Zainstalowałem sysbench poleceniem:
sudo apt-get install sysbenchi wykonałem dla każdego z kerneli następujące pomiary:
sysbench --num-threads=$(grep 'model name' /proc/cpuinfo | wc -l ) --test=cpu run > testPierwsze polecenie bada CPU, a drugie Pamięć. Ten tajemniczy na pierwszy rzut oka zapis:
sysbench --num-threads=$(grep 'model name' /proc/cpuinfo | wc -l ) --test=memory run >> test
$(grep 'model name' /proc/cpuinfo | wc -l )odpowiedzialny jest za sprawdzenie ilu wątkowa jest nasza maszyna, teraz całe polecenie powinno być już zrozumiałe:)
A oto wspomniane wyniki:
Linux yoga 3.15.0-ext73-30.0-i7-pro #17 SMP PREEMPT Sun Jun 15 16:57:55 CEST 2014 x86_64 x86_64 x86_64 GNU/Linux
sysbench 0.4.12: multi-threaded system evaluation benchmarkLinux yoga 3.14.6-ext73-f1-29.6-i7-pro-ags-cfs #5 SMP PREEMPT Mon Jun 9 18:29:44 CEST 2014 x86_64 x86_64 x86_64 GNU/Linux
Running the test with following options:
Number of threads: 4
Doing CPU performance benchmark
Threads started!
Done.
Maximum prime number checked in CPU test: 10000
Test execution summary:
total time: 3.8329s
total number of events: 10000
total time taken by event execution: 15.3234
per-request statistics:
min: 1.35ms
avg: 1.53ms
max: 16.31ms
approx. 95 percentile: 1.54ms
Threads fairness:
events (avg/stddev): 2500.0000/15.54
execution time (avg/stddev): 3.8309/0.00
sysbench 0.4.12: multi-threaded system evaluation benchmark
Running the test with following options:
Number of threads: 4
Doing memory operations speed test
Memory block size: 1K
Memory transfer size: 102400M
Memory operations type: write
Memory scope type: global
Threads started!
Done.
Operations performed: 104857600 (3901287.11 ops/sec)
102400.00 MB transferred (3809.85 MB/sec)
Test execution summary:
total time: 26.8777s
total number of events: 104857600
total time taken by event execution: 71.9798
per-request statistics:
min: 0.00ms
avg: 0.00ms
max: 16.44ms
approx. 95 percentile: 0.00ms
Threads fairness:
events (avg/stddev): 26214400.0000/297309.43
execution time (avg/stddev): 17.9949/0.08
sysbench 0.4.12: multi-threaded system evaluation benchmarkLinux yoga 3.14.6-ext73-f1-29.6-k8-ags-cfs #8 SMP PREEMPT Mon Jun 9 21:49:51 CEST 2014 x86_64 x86_64 x86_64 GNU/Linux
Running the test with following options:
Number of threads: 4
Doing CPU performance benchmark
Threads started!
Done.
Maximum prime number checked in CPU test: 10000
Test execution summary:
total time: 5.0742s
total number of events: 10000
total time taken by event execution: 20.2616
per-request statistics:
min: 1.48ms
avg: 2.03ms
max: 40.00ms
approx. 95 percentile: 4.53ms
Threads fairness:
events (avg/stddev): 2500.0000/230.04
execution time (avg/stddev): 5.0654/0.01
sysbench 0.4.12: multi-threaded system evaluation benchmark
Running the test with following options:
Number of threads: 4
Doing memory operations speed test
Memory block size: 1K
Memory transfer size: 102400M
Memory operations type: write
Memory scope type: global
Threads started!
Done.
Operations performed: 104857600 (3834313.32 ops/sec)
102400.00 MB transferred (3744.45 MB/sec)
Test execution summary:
total time: 27.3472s
total number of events: 104857600
total time taken by event execution: 75.2566
per-request statistics:
min: 0.00ms
avg: 0.00ms
max: 10.01ms
approx. 95 percentile: 0.00ms
Threads fairness:
events (avg/stddev): 26214400.0000/598046.46
execution time (avg/stddev): 18.8141/0.15
sysbench 0.4.12: multi-threaded system evaluation benchmark
Running the test with following options:
Number of threads: 4
Doing CPU performance benchmark
Threads started!
Done.
Maximum prime number checked in CPU test: 10000
Test execution summary:
total time: 5.1035s
total number of events: 10000
total time taken by event execution: 20.3904
per-request statistics:
min: 1.46ms
avg: 2.04ms
max: 21.52ms
approx. 95 percentile: 6.31ms
Threads fairness:
events (avg/stddev): 2500.0000/239.54
execution time (avg/stddev): 5.0976/0.01
sysbench 0.4.12: multi-threaded system evaluation benchmark
Running the test with following options:
Number of threads: 4
Doing memory operations speed test
Memory block size: 1K
Memory transfer size: 102400M
Memory operations type: write
Memory scope type: global
Threads started!
Done.
Operations performed: 104857600 (3823292.28 ops/sec)
102400.00 MB transferred (3733.68 MB/sec)
Test execution summary:
total time: 27.4260s
total number of events: 104857600
total time taken by event execution: 74.2108
per-request statistics:
min: 0.00ms
avg: 0.00ms
max: 19.12ms
approx. 95 percentile: 0.00ms
Threads fairness:
events (avg/stddev): 26214400.0000/389813.55
execution time (avg/stddev): 18.5527/0.08
Linux yoga 3.13.0-29-generic #53-Ubuntu SMP Wed Jun 4 21:00:20 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
sysbench 0.4.12: multi-threaded system evaluation benchmarkróżnice na moim sprzęcie mnie zaskoczyły, fakt, że są widoczne gołym okiem przy zestawieniu z tymi liczbami jest zrozumiały:)
Running the test with following options:
Number of threads: 4
Doing CPU performance benchmark
Threads started!
Done.
Maximum prime number checked in CPU test: 10000
Test execution summary:
total time: 15.5585s
total number of events: 10000
total time taken by event execution: 62.2050
per-request statistics:
min: 4.29ms
avg: 6.22ms
max: 40.43ms
approx. 95 percentile: 16.35ms
Threads fairness:
events (avg/stddev): 2500.0000/78.06
execution time (avg/stddev): 15.5512/0.00
sysbench 0.4.12: multi-threaded system evaluation benchmark
Running the test with following options:
Number of threads: 4
Doing memory operations speed test
Memory block size: 1K
Memory transfer size: 102400M
Memory operations type: write
Memory scope type: global
Threads started!
Done.
Operations performed: 104857600 (1121182.06 ops/sec)
102400.00 MB transferred (1094.90 MB/sec)
Test execution summary:
total time: 93.5242s
total number of events: 104857600
total time taken by event execution: 237.4461
per-request statistics:
min: 0.00ms
avg: 0.00ms
max: 30.95ms
approx. 95 percentile: 0.00ms
Threads fairness:
events (avg/stddev): 26214400.0000/804010.20
execution time (avg/stddev): 59.3615/0.83
Namawiam do sprawdzenia na waszych zabawkach. Może zdobędę się na heroiczny wyczyn i zbuduję coś co ładnie to zbierze w tabelkę:) może jakiś "ficzer" NeteXta powstanie w ten sposób w przyszłości :)
