(yet another) UnixBench on Bash on Ubuntu on Windows

More than 1 year has passed since last update.

fujiedaさんの「UnixBench on Bash on Ubuntu on Windows」を見て自分の環境(Core i5-2410M、メモリ8GB、ハイブリッドHDD)でもやってみた。


Build 16251 (16251.rs3_release.170721-2122)

このビルドからWindows Subsystems for Linuxのベータが取れました。UnixBenchはいつものようにエラーが出たので-i 3オプション付きで実行して、スコアは185.1。また少し回復してますが、200台に戻すことはなさそうですね。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 4.4.0-43-Microsoft -- #1-Microsoft Wed Dec 31 14:42:53 PST 2014
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
11:23:20 up 9:54, 0 users, load average: 0.52, 0.58, 0.59; runlevel

------------------------------------------------------------------------
Benchmark Run: 土 7月 29 2017 11:32:25 - 11:41:39
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 64611328.4 lps (10.0 s, 2 samples)
Double-Precision Whetstone 10654.8 MWIPS (9.8 s, 2 samples)
Execl Throughput 340.5 lps (29.2 s, 1 samples)
File Copy 1024 bufsize 2000 maxblocks 28695.0 KBps (30.0 s, 1 samples)
File Copy 256 bufsize 500 maxblocks 7502.0 KBps (30.0 s, 1 samples)
File Copy 4096 bufsize 8000 maxblocks 102182.0 KBps (30.0 s, 1 samples)
Pipe Throughput 120870.2 lps (10.0 s, 2 samples)
Pipe-based Context Switching 51406.8 lps (10.0 s, 2 samples)
Process Creation 754.5 lps (30.0 s, 1 samples)
Shell Scripts (1 concurrent) 848.0 lpm (60.1 s, 1 samples)
Shell Scripts (8 concurrent) 143.5 lpm (61.0 s, 1 samples)
System Call Overhead 138146.4 lps (10.0 s, 2 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 64611328.4 5536.5
Double-Precision Whetstone 55.0 10654.8 1937.2
Execl Throughput 43.0 340.5 79.2
File Copy 1024 bufsize 2000 maxblocks 3960.0 28695.0 72.5
File Copy 256 bufsize 500 maxblocks 1655.0 7502.0 45.3
File Copy 4096 bufsize 8000 maxblocks 5800.0 102182.0 176.2
Pipe Throughput 12440.0 120870.2 97.2
Pipe-based Context Switching 4000.0 51406.8 128.5
Process Creation 126.0 754.5 59.9
Shell Scripts (1 concurrent) 42.4 848.0 200.0
Shell Scripts (8 concurrent) 6.0 143.5 239.2
System Call Overhead 15000.0 138146.4 92.1
========
System Benchmarks Index Score 185.1


Build 16188 (16188.rs_prerelease.170430-1928)

Creator's Update後のInsider Previewです。-i 5オプション付きで実行して、スコアは166.5。

WSLをインストールし直したのが良かったのかわからないが、少し回復。とはいえ、初期のスコアには到底届かないですね。いろいろ実装するとやっぱり遅くなってしまうのでしょう。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 4.4.0-43-Microsoft -- #1-Microsoft Wed Dec 31 14:42:53 PST 2014
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
10:28:39 up 1:00, 0 users, load average: 0.52, 0.58, 0.59; runlevel unknown

------------------------------------------------------------------------
Benchmark Run: Sat May 06 2017 10:45:30 - 11:02:34
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 62009078.8 lps (10.1 s, 4 samples)
Double-Precision Whetstone 7186.5 MWIPS (10.1 s, 4 samples)
Execl Throughput 286.0 lps (29.4 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 31765.4 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 8143.0 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 112893.4 KBps (30.0 s, 2 samples)
Pipe Throughput 123076.8 lps (10.1 s, 4 samples)
Pipe-based Context Switching 52033.2 lps (10.0 s, 4 samples)
Process Creation 396.1 lps (30.1 s, 2 samples)
Shell Scripts (1 concurrent) 786.1 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 110.3 lpm (60.8 s, 2 samples)
System Call Overhead 138638.6 lps (10.1 s, 4 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 62009078.8 5313.5
Double-Precision Whetstone 55.0 7186.5 1306.6
Execl Throughput 43.0 286.0 66.5
File Copy 1024 bufsize 2000 maxblocks 3960.0 31765.4 80.2
File Copy 256 bufsize 500 maxblocks 1655.0 8143.0 49.2
File Copy 4096 bufsize 8000 maxblocks 5800.0 112893.4 194.6
Pipe Throughput 12440.0 123076.8 98.9
Pipe-based Context Switching 4000.0 52033.2 130.1
Process Creation 126.0 396.1 31.4
Shell Scripts (1 concurrent) 42.4 786.1 185.4
Shell Scripts (8 concurrent) 6.0 110.3 183.9
System Call Overhead 15000.0 138638.6 92.4
========
System Benchmarks Index Score 166.5


Build 15025 (15025.rs_prerelease.170127-1750)

-i 4 オプションを指定するとエラーが出たので、 -i 3 オプションを指定。スコアは148.4。実装が進んできているようだが、どんどん遅くなっている。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 4.4.0-43-Microsoft -- #1-Microsoft Wed Dec 31 14:42:53 PST 2014
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
22:39:05 up 38 min, 0 users, load average: 0.52, 0.58, 0.59; runlevel unknown

------------------------------------------------------------------------
Benchmark Run: Mon Feb 06 2017 22:48:04 - 22:57:10
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 60789432.6 lps (10.1 s, 2 samples)
Double-Precision Whetstone 7096.4 MWIPS (10.1 s, 2 samples)
Execl Throughput 291.2 lps (29.3 s, 1 samples)
File Copy 1024 bufsize 2000 maxblocks 18725.0 KBps (30.0 s, 1 samples)
File Copy 256 bufsize 500 maxblocks 5313.0 KBps (30.0 s, 1 samples)
File Copy 4096 bufsize 8000 maxblocks 79636.0 KBps (30.0 s, 1 samples)
Pipe Throughput 139019.4 lps (10.0 s, 2 samples)
Pipe-based Context Switching 37107.3 lps (10.0 s, 2 samples)
Process Creation 405.8 lps (30.1 s, 1 samples)
Shell Scripts (1 concurrent) 772.4 lpm (60.2 s, 1 samples)
Shell Scripts (8 concurrent) 110.3 lpm (60.9 s, 1 samples)
System Call Overhead 156715.5 lps (10.0 s, 2 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 60789432.6 5209.0
Double-Precision Whetstone 55.0 7096.4 1290.3
Execl Throughput 43.0 291.2 67.7
File Copy 1024 bufsize 2000 maxblocks 3960.0 18725.0 47.3
File Copy 256 bufsize 500 maxblocks 1655.0 5313.0 32.1
File Copy 4096 bufsize 8000 maxblocks 5800.0 79636.0 137.3
Pipe Throughput 12440.0 139019.4 111.8
Pipe-based Context Switching 4000.0 37107.3 92.8
Process Creation 126.0 405.8 32.2
Shell Scripts (1 concurrent) 42.4 772.4 182.2
Shell Scripts (8 concurrent) 6.0 110.3 183.8
System Call Overhead 15000.0 156715.5 104.5
========
System Benchmarks Index Score 148.0


Build 15002 (15002.rs_prerelease.170102-1700)

ここでカーネルバージョン表示が 3.4.0+ からしれっと 4.4.0-43-Microsoft になってますね。スコアは168.4。File Copyは少し回復したが、それ以外は遅くなった。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 4.4.0-43-Microsoft -- #1-Microsoft Wed Dec 31 14:42:53 PST 2014
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
22:27:01 up 18 min, 0 users, load average: 0.52, 0.58, 0.59; runlevel unknown

------------------------------------------------------------------------
Benchmark Run: Mon Jan 09 2017 22:55:05 - 23:23:53
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 51658132.1 lps (10.0 s, 7 samples)
Double-Precision Whetstone 7204.1 MWIPS (10.0 s, 7 samples)
Execl Throughput 426.4 lps (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 29189.6 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 7266.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 104499.9 KBps (30.0 s, 2 samples)
Pipe Throughput 102834.5 lps (10.0 s, 7 samples)
Pipe-based Context Switching 41627.9 lps (10.0 s, 7 samples)
Process Creation 509.9 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1000.7 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 143.9 lpm (61.1 s, 2 samples)
System Call Overhead 117740.4 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 51658132.1 4426.6
Double-Precision Whetstone 55.0 7204.1 1309.8
Execl Throughput 43.0 426.4 99.2
File Copy 1024 bufsize 2000 maxblocks 3960.0 29189.6 73.7
File Copy 256 bufsize 500 maxblocks 1655.0 7266.5 43.9
File Copy 4096 bufsize 8000 maxblocks 5800.0 104499.9 180.2
Pipe Throughput 12440.0 102834.5 82.7
Pipe-based Context Switching 4000.0 41627.9 104.1
Process Creation 126.0 509.9 40.5
Shell Scripts (1 concurrent) 42.4 1000.7 236.0
Shell Scripts (8 concurrent) 6.0 143.9 239.8
System Call Overhead 15000.0 117740.4 78.5
========
System Benchmarks Index Score 168.4


Build 14971 (14971.rs_prerelease.161111-1700)

178.7。DhrystoneとPipeとShell Scriptsが少し良くなったのでちょっとだけ戻したけど、まだ遅い、

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 3.4.0+ -- #1 PREEMPT Thu Aug 1 17:06:05 CST 2013
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
17:52:20 up 8:45, 0 users, load average: 0.52, 0.58, 0.59; runlevel unknown

------------------------------------------------------------------------
Benchmark Run: Fri Nov 18 2016 18:20:21 - 18:48:44
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 62447708.2 lps (10.0 s, 7 samples)
Double-Precision Whetstone 7128.1 MWIPS (10.1 s, 7 samples)
Execl Throughput 500.3 lps (29.6 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 22637.2 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 5696.4 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 80877.5 KBps (30.0 s, 2 samples)
Pipe Throughput 149728.2 lps (10.0 s, 7 samples)
Pipe-based Context Switching 39295.6 lps (10.0 s, 7 samples)
Process Creation 619.6 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1173.5 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 153.0 lpm (60.8 s, 2 samples)
System Call Overhead 174847.9 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 62447708.2 5351.1
Double-Precision Whetstone 55.0 7128.1 1296.0
Execl Throughput 43.0 500.3 116.4
File Copy 1024 bufsize 2000 maxblocks 3960.0 22637.2 57.2
File Copy 256 bufsize 500 maxblocks 1655.0 5696.4 34.4
File Copy 4096 bufsize 8000 maxblocks 5800.0 80877.5 139.4
Pipe Throughput 12440.0 149728.2 120.4
Pipe-based Context Switching 4000.0 39295.6 98.2
Process Creation 126.0 619.6 49.2
Shell Scripts (1 concurrent) 42.4 1173.5 276.8
Shell Scripts (8 concurrent) 6.0 153.0 254.9
System Call Overhead 15000.0 174847.9 116.6
========
System Benchmarks Index Score 178.7


Build 14965 (14965.rs_prerelease.161104-1700)

170.7。前回もそうでしたが今回もがっつり遅くなってます。Dhrystone 2、ファイルコピー、パイプ、システムコールが遅い。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 3.4.0+ -- #1 PREEMPT Thu Aug 1 17:06:05 CST 2013
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
06:32:10 up 6:30, 0 users, load average: 0.52, 0.58, 0.59; runlevel unknown

------------------------------------------------------------------------
Benchmark Run: Fri Nov 11 2016 07:00:12 - 07:28:41
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 58270259.8 lps (10.0 s, 7 samples)
Double-Precision Whetstone 7127.1 MWIPS (10.1 s, 7 samples)
Execl Throughput 476.0 lps (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 22720.7 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 5878.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 84084.0 KBps (30.0 s, 2 samples)
Pipe Throughput 135163.6 lps (10.0 s, 7 samples)
Pipe-based Context Switching 35269.4 lps (10.0 s, 7 samples)
Process Creation 580.2 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1088.8 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 142.1 lpm (60.8 s, 2 samples)
System Call Overhead 162108.0 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 58270259.8 4993.2
Double-Precision Whetstone 55.0 7127.1 1295.8
Execl Throughput 43.0 476.0 110.7
File Copy 1024 bufsize 2000 maxblocks 3960.0 22720.7 57.4
File Copy 256 bufsize 500 maxblocks 1655.0 5878.5 35.5
File Copy 4096 bufsize 8000 maxblocks 5800.0 84084.0 145.0
Pipe Throughput 12440.0 135163.6 108.7
Pipe-based Context Switching 4000.0 35269.4 88.2
Process Creation 126.0 580.2 46.0
Shell Scripts (1 concurrent) 42.4 1088.8 256.8
Shell Scripts (8 concurrent) 6.0 142.1 236.8
System Call Overhead 15000.0 162108.0 108.1
========
System Benchmarks Index Score 170.7


Build 14959 (14959.rs_prerelease.161026-1700)

213.8。またけっこう遅くなりました。ファイルコピー、パイプ、システムコールオーバーヘッドあたりでしょうか。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 3.4.0+ -- #1 PREEMPT Thu Aug 1 17:06:05 CST 2013
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
08:35:42 up 23 min, 0 users, load average: 0.52, 0.58, 0.59; runlevel

------------------------------------------------------------------------
Benchmark Run: Fri Nov 04 2016 09:03:41 - 09:31:57
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 62566382.2 lps (10.0 s, 7 samples)
Double-Precision Whetstone 7168.5 MWIPS (10.2 s, 7 samples)
Execl Throughput 516.7 lps (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 34040.5 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 9407.4 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 127684.1 KBps (30.0 s, 2 samples)
Pipe Throughput 147741.5 lps (10.0 s, 7 samples)
Pipe-based Context Switching 63509.2 lps (10.0 s, 7 samples)
Process Creation 670.3 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1311.9 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 170.1 lpm (60.5 s, 2 samples)
System Call Overhead 172098.4 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 62566382.2 5361.3
Double-Precision Whetstone 55.0 7168.5 1303.4
Execl Throughput 43.0 516.7 120.2
File Copy 1024 bufsize 2000 maxblocks 3960.0 34040.5 86.0
File Copy 256 bufsize 500 maxblocks 1655.0 9407.4 56.8
File Copy 4096 bufsize 8000 maxblocks 5800.0 127684.1 220.1
Pipe Throughput 12440.0 147741.5 118.8
Pipe-based Context Switching 4000.0 63509.2 158.8
Process Creation 126.0 670.3 53.2
Shell Scripts (1 concurrent) 42.4 1311.9 309.4
Shell Scripts (8 concurrent) 6.0 170.1 283.5
System Call Overhead 15000.0 172098.4 114.7
========
System Benchmarks Index Score 213.8


Build 14955 (14955.rs_prerelease.161020-1700)

4並列で246.0。少し戻しましたね。Double-Precision Whetstoneはがっつり下がってるけど、ファイルコピー、パイプ、シェルスクリプト、システムコールなどが前より良くなった。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 3.4.0+ -- #1 PREEMPT Thu Aug 1 17:06:05 CST 2013
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
16:39:10 up 3 min, 0 users, load average: 0.52, 0.58, 0.59; runlevel

------------------------------------------------------------------------
Benchmark Run: 水 10月 26 2016 17:07:09 - 17:35:29
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 61628732.0 lps (10.0 s, 7 samples)
Double-Precision Whetstone 7131.1 MWIPS (10.0 s, 7 samples)
Execl Throughput 167.3 lps (29.6 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 80173.0 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 20550.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 276870.1 KBps (30.0 s, 2 samples)
Pipe Throughput 208905.9 lps (10.0 s, 7 samples)
Pipe-based Context Switching 48255.3 lps (10.0 s, 7 samples)
Process Creation 589.3 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1353.0 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 178.7 lpm (60.9 s, 2 samples)
System Call Overhead 255308.3 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 61628732.0 5281.0
Double-Precision Whetstone 55.0 7131.1 1296.6
Execl Throughput 43.0 167.3 38.9
File Copy 1024 bufsize 2000 maxblocks 3960.0 80173.0 202.5
File Copy 256 bufsize 500 maxblocks 1655.0 20550.5 124.2
File Copy 4096 bufsize 8000 maxblocks 5800.0 276870.1 477.4
Pipe Throughput 12440.0 208905.9 167.9
Pipe-based Context Switching 4000.0 48255.3 120.6
Process Creation 126.0 589.3 46.8
Shell Scripts (1 concurrent) 42.4 1353.0 319.1
Shell Scripts (8 concurrent) 6.0 178.7 297.8
System Call Overhead 15000.0 255308.3 170.2
========
System Benchmarks Index Score 246.0


Build 14951

なぜかローカルに降ってこなかったので、ベンチマークはなし。


Build 14946 (14946.rs_prerelease.161007-1700)

4並列のスコアは220.2。どんどん下がる一方ですねえ。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 3.4.0+ -- #1 PREEMPT Thu Aug 1 17:06:05 CST 2013
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
23:56:42 up 0 min, 0 users, load average: 0.52, 0.58, 0.59; runlevel

------------------------------------------------------------------------
Benchmark Run: 日 10月 16 2016 00:25:03 - 00:53:44
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 65852667.8 lps (10.0 s, 7 samples)
Double-Precision Whetstone 11641.9 MWIPS (9.9 s, 7 samples)
Execl Throughput 167.4 lps (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 60928.0 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 15280.4 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 209971.3 KBps (30.0 s, 2 samples)
Pipe Throughput 147251.0 lps (10.0 s, 7 samples)
Pipe-based Context Switching 37652.3 lps (10.0 s, 7 samples)
Process Creation 603.7 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1313.4 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 172.0 lpm (60.4 s, 2 samples)
System Call Overhead 172385.7 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 65852667.8 5642.9
Double-Precision Whetstone 55.0 11641.9 2116.7
Execl Throughput 43.0 167.4 38.9
File Copy 1024 bufsize 2000 maxblocks 3960.0 60928.0 153.9
File Copy 256 bufsize 500 maxblocks 1655.0 15280.4 92.3
File Copy 4096 bufsize 8000 maxblocks 5800.0 209971.3 362.0
Pipe Throughput 12440.0 147251.0 118.4
Pipe-based Context Switching 4000.0 37652.3 94.1
Process Creation 126.0 603.7 47.9
Shell Scripts (1 concurrent) 42.4 1313.4 309.8
Shell Scripts (8 concurrent) 6.0 172.0 286.6
System Call Overhead 15000.0 172385.7 114.9
========
System Benchmarks Index Score 220.2


Build 14942 (14942.rs_prerelease.161003-1929)

WSLを削除して再セットアップして試した。テストも落ちなくなったしスコアも少し上がった。ぐぬぬ。これまでのテストスコアの信頼度が低いことがわかった。

4並列のスコアは228.6。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 3.4.0+ -- #1 PREEMPT Thu Aug 1 17:06:05 CST 2013
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
07:24:40 up 1:37, 0 users, load average: 0.52, 0.58, 0.59; runlevel

------------------------------------------------------------------------
Benchmark Run: 火 10月 11 2016 07:52:55 - 08:21:33
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 65663926.0 lps (10.0 s, 7 samples)
Double-Precision Whetstone 11623.5 MWIPS (9.9 s, 7 samples)
Execl Throughput 165.8 lps (29.8 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 57023.3 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 14611.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 194312.1 KBps (30.0 s, 2 samples)
Pipe Throughput 149040.2 lps (10.0 s, 7 samples)
Pipe-based Context Switching 63197.9 lps (10.0 s, 7 samples)
Process Creation 600.5 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1407.7 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 183.1 lpm (60.8 s, 2 samples)
System Call Overhead 171390.6 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 65663926.0 5626.7
Double-Precision Whetstone 55.0 11623.5 2113.4
Execl Throughput 43.0 165.8 38.6
File Copy 1024 bufsize 2000 maxblocks 3960.0 57023.3 144.0
File Copy 256 bufsize 500 maxblocks 1655.0 14611.5 88.3
File Copy 4096 bufsize 8000 maxblocks 5800.0 194312.1 335.0
Pipe Throughput 12440.0 149040.2 119.8
Pipe-based Context Switching 4000.0 63197.9 158.0
Process Creation 126.0 600.5 47.7
Shell Scripts (1 concurrent) 42.4 1407.7 332.0
Shell Scripts (8 concurrent) 6.0 183.1 305.1
System Call Overhead 15000.0 171390.6 114.3
========
System Benchmarks Index Score 228.6

以下、再セットアップ前

繰り返し回数を5回にして (-i 5オプションを与えて) 実行。

4並列のスコアは220.5。ひたすら数字が下がる一方なのはPCの問題もあるのかもしれない。そろそろ一度WSLを削除して再セットアップして試したほうがいいのかも。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 3.4.0+ -- #1 PREEMPT Thu Aug 1 17:06:05 CST 2013
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
22:03:56 up 3:49, 0 users, load average: 0.52, 0.58, 0.59; runlevel unknown

------------------------------------------------------------------------
Benchmark Run: Sat Oct 08 2016 22:20:44 - 22:37:42
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 63160479.1 lps (10.0 s, 4 samples)
Double-Precision Whetstone 6339.7 MWIPS (10.1 s, 4 samples)
Execl Throughput 168.2 lps (29.8 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 58421.0 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 14822.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 203868.0 KBps (30.0 s, 2 samples)
Pipe Throughput 150761.4 lps (10.0 s, 4 samples)
Pipe-based Context Switching 64067.4 lps (10.0 s, 4 samples)
Process Creation 622.6 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1430.7 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 187.4 lpm (60.4 s, 2 samples)
System Call Overhead 173241.2 lps (10.0 s, 4 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 63160479.1 5412.2
Double-Precision Whetstone 55.0 6339.7 1152.7
Execl Throughput 43.0 168.2 39.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 58421.0 147.5
File Copy 256 bufsize 500 maxblocks 1655.0 14822.5 89.6
File Copy 4096 bufsize 8000 maxblocks 5800.0 203868.0 351.5
Pipe Throughput 12440.0 150761.4 121.2
Pipe-based Context Switching 4000.0 64067.4 160.2
Process Creation 126.0 622.6 49.4
Shell Scripts (1 concurrent) 42.4 1430.7 337.4
Shell Scripts (8 concurrent) 6.0 187.4 312.3
System Call Overhead 15000.0 173241.2 115.5
========
System Benchmarks Index Score 220.5


Build 14936 (14936.rs_prerelease.160923-1700)

このビルドから正式にXenialに対応したとのことで、do-release-upgradeを実行してバージョンを上げました。

でもやはりベンチ途中で落ちたので、繰り返し回数を4回にして (-i 4オプションを与えて) 実行してます。

4並列のスコアは222.5。倍精度浮動小数点演算が大きく下がり、整数・文字列演算も下がった影響でしょうか。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 3.4.0+ -- #1 PREEMPT Thu Aug 1 17:06:05 CST 2013
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
21:13:14 up 28 min, 0 users, load average: 0.52, 0.58, 0.59; runlevel unknown

------------------------------------------------------------------------
Benchmark Run: Thu Sep 29 2016 21:23:19 - 21:33:30
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 59889363.9 lps (10.0 s, 3 samples)
Double-Precision Whetstone 6356.9 MWIPS (10.0 s, 3 samples)
Execl Throughput 165.6 lps (29.8 s, 1 samples)
File Copy 1024 bufsize 2000 maxblocks 60651.0 KBps (30.0 s, 1 samples)
File Copy 256 bufsize 500 maxblocks 15124.0 KBps (30.0 s, 1 samples)
File Copy 4096 bufsize 8000 maxblocks 201167.0 KBps (30.0 s, 1 samples)
Pipe Throughput 152302.2 lps (10.0 s, 3 samples)
Pipe-based Context Switching 65603.3 lps (10.0 s, 3 samples)
Process Creation 651.0 lps (30.0 s, 1 samples)
Shell Scripts (1 concurrent) 1473.6 lpm (60.1 s, 1 samples)
Shell Scripts (8 concurrent) 190.2 lpm (60.6 s, 1 samples)
System Call Overhead 174132.5 lps (10.0 s, 3 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 59889363.9 5131.9
Double-Precision Whetstone 55.0 6356.9 1155.8
Execl Throughput 43.0 165.6 38.5
File Copy 1024 bufsize 2000 maxblocks 3960.0 60651.0 153.2
File Copy 256 bufsize 500 maxblocks 1655.0 15124.0 91.4
File Copy 4096 bufsize 8000 maxblocks 5800.0 201167.0 346.8
Pipe Throughput 12440.0 152302.2 122.4
Pipe-based Context Switching 4000.0 65603.3 164.0
Process Creation 126.0 651.0 51.7
Shell Scripts (1 concurrent) 42.4 1473.6 347.5
Shell Scripts (8 concurrent) 6.0 190.2 317.0
System Call Overhead 15000.0 174132.5 116.1
========
System Benchmarks Index Score 222.5


Build 14931 (14931.rs_prerelease.160916-1700)

Build 14926と同じくエラーで落ちたので、繰り返し回数を5回にして (-i 5オプションを与えて)実行しました。

それが影響したのかしないのか、4並列のスコアは235.1。ファイルコピーが遅くなってスコアの足を引っ張った感じ。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 3.4.0+ -- #1 PREEMPT Thu Aug 1 17:06:05 CST 2013
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
00:14:43 up 26 min, 0 users, load average: 0.52, 0.58, 0.59; runlevel

------------------------------------------------------------------------
Benchmark Run: Sat Sep 24 2016 00:31:38 - 00:48:44
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 63640830.7 lps (10.0 s, 4 samples)
Double-Precision Whetstone 11388.6 MWIPS (9.9 s, 4 samples)
Execl Throughput 167.5 lps (29.5 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 59803.2 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 15152.0 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 203266.4 KBps (30.0 s, 2 samples)
Pipe Throughput 154718.1 lps (10.0 s, 4 samples)
Pipe-based Context Switching 66701.7 lps (10.0 s, 4 samples)
Process Creation 637.9 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1455.7 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 190.4 lpm (60.5 s, 2 samples)
System Call Overhead 175181.2 lps (10.1 s, 4 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 63640830.7 5453.4
Double-Precision Whetstone 55.0 11388.6 2070.7
Execl Throughput 43.0 167.5 38.9
File Copy 1024 bufsize 2000 maxblocks 3960.0 59803.2 151.0
File Copy 256 bufsize 500 maxblocks 1655.0 15152.0 91.6
File Copy 4096 bufsize 8000 maxblocks 5800.0 203266.4 350.5
Pipe Throughput 12440.0 154718.1 124.4
Pipe-based Context Switching 4000.0 66701.7 166.8
Process Creation 126.0 637.9 50.6
Shell Scripts (1 concurrent) 42.4 1455.7 343.3
Shell Scripts (8 concurrent) 6.0 190.4 317.3
System Call Overhead 15000.0 175181.2 116.8
========
System Benchmarks Index Score 235.1


Build 14926 (14926.rs_prerelease.160910-1529)

エラーで落ちました……うまく動いたらまた更新します。

Run: "Shell Scripts (1 concurrent)": Bad wait status: 0x8000; aborting


Build 14915 (14915.rs_prerelease.160826-1902)

4並列で259.9。Process Creationが54.1に上昇し、他も全体的に向上しているが、System Call Overheadだけは結構スコアが下がった(オーバーヘッドが大きくなった)。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 3.4.0+ -- #1 PREEMPT Thu Aug 1 17:06:05 CST 2013
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
07:37:17 up 11 min, 0 users, load average: 0.52, 0.58, 0.59; runlevel

------------------------------------------------------------------------
Benchmark Run: Thu Sep 01 2016 08:05:34 - 08:34:08
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 63203092.1 lps (10.0 s, 7 samples)
Double-Precision Whetstone 11486.0 MWIPS (9.8 s, 7 samples)
Execl Throughput 166.7 lps (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 89790.4 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 22899.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 279349.6 KBps (30.0 s, 2 samples)
Pipe Throughput 151258.8 lps (10.0 s, 7 samples)
Pipe-based Context Switching 66193.6 lps (10.0 s, 7 samples)
Process Creation 681.1 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1505.9 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 193.4 lpm (60.3 s, 2 samples)
System Call Overhead 172726.0 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 63203092.1 5415.9
Double-Precision Whetstone 55.0 11486.0 2088.4
Execl Throughput 43.0 166.7 38.8
File Copy 1024 bufsize 2000 maxblocks 3960.0 89790.4 226.7
File Copy 256 bufsize 500 maxblocks 1655.0 22899.5 138.4
File Copy 4096 bufsize 8000 maxblocks 5800.0 279349.6 481.6
Pipe Throughput 12440.0 151258.8 121.6
Pipe-based Context Switching 4000.0 66193.6 165.5
Process Creation 126.0 681.1 54.1
Shell Scripts (1 concurrent) 42.4 1505.9 355.2
Shell Scripts (8 concurrent) 6.0 193.4 322.4
System Call Overhead 15000.0 172726.0 115.2
========
System Benchmarks Index Score 259.9


Build 14905 (14905.rs_prerelease.160811-1739)

4並列で256.9。もちろんfujiedaさんのベンチマーク結果とは比べられないが、後述するCygwinと比べるとファイルコピー、パイプのスループット、システムコールのオーバーヘッドが遅い。プロセス生成とシェルスクリプトは速かった。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: GNU/Linux
OS: GNU/Linux -- 3.4.0+ -- #1 PREEMPT Thu Aug 1 17:06:05 CST 2013
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (4602.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
21:32:07 up 12 min, 0 users, load average: 0.52, 0.58, 0.59; runlevel

------------------------------------------------------------------------
Benchmark Run: Wed Aug 24 2016 22:00:23 - 22:29:01
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 63039929.9 lps (10.0 s, 7 samples)
Double-Precision Whetstone 11436.5 MWIPS (9.8 s, 7 samples)
Execl Throughput 164.7 lps (29.6 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 85246.5 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 21796.3 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 269285.4 KBps (30.0 s, 2 samples)
Pipe Throughput 149504.0 lps (10.0 s, 7 samples)
Pipe-based Context Switching 65073.2 lps (10.0 s, 7 samples)
Process Creation 471.8 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1336.4 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 181.8 lpm (60.6 s, 2 samples)
System Call Overhead 312309.5 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 63039929.9 5401.9
Double-Precision Whetstone 55.0 11436.5 2079.4
Execl Throughput 43.0 164.7 38.3
File Copy 1024 bufsize 2000 maxblocks 3960.0 85246.5 215.3
File Copy 256 bufsize 500 maxblocks 1655.0 21796.3 131.7
File Copy 4096 bufsize 8000 maxblocks 5800.0 269285.4 464.3
Pipe Throughput 12440.0 149504.0 120.2
Pipe-based Context Switching 4000.0 65073.2 162.7
Process Creation 126.0 471.8 37.4
Shell Scripts (1 concurrent) 42.4 1336.4 315.2
Shell Scripts (8 concurrent) 6.0 181.8 303.0
System Call Overhead 15000.0 312309.5 208.2
========
System Benchmarks Index Score 256.9


Cygwin

4並列で332.8。ファイルシステムはデフォルトのまま(noaclはつけていない)。

Bash on Ubuntu on WindowsとCygwinとは用途(というか目的)が違うので、Cygwinより遅いから使わないとか、Cygwinより速いから乗り換える、みたいなことはお勧めしない。

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: ASPIRE3830T: Cygwin
OS: Cygwin -- 2.5.2(0.297/5/3) -- 2016-06-23 14:29
Machine: x86_64 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (0.0 bogomips)
Hyper-Threading, MMX, Physical Address Ext, SYSENTER/SYSEXIT
CPU 1: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (0.0 bogomips)
Hyper-Threading, MMX, Physical Address Ext, SYSENTER/SYSEXIT
CPU 2: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (0.0 bogomips)
Hyper-Threading, MMX, Physical Address Ext, SYSENTER/SYSEXIT
CPU 3: Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz (0.0 bogomips)
Hyper-Threading, MMX, Physical Address Ext, SYSENTER/SYSEXIT
; runlevel

------------------------------------------------------------------------
Benchmark Run: 火 8月 23 2016 22:59:21 - 23:28:28
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 64400294.4 lps (10.1 s, 7 samples)
Double-Precision Whetstone 14290.8 MWIPS (9.8 s, 7 samples)
Execl Throughput 184.5 lps (29.4 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 338765.0 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 87103.9 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 728546.8 KBps (30.0 s, 2 samples)
Pipe Throughput 572247.6 lps (10.1 s, 7 samples)
Pipe-based Context Switching 41226.9 lps (10.5 s, 7 samples)
Process Creation 203.5 lps (30.1 s, 2 samples)
Shell Scripts (1 concurrent) 371.3 lpm (60.3 s, 2 samples)
Shell Scripts (8 concurrent) 48.0 lpm (61.9 s, 2 samples)
System Call Overhead 1483903.8 lps (10.1 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 64400294.4 5518.4
Double-Precision Whetstone 55.0 14290.8 2598.3
Execl Throughput 43.0 184.5 42.9
File Copy 1024 bufsize 2000 maxblocks 3960.0 338765.0 855.5
File Copy 256 bufsize 500 maxblocks 1655.0 87103.9 526.3
File Copy 4096 bufsize 8000 maxblocks 5800.0 728546.8 1256.1
Pipe Throughput 12440.0 572247.6 460.0
Pipe-based Context Switching 4000.0 41226.9 103.1
Process Creation 126.0 203.5 16.1
Shell Scripts (1 concurrent) 42.4 371.3 87.6
Shell Scripts (8 concurrent) 6.0 48.0 79.9
System Call Overhead 15000.0 1483903.8 989.3
========
System Benchmarks Index Score 332.8