前回の投稿で、RaspberryPi3が手に入ったので、UnixBenchを用いてRaspberryPi2とRaspberryPi3のベンチマーク比較をしたいと思います。
ベンチマークした Raspberry Pi 2のOS
$ cat /etc/issue Raspbian GNU/Linux8\n\l
$ cat /etc/debian_version 8.0
Raspberry Pi 2 のベンチマーク結果
「Raspberry Pi 2」の「System Benchmarks Index Score」は「301.2」。
Benchmark Run: Mon Apr 04 2016 13:19:37 - 13:47:49
4 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 2601790.1 lps (10.0 s, 7 samples)
Double-Precision Whetstone 422.0 MWIPS (9.7 s, 7 samples)
Execl Throughput 331.3 lps (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 58975.8 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 16781.0 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 141826.6 KBps (30.0 s, 2 samples)
Pipe Throughput 185027.4 lps (10.0 s, 7 samples)
Pipe-based Context Switching 32491.1 lps (10.0 s, 7 samples)
Process Creation 904.2 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 792.0 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 224.0 lpm (60.1 s, 2 samples)
System Call Overhead 429559.1 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 2601790.1 222.9
Double-Precision Whetstone 55.0 422.0 76.7
Execl Throughput 43.0 331.3 77.0
File Copy 1024 bufsize 2000 maxblocks 3960.0 58975.8 148.9
File Copy 256 bufsize 500 maxblocks 1655.0 16781.0 101.4
File Copy 4096 bufsize 8000 maxblocks 5800.0 141826.6 244.5
Pipe Throughput 12440.0 185027.4 148.7
Pipe-based Context Switching 4000.0 32491.1 81.2
Process Creation 126.0 904.2 71.8
Shell Scripts (1 concurrent) 42.4 792.0 186.8
Shell Scripts (8 concurrent) 6.0 224.0 373.3
System Call Overhead 15000.0 429559.1 286.4
========
System Benchmarks Index Score 144.7
------------------------------------------------------------------------
Benchmark Run: Mon Apr 04 2016 13:47:49 - 14:16:13
4 CPUs in system; running 4 parallel copies of tests
Dhrystone 2 using register variables 8014979.3 lps (10.0 s, 7 samples)
Double-Precision Whetstone 1171.0 MWIPS (9.9 s, 7 samples)
Execl Throughput 899.6 lps (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 82349.3 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 22653.7 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 231590.8 KBps (30.0 s, 2 samples)
Pipe Throughput 514012.5 lps (10.0 s, 7 samples)
Pipe-based Context Switching 81857.2 lps (10.0 s, 7 samples)
Process Creation 1934.6 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1787.8 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 235.0 lpm (60.5 s, 2 samples)
System Call Overhead 1122629.1 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 8014979.3 686.8
Double-Precision Whetstone 55.0 1171.0 212.9
Execl Throughput 43.0 899.6 209.2
File Copy 1024 bufsize 2000 maxblocks 3960.0 82349.3 208.0
File Copy 256 bufsize 500 maxblocks 1655.0 22653.7 136.9
File Copy 4096 bufsize 8000 maxblocks 5800.0 231590.8 399.3
Pipe Throughput 12440.0 514012.5 413.2
Pipe-based Context Switching 4000.0 81857.2 204.6
Process Creation 126.0 1934.6 153.5
Shell Scripts (1 concurrent) 42.4 1787.8 421.7
Shell Scripts (8 concurrent) 6.0 235.0 391.7
System Call Overhead 15000.0 1122629.1 748.4
========
System Benchmarks Index Score 301.2
Raspberry Pi 3 のベンチマーク結果
「Raspberry Pi 3」の「System Benchmarks Index Score」は「628.3」。
Benchmark Run: Thu Mar 31 2016 11:08:57 - 11:37:18
4 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 4361979.0 lps (10.0 s, 7 samples)
Double-Precision Whetstone 740.8 MWIPS (9.8 s, 7 samples)
Execl Throughput 491.7 lps (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 143567.0 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 42055.4 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 348281.5 KBps (30.0 s, 2 samples)
Pipe Throughput 308168.0 lps (10.0 s, 7 samples)
Pipe-based Context Switching 60826.0 lps (10.0 s, 7 samples)
Process Creation 2582.8 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1997.8 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 525.6 lpm (60.1 s, 2 samples)
System Call Overhead 656890.3 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 4361979.0 373.8
Double-Precision Whetstone 55.0 740.8 134.7
Execl Throughput 43.0 491.7 114.4
File Copy 1024 bufsize 2000 maxblocks 3960.0 143567.0 362.5
File Copy 256 bufsize 500 maxblocks 1655.0 42055.4 254.1
File Copy 4096 bufsize 8000 maxblocks 5800.0 348281.5 600.5
Pipe Throughput 12440.0 308168.0 247.7
Pipe-based Context Switching 4000.0 60826.0 152.1
Process Creation 126.0 2582.8 205.0
Shell Scripts (1 concurrent) 42.4 1997.8 471.2
Shell Scripts (8 concurrent) 6.0 525.6 876.1
System Call Overhead 15000.0 656890.3 437.9
========
System Benchmarks Index Score 295.6
------------------------------------------------------------------------
Benchmark Run: Thu Mar 31 2016 11:37:18 - 12:05:55
4 CPUs in system; running 4 parallel copies of tests
Dhrystone 2 using register variables 14287992.5 lps (10.0 s, 7 samples)
Double-Precision Whetstone 2557.7 MWIPS (11.4 s, 7 samples)
Execl Throughput 2019.6 lps (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 185498.2 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 51870.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 497220.5 KBps (30.0 s, 2 samples)
Pipe Throughput 1003033.9 lps (10.0 s, 7 samples)
Pipe-based Context Switching 178203.7 lps (10.0 s, 7 samples)
Process Creation 4342.2 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 3505.8 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 451.5 lpm (60.3 s, 2 samples)
System Call Overhead 2190375.0 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 14287992.5 1224.3
Double-Precision Whetstone 55.0 2557.7 465.0
Execl Throughput 43.0 2019.6 469.7
File Copy 1024 bufsize 2000 maxblocks 3960.0 185498.2 468.4
File Copy 256 bufsize 500 maxblocks 1655.0 51870.5 313.4
File Copy 4096 bufsize 8000 maxblocks 5800.0 497220.5 857.3
Pipe Throughput 12440.0 1003033.9 806.3
Pipe-based Context Switching 4000.0 178203.7 445.5
Process Creation 126.0 4342.2 344.6
Shell Scripts (1 concurrent) 42.4 3505.8 826.8
Shell Scripts (8 concurrent) 6.0 451.5 752.6
System Call Overhead 15000.0 2190375.0 1460.2
========
System Benchmarks Index Score 628.3
RPi3のCPU性能はRPi2と比較して50％処理速度が向上らしいです。
個人的には処理速度向上より、Wi-Fiになったことで、LAN線の取り回しが必要なく非常に便利になりました。
