1 This is cpufreq-bench, a microbenchmark for th 1 This is cpufreq-bench, a microbenchmark for the cpufreq framework.
2 2
3 Purpose 3 Purpose
4 ======= 4 =======
5 5
6 What is this benchmark for: 6 What is this benchmark for:
7 - Identify worst case performance loss when 7 - Identify worst case performance loss when doing dynamic frequency
8 scaling using Linux kernel governors 8 scaling using Linux kernel governors
9 - Identify average reaction time of a govern 9 - Identify average reaction time of a governor to CPU load changes
10 - (Stress) Testing whether a cpufreq low lev 10 - (Stress) Testing whether a cpufreq low level driver or governor works
11 as expected 11 as expected
12 - Identify cpufreq related performance regre 12 - Identify cpufreq related performance regressions between kernels
13 - Possibly Real time priority testing? -> wh 13 - Possibly Real time priority testing? -> what happens if there are
14 processes with a higher prio than the gove 14 processes with a higher prio than the governor's kernel thread
15 - ... 15 - ...
16 16
17 What this benchmark does *not* cover: 17 What this benchmark does *not* cover:
18 - Power saving related regressions (In fact 18 - Power saving related regressions (In fact as better the performance
19 throughput is, the worse the power savings 19 throughput is, the worse the power savings will be, but the first should
20 mostly count more...) 20 mostly count more...)
21 - Real world (workloads) 21 - Real world (workloads)
22 22
23 23
24 Description 24 Description
25 =========== 25 ===========
26 26
27 cpufreq-bench helps to test the condition of a 27 cpufreq-bench helps to test the condition of a given cpufreq governor.
28 For that purpose, it compares the performance 28 For that purpose, it compares the performance governor to a configured
29 powersave module. 29 powersave module.
30 30
31 31
32 How it works 32 How it works
33 ============ 33 ============
34 You can specify load (100% CPU load) and sleep 34 You can specify load (100% CPU load) and sleep (0% CPU load) times in us which
35 will be run X time in a row (cycles): 35 will be run X time in a row (cycles):
36 36
37 sleep=25000 37 sleep=25000
38 load=25000 38 load=25000
39 cycles=20 39 cycles=20
40 40
41 This part of the configuration file will creat 41 This part of the configuration file will create 25ms load/sleep turns,
42 repeated 20 times. 42 repeated 20 times.
43 43
44 Adding this: 44 Adding this:
45 sleep_step=25000 45 sleep_step=25000
46 load_step=25000 46 load_step=25000
47 rounds=5 47 rounds=5
48 Will increase load and sleep time by 25ms 5 ti 48 Will increase load and sleep time by 25ms 5 times.
49 Together you get following test: 49 Together you get following test:
50 25ms load/sleep time repeated 20 times (cycle 50 25ms load/sleep time repeated 20 times (cycles).
51 50ms load/sleep time repeated 20 times (cycle 51 50ms load/sleep time repeated 20 times (cycles).
52 .. 52 ..
53 100ms load/sleep time repeated 20 times (cycle 53 100ms load/sleep time repeated 20 times (cycles).
54 54
55 First it is calibrated how long a specific CPU 55 First it is calibrated how long a specific CPU intensive calculation
56 takes on this machine and needs to be run in a 56 takes on this machine and needs to be run in a loop using the performance
57 governor. 57 governor.
58 Then the above test runs are processed using t 58 Then the above test runs are processed using the performance governor
59 and the governor to test. The time the calcula 59 and the governor to test. The time the calculation really needed
60 with the dynamic freq scaling governor is comp 60 with the dynamic freq scaling governor is compared with the time needed
61 on full performance and you get the overall pe 61 on full performance and you get the overall performance loss.
62 62
63 63
64 Example of expected results with ondemand gove 64 Example of expected results with ondemand governor:
65 65
66 This shows expected results of the first two t 66 This shows expected results of the first two test run rounds from
67 above config, you there have: 67 above config, you there have:
68 68
69 100% CPU load (load) | 0 % CPU load (sleep) | 69 100% CPU load (load) | 0 % CPU load (sleep) | round
70 25 ms | 25 ms | 70 25 ms | 25 ms | 1
71 50 ms | 50 ms | 71 50 ms | 50 ms | 2
72 72
73 For example if ondemand governor is configured 73 For example if ondemand governor is configured to have a 50ms
74 sampling rate you get: 74 sampling rate you get:
75 75
76 In round 1, ondemand should have rather static 76 In round 1, ondemand should have rather static 50% load and probably
77 won't ever switch up (as long as up_threshold 77 won't ever switch up (as long as up_threshold is above).
78 78
79 In round 2, if the ondemand sampling times exa 79 In round 2, if the ondemand sampling times exactly match the load/sleep
80 trigger of the cpufreq-bench, you will see no 80 trigger of the cpufreq-bench, you will see no performance loss (compare with
81 below possible ondemand sample kick ins (1)): 81 below possible ondemand sample kick ins (1)):
82 82
83 But if ondemand always kicks in in the middle 83 But if ondemand always kicks in in the middle of the load sleep cycles, it
84 will always see 50% loads and you get worst pe 84 will always see 50% loads and you get worst performance impact never
85 switching up (compare with below possible onde 85 switching up (compare with below possible ondemand sample kick ins (2))::
86 86
87 50 50 50 50ms ->time 87 50 50 50 50ms ->time
88 load -----| |-----| |-----| |----- 88 load -----| |-----| |-----| |-----|
89 | | | | | | 89 | | | | | | |
90 sleep |-----| |-----| |-----| 90 sleep |-----| |-----| |-----| |----
91 |-----|-----|-----|-----|-----|-----|----- 91 |-----|-----|-----|-----|-----|-----|-----|---- ondemand sampling (1)
92 100 0 100 0 100 0 1 92 100 0 100 0 100 0 100 load seen by ondemand(%)
93 |-----|-----|-----|-----|-----|-----|-- 93 |-----|-----|-----|-----|-----|-----|-----|-- ondemand sampling (2)
94 50 50 50 50 50 50 50 94 50 50 50 50 50 50 50 load seen by ondemand(%)
95 95
96 You can easily test all kind of load/sleep tim 96 You can easily test all kind of load/sleep times and check whether your
97 governor in average behaves as expected. 97 governor in average behaves as expected.
98 98
99 99
100 ToDo 100 ToDo
101 ==== 101 ====
102 102
103 Provide a gnuplot utility script for easy gene 103 Provide a gnuplot utility script for easy generation of plots to present
104 the outcome nicely. 104 the outcome nicely.
105 105
106 106
107 cpufreq-bench Command Usage 107 cpufreq-bench Command Usage
108 =========================== 108 ===========================
109 -l, --load=<long int> initial load t 109 -l, --load=<long int> initial load time in us
110 -s, --sleep=<long int> initial sleep 110 -s, --sleep=<long int> initial sleep time in us
111 -x, --load-step=<long int> time to be add 111 -x, --load-step=<long int> time to be added to load time, in us
112 -y, --sleep-step=<long int> time to be add 112 -y, --sleep-step=<long int> time to be added to sleep time, in us
113 -c, --cpu=<unsigned int> CPU Number to 113 -c, --cpu=<unsigned int> CPU Number to use, starting at 0
114 -p, --prio=<priority> scheduler prio 114 -p, --prio=<priority> scheduler priority, HIGH, LOW or DEFAULT
115 -g, --governor=<governor> cpufreq govern 115 -g, --governor=<governor> cpufreq governor to test
116 -n, --cycles=<int> load/sleep cyc !! 116 -n, --cycles=<int> load/sleep cycles to get an avarage value to compare
117 -r, --rounds<int> load/sleep rou 117 -r, --rounds<int> load/sleep rounds
118 -f, --file=<configfile> config file to 118 -f, --file=<configfile> config file to use
119 -o, --output=<dir> output dir, mu 119 -o, --output=<dir> output dir, must exist
120 -v, --verbose verbose output 120 -v, --verbose verbose output on/off
121 121
122 Due to the high priority, the application may 122 Due to the high priority, the application may not be responsible for some time.
123 After the benchmark, the logfile is saved in O 123 After the benchmark, the logfile is saved in OUTPUTDIR/benchmark_TIMESTAMP.log
124 124
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.