9,280,007,000, 14,720,008,000) 136.000 KiB |* |
[14,720,008,000, 20,160,009,000) 40.000 KiB |* |
[20,160,009,000, 25,600,010,000) 11.188 GiB |*************** |
[25,600,010,000, 31,040,011,000) 4.000 KiB |* |
total size: 62.000 GiB접근 빈도가 0이 아닌 region을 더 많이 찾았습니다. Region 수는 여전히 `min_nr_regions`보다 훨씬 많지만 이전 설정보다 줄었고, 분포가 hot region 쪽으로 조금 편향된 것으로 보입니다.
800ms/16s intervals: Another bias
=================================
Further double the intervals (800 milliseconds and 16 seconds for sampling
and aggregation intervals, respectively). The results is more improved for the
hot regions detection, but starts looking degrading cold regions detection. ::
# damo start -s 800ms -a 16s
# sleep 600
# damo record --snapshot 0 1
# damo stop
# damo report access --sort_regions_by temperature
0 addr 64.781 GiB size 1.219 GiB access 0 % age 4 m 48 s
1 addr 24.505 GiB size 2.475 GiB access 0 % age 4 m 16 s
2 addr 26.980 GiB size 504.273 MiB access 0 % age 4 m
3 addr 29.443 GiB size 2.462 GiB access 0 % age 4 m
4 addr 37.264 GiB size 5.645 GiB access 0 % age 4 m
5 addr 31.905 GiB size 5.359 GiB access 0 % age 3 m 44 s
[...]
20 addr 8.711 GiB size 40.000 KiB access 5 % age 2 m 40 s
21 addr 27.473 GiB size 1.970 GiB access 5 % age 4 m
22 addr 48.185 GiB size 4.625 GiB access 5 % age 4 m
23 addr 47.304 GiB size 902.117 MiB access 10 % age 4 m
24 addr 8.711 GiB size 4.000 KiB access 100 % age 4 m
25 addr 20.793 GiB size 3.713 GiB access 5 % age 4 m 16 s
26 addr 8.773 GiB size 4.000 KiB access 100 % age 4 m 16 s
total size: 62.000 GiB
# damo report access --style temperature-sz-hist
<temperature> <total size>
[-28,800,000,000, -23,359,999,000) 12.294 GiB |***************** |
[-23,359,999,000, -17,919,998,000) 9.753 GiB |************* |
[-17,919,998,000, -12,479,997,000) 15.131 GiB |********************|
[-12,479,997,000, -7,039,996,000) 0 B | |
[-7,039,996,000, -1,599,995,000) 7.506 GiB |********** |
[-1,599,995,000, 3,840,006,000) 6.127 GiB |********* |
[3,840,006,000, 9,280,007,000) 0 B | |
[9,280,007,000, 14,720,008,000) 136.000 KiB |* |
[14,720,008,000, 20,160,009,000) 40.000 KiB |* |
[20,160,009,000, 25,600,010,000) 11.188 GiB |*************** |
[25,600,010,000, 31,040,011,000) 4.000 KiB |* |
total size: 62.000 GiB
It found more non-zero access frequency regions. The number of regions is still
much higher than the ``min_nr_regions``, but it is reduced from that of the
previous setup. And apparently the distribution seems bit biased to hot
regions.
결론
242-247위 실험 조정 결과를 바탕으로 DAMON interval 조정 이론과 지침은 적어도 이 workload에서 타당하며 비슷한 사례에도 적용할 수 있다고 결론 내릴 수 있습니다.
Conclusion
==========
With the above experimental tuning results, we can conclude the theory and the
guide makes sense to at least this workload, and could be applied to similar
cases.
요약·해설
monitoring_intervals_tuning_example.rst:1-247같은 1:20 비율을 유지해도 absolute interval이 너무 짧으면 aggregation마다 의미 있는 접근량이 쌓이지 않아 region이 기본 최소 개수와 비슷한 큰 덩어리에 머뭅니다. Interval을 늘리면 작은 hot region과 오래 유지된 cold region이 분리되지만, 지나치게 늘리면 hot 쪽 편향이 생길 수 있습니다.
62 GiB 물리 메모리에서 네 설정의 region 분해와 탐지 특성을 비교합니다.
Temperature는 접근 빈도와 age의 가중합이며 0% 접근에는 음수 부호가 붙습니다.
Region 다양성과 hot·cold 분리 품질을 함께 보며 aggregation interval을 조정합니다.
출력에서 조정 성공 여부를 판단할 때 함께 확인할 항목입니다.
이 예제의 측정 절차를 재사용 가능한 단계로 정리했습니다.