KVMのゲストマシンとさくらのVPSでUnixBench採ってみました
みなさん、こんにちは(^^)/
楽しかったゴールデンウィークも終わってしましたが、いかがお過ごしでしょうか？
私はというと、行楽日和の晴天を横目に、薄暗い事務所の中でコンピューターに向かって黙々と仕事をこなす毎日ですw
さて、仕事でUnixBenchを採る機会がありましたので、個人的な備忘録の意味も含め、ここに記録して(晒して)おきたいと思います。
UnixBenchは以下のページから取ってきます。
https://code.google.com/p/byte-unixbench/
バージョン5.1から新しく用意されたGRAPHICテストを実行したい場合には、Makefileを編集し"GRAPHIC_TESTS = defined"の行を有効にしてmakeする必要がありますが、今回は不要です。
CPUが二つ以上のシステムの場合、テストは二回実行されます。一回目は一セットのテストが同時実行され、二回目はCPUの数をNとすると、Nセットのテストが並列に同時実行されます。
テストの実行時間は一セットおよそ29分で、マルチCPUマシン上で、GRAPHICテストを行わない場合、およそ、58分程度かかるということです。
今回UnixBenchを実行したのは以下のマシンです。
(1)さくらのVPS512@980円
(2)内製KVMゲストマシン1 (ホストCPU Xeon X5650 @ 2.67GHz)
(3)内製KVMゲストマシン2 (ホストCPU Xeon X3440 @ 2.53GHz)
(4)(3)のホストマシン
(1)さくらのVPS512の結果
(2)内製KVMゲストマシン1 (ホストCPU Xeon X5650 @ 2.67GHz)
以上の結果を見ると、CPU8個のサーバより、2CPUの仮想マシンゲストの方が最終スコアが高かったり、FileCopy関連のインデックスが仮想マシンの方が軒並み高かったり、理解に苦しむ結果が出ています。
今後、別のベンチマークツールで試してみたいと思います。
追記 System Benchmarks Index Scoreではなく、個別のテストごとに比較すべきか。
楽しかったゴールデンウィークも終わってしましたが、いかがお過ごしでしょうか？
私はというと、行楽日和の晴天を横目に、薄暗い事務所の中でコンピューターに向かって黙々と仕事をこなす毎日ですw
さて、仕事でUnixBenchを採る機会がありましたので、個人的な備忘録の意味も含め、ここに記録して(晒して)おきたいと思います。
UnixBenchは以下のページから取ってきます。
https://code.google.com/p/byte-unixbench/
wget http://byte-unixbench.googlecode.com/files/UnixBench5.1.3.tgzUSAGEをざっと眺めると、makeして./Runスクリプトを実行すれば用意されたすべてのテストが実行されるらしいです。
tar xvf UnixBench5.1.3.tgz
cd UnixBench
バージョン5.1から新しく用意されたGRAPHICテストを実行したい場合には、Makefileを編集し"GRAPHIC_TESTS = defined"の行を有効にしてmakeする必要がありますが、今回は不要です。
CPUが二つ以上のシステムの場合、テストは二回実行されます。一回目は一セットのテストが同時実行され、二回目はCPUの数をNとすると、Nセットのテストが並列に同時実行されます。
テストの実行時間は一セットおよそ29分で、マルチCPUマシン上で、GRAPHICテストを行わない場合、およそ、58分程度かかるということです。
今回UnixBenchを実行したのは以下のマシンです。
(1)さくらのVPS512@980円
CPU Intel(R) Core(TM)2 Duo CPU T7700 @ 2.40GHz x 2
メモリー 1GByte
OS Debian 6.0.7 Kernel 2.6.32 SMP x86_64
(2)内製KVMゲストマシン1 (ホストCPU Xeon X5650 @ 2.67GHz)
CPU QEMU Virtual CPU version 1.0 2.67GHz x 2
メモリー 512MByte
OS Debian 7.0 Kernel 3.9.1 SMP x86_64
(3)内製KVMゲストマシン2 (ホストCPU Xeon X3440 @ 2.53GHz)
CPU QEMU Virtual CPU version 1.0 2.52GHz x 2
メモリー 512MByte
OS Debian 7.0 Kernel 3.8.8 SMP x86_64
(4)(3)のホストマシン
CPU Intel(R) Xeon(R) CPU X3440 @ 2.53GHz 4core 8thread
メモリー 4GByte
OS Debian 6.0.7 Kernel 3.4.0 SMP x86_64
(1)さくらのVPS512の結果
BYTE UNIX Benchmarks (Version 5.1.3)
System: sak1: GNU/Linux
OS: GNU/Linux -- 2.6.32-5-amd64 -- #1 SMP Mon Jan 9 20:49:59 UTC 2012
Machine: x86_64 (unknown)
Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
CPU 0: Intel(R) Core(TM)2 Duo CPU T7700 @ 2.40GHz (5320.2 bogomips)
x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM)2 Duo CPU T7700 @ 2.40GHz (5320.2 bogomips)
x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
13:48:04 up 146 days, 11:37, 1 user, load average: 0.22, 0.12, 0.04; runlevel
------------------------------------------------------------------------
Benchmark Run: Wed May 08 2013 13:48:04 - 14:16:18
2 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 24085942.4 lps (10.0 s, 7 samples)
Double-Precision Whetstone 3098.5 MWIPS (9.9 s, 7 samples)
Execl Throughput 2249.6 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 671875.9 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 196527.3 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1413265.2 KBps (30.0 s, 2 samples)
Pipe Throughput 1897140.7 lps (10.0 s, 7 samples)
Pipe-based Context Switching 287952.8 lps (10.0 s, 7 samples)
Process Creation 5839.1 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 5899.8 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1365.4 lpm (60.0 s, 2 samples)
System Call Overhead 3301152.6 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 24085942.4 2063.9
Double-Precision Whetstone 55.0 3098.5 563.4
Execl Throughput 43.0 2249.6 523.2
File Copy 1024 bufsize 2000 maxblocks 3960.0 671875.9 1696.7
File Copy 256 bufsize 500 maxblocks 1655.0 196527.3 1187.5
File Copy 4096 bufsize 8000 maxblocks 5800.0 1413265.2 2436.7
Pipe Throughput 12440.0 1897140.7 1525.0
Pipe-based Context Switching 4000.0 287952.8 719.9
Process Creation 126.0 5839.1 463.4
Shell Scripts (1 concurrent) 42.4 5899.8 1391.5
Shell Scripts (8 concurrent) 6.0 1365.4 2275.7
System Call Overhead 15000.0 3301152.6 2200.8
========
System Benchmarks Index Score 1217.3
------------------------------------------------------------------------
Benchmark Run: Wed May 08 2013 14:16:18 - 14:44:33
2 CPUs in system; running 2 parallel copies of tests
Dhrystone 2 using register variables 48543864.7 lps (10.0 s, 7 samples)
Double-Precision Whetstone 6093.3 MWIPS (10.0 s, 7 samples)
Execl Throughput 6641.9 lps (29.6 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 775102.5 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 214100.7 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 2170934.3 KBps (30.0 s, 2 samples)
Pipe Throughput 3701877.8 lps (10.0 s, 7 samples)
Pipe-based Context Switching 603992.9 lps (10.0 s, 7 samples)
Process Creation 21862.8 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 10135.8 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1410.2 lpm (60.0 s, 2 samples)
System Call Overhead 5432476.9 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 48543864.7 4159.7
Double-Precision Whetstone 55.0 6093.3 1107.9
Execl Throughput 43.0 6641.9 1544.6
File Copy 1024 bufsize 2000 maxblocks 3960.0 775102.5 1957.3
File Copy 256 bufsize 500 maxblocks 1655.0 214100.7 1293.7
File Copy 4096 bufsize 8000 maxblocks 5800.0 2170934.3 3743.0
Pipe Throughput 12440.0 3701877.8 2975.8
Pipe-based Context Switching 4000.0 603992.9 1510.0
Process Creation 126.0 21862.8 1735.1
Shell Scripts (1 concurrent) 42.4 10135.8 2390.5
Shell Scripts (8 concurrent) 6.0 1410.2 2350.3
System Call Overhead 15000.0 5432476.9 3621.7
========
System Benchmarks Index Score 2166.7
(2)内製KVMゲストマシン1 (ホストCPU Xeon X5650 @ 2.67GHz)
BYTE UNIX Benchmarks (Version 5.1.3)
System: v147: GNU/Linux
OS: GNU/Linux -- 3.9.1-64kvmg01 -- #1 SMP Wed May 8 22:00:45 JST 2013
Machine: x86_64 (unknown)
Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
CPU 0: QEMU Virtual CPU version 1.0 (5333.3 bogomips)
x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: QEMU Virtual CPU version 1.0 (5333.3 bogomips)
x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
15:34:32 up 0 min, 1 user, load average: 0.16, 0.03, 0.01; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Thu May 09 2013 15:34:32 - 16:02:31
2 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 30879094.2 lps (10.0 s, 7 samples)
Double-Precision Whetstone 3749.0 MWIPS (9.1 s, 7 samples)
Execl Throughput 4534.2 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 1097888.6 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 301066.2 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 2055234.3 KBps (30.0 s, 2 samples)
Pipe Throughput 2216106.5 lps (10.0 s, 7 samples)
Pipe-based Context Switching 381391.5 lps (10.0 s, 7 samples)
Process Creation 5417.9 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 9541.0 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 2040.2 lpm (60.0 s, 2 samples)
System Call Overhead 4035159.8 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 30879094.2 2646.0
Double-Precision Whetstone 55.0 3749.0 681.6
Execl Throughput 43.0 4534.2 1054.5
File Copy 1024 bufsize 2000 maxblocks 3960.0 1097888.6 2772.4
File Copy 256 bufsize 500 maxblocks 1655.0 301066.2 1819.1
File Copy 4096 bufsize 8000 maxblocks 5800.0 2055234.3 3543.5
Pipe Throughput 12440.0 2216106.5 1781.4
Pipe-based Context Switching 4000.0 381391.5 953.5
Process Creation 126.0 5417.9 430.0
Shell Scripts (1 concurrent) 42.4 9541.0 2250.2
Shell Scripts (8 concurrent) 6.0 2040.2 3400.3
System Call Overhead 15000.0 4035159.8 2690.1
========
System Benchmarks Index Score 1681.5
------------------------------------------------------------------------
Benchmark Run: Thu May 09 2013 16:02:31 - 16:30:31
2 CPUs in system; running 2 parallel copies of tests
Dhrystone 2 using register variables 61886737.4 lps (10.0 s, 7 samples)
Double-Precision Whetstone 7438.8 MWIPS (9.1 s, 7 samples)
Execl Throughput 8960.8 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 1356632.6 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 228901.8 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 3404779.9 KBps (30.0 s, 2 samples)
Pipe Throughput 4411432.3 lps (10.0 s, 7 samples)
Pipe-based Context Switching 792350.4 lps (10.0 s, 7 samples)
Process Creation 29506.4 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 16737.4 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 2339.6 lpm (60.0 s, 2 samples)
System Call Overhead 6544409.1 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 61886737.4 5303.1
Double-Precision Whetstone 55.0 7438.8 1352.5
Execl Throughput 43.0 8960.8 2083.9
File Copy 1024 bufsize 2000 maxblocks 3960.0 1356632.6 3425.8
File Copy 256 bufsize 500 maxblocks 1655.0 228901.8 1383.1
File Copy 4096 bufsize 8000 maxblocks 5800.0 3404779.9 5870.3
Pipe Throughput 12440.0 4411432.3 3546.2
Pipe-based Context Switching 4000.0 792350.4 1980.9
Process Creation 126.0 29506.4 2341.8
Shell Scripts (1 concurrent) 42.4 16737.4 3947.5
Shell Scripts (8 concurrent) 6.0 2339.6 3899.4
System Call Overhead 15000.0 6544409.1 4362.9
========
System Benchmarks Index Score 2963.7
(3)内製KVMゲストマシン2 (ホストCPU Xeon X3440 @ 2.53GHz)
BYTE UNIX Benchmarks (Version 5.1.3)
System: v153: GNU/Linux
OS: GNU/Linux -- 3.8.8-64kvmg01 -- #1 SMP Mon Apr 22 23:44:25 JST 2013
Machine: x86_64 (unknown)
Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
CPU 0: QEMU Virtual CPU version 1.4.1 (5066.6 bogomips)
x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: QEMU Virtual CPU version 1.4.1 (5066.6 bogomips)
x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
04:21:48 up 2 min, 2 users, load average: 0.06, 0.02, 0.01; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Fri May 10 2013 04:21:48 - 04:49:56
2 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 28700468.1 lps (10.0 s, 7 samples)
Double-Precision Whetstone 3603.2 MWIPS (10.0 s, 7 samples)
Execl Throughput 4820.8 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 1136729.9 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 313902.4 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 2329713.1 KBps (30.0 s, 2 samples)
Pipe Throughput 1937085.6 lps (10.0 s, 7 samples)
Pipe-based Context Switching 416442.2 lps (10.0 s, 7 samples)
Process Creation 4905.0 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 5206.7 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 2342.7 lpm (60.0 s, 2 samples)
System Call Overhead 3696866.5 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 28700468.1 2459.3
Double-Precision Whetstone 55.0 3603.2 655.1
Execl Throughput 43.0 4820.8 1121.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 1136729.9 2870.5
File Copy 256 bufsize 500 maxblocks 1655.0 313902.4 1896.7
File Copy 4096 bufsize 8000 maxblocks 5800.0 2329713.1 4016.7
Pipe Throughput 12440.0 1937085.6 1557.1
Pipe-based Context Switching 4000.0 416442.2 1041.1
Process Creation 126.0 4905.0 389.3
Shell Scripts (1 concurrent) 42.4 5206.7 1228.0
Shell Scripts (8 concurrent) 6.0 2342.7 3904.5
System Call Overhead 15000.0 3696866.5 2464.6
========
System Benchmarks Index Score 1606.1
------------------------------------------------------------------------
Benchmark Run: Fri May 10 2013 04:49:56 - 05:18:04
2 CPUs in system; running 2 parallel copies of tests
Dhrystone 2 using register variables 54903785.4 lps (10.0 s, 7 samples)
Double-Precision Whetstone 6894.6 MWIPS (10.0 s, 7 samples)
Execl Throughput 9306.8 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 1347898.4 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 354944.7 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 3256880.8 KBps (30.0 s, 2 samples)
Pipe Throughput 3680672.6 lps (10.0 s, 7 samples)
Pipe-based Context Switching 740043.1 lps (10.0 s, 7 samples)
Process Creation 34071.7 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 17647.1 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 2415.1 lpm (60.0 s, 2 samples)
System Call Overhead 5929544.7 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 54903785.4 4704.7
Double-Precision Whetstone 55.0 6894.6 1253.6
Execl Throughput 43.0 9306.8 2164.4
File Copy 1024 bufsize 2000 maxblocks 3960.0 1347898.4 3403.8
File Copy 256 bufsize 500 maxblocks 1655.0 354944.7 2144.7
File Copy 4096 bufsize 8000 maxblocks 5800.0 3256880.8 5615.3
Pipe Throughput 12440.0 3680672.6 2958.7
Pipe-based Context Switching 4000.0 740043.1 1850.1
Process Creation 126.0 34071.7 2704.1
Shell Scripts (1 concurrent) 42.4 17647.1 4162.0
Shell Scripts (8 concurrent) 6.0 2415.1 4025.1
System Call Overhead 15000.0 5929544.7 3953.0
========
System Benchmarks Index Score 2991.1
(4)(3)のホストマシン
BYTE UNIX Benchmarks (Version 5.1.3)
System: kvm: GNU/Linux
OS: GNU/Linux -- 3.4.0-64kvmh01 -- #1 SMP Fri May 25 14:59:18 JST 2012
Machine: x86_64 (unknown)
Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
CPU 0: Intel(R) Xeon(R) CPU X3440 @ 2.53GHz (5066.2 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 1: Intel(R) Xeon(R) CPU X3440 @ 2.53GHz (5066.2 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 2: Intel(R) Xeon(R) CPU X3440 @ 2.53GHz (5066.2 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 3: Intel(R) Xeon(R) CPU X3440 @ 2.53GHz (5066.2 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 4: Intel(R) Xeon(R) CPU X3440 @ 2.53GHz (5066.2 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 5: Intel(R) Xeon(R) CPU X3440 @ 2.53GHz (5066.2 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 6: Intel(R) Xeon(R) CPU X3440 @ 2.53GHz (5066.2 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 7: Intel(R) Xeon(R) CPU X3440 @ 2.53GHz (5066.2 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
20:37:04 up 113 days, 7:51, 2 users, load average: 0.50, 0.22, 0.52; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Thu May 09 2013 20:37:04 - 21:05:17
8 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 28301315.7 lps (10.0 s, 7 samples)
Double-Precision Whetstone 3470.1 MWIPS (10.0 s, 7 samples)
Execl Throughput 1569.0 lps (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 841958.4 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 235274.3 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1938405.5 KBps (30.0 s, 2 samples)
Pipe Throughput 1827474.3 lps (10.0 s, 7 samples)
Pipe-based Context Switching 54481.2 lps (10.0 s, 7 samples)
Process Creation 3844.0 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 3859.0 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 2910.4 lpm (60.0 s, 2 samples)
System Call Overhead 3240538.1 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 28301315.7 2425.1
Double-Precision Whetstone 55.0 3470.1 630.9
Execl Throughput 43.0 1569.0 364.9
File Copy 1024 bufsize 2000 maxblocks 3960.0 841958.4 2126.2
File Copy 256 bufsize 500 maxblocks 1655.0 235274.3 1421.6
File Copy 4096 bufsize 8000 maxblocks 5800.0 1938405.5 3342.1
Pipe Throughput 12440.0 1827474.3 1469.0
Pipe-based Context Switching 4000.0 54481.2 136.2
Process Creation 126.0 3844.0 305.1
Shell Scripts (1 concurrent) 42.4 3859.0 910.1
Shell Scripts (8 concurrent) 6.0 2910.4 4850.6
System Call Overhead 15000.0 3240538.1 2160.4
========
System Benchmarks Index Score 1104.1
------------------------------------------------------------------------
Benchmark Run: Thu May 09 2013 21:05:17 - 21:33:41
8 CPUs in system; running 8 parallel copies of tests
Dhrystone 2 using register variables 110467624.7 lps (10.0 s, 7 samples)
Double-Precision Whetstone 21211.4 MWIPS (9.9 s, 7 samples)
Execl Throughput 12952.9 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 495833.5 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 134897.1 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1547822.4 KBps (30.0 s, 2 samples)
Pipe Throughput 2276741.3 lps (10.0 s, 7 samples)
Pipe-based Context Switching 455844.3 lps (10.0 s, 7 samples)
Process Creation 30000.0 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 28230.5 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 3724.5 lpm (60.0 s, 2 samples)
System Call Overhead 3454619.9 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 110467624.7 9465.9
Double-Precision Whetstone 55.0 21211.4 3856.6
Execl Throughput 43.0 12952.9 3012.3
File Copy 1024 bufsize 2000 maxblocks 3960.0 495833.5 1252.1
File Copy 256 bufsize 500 maxblocks 1655.0 134897.1 815.1
File Copy 4096 bufsize 8000 maxblocks 5800.0 1547822.4 2668.7
Pipe Throughput 12440.0 2276741.3 1830.2
Pipe-based Context Switching 4000.0 455844.3 1139.6
Process Creation 126.0 30000.0 2381.0
Shell Scripts (1 concurrent) 42.4 28230.5 6658.1
Shell Scripts (8 concurrent) 6.0 3724.5 6207.5
System Call Overhead 15000.0 3454619.9 2303.1
========
System Benchmarks Index Score 2686.9
以上の結果を見ると、CPU8個のサーバより、2CPUの仮想マシンゲストの方が最終スコアが高かったり、FileCopy関連のインデックスが仮想マシンの方が軒並み高かったり、理解に苦しむ結果が出ています。
今後、別のベンチマークツールで試してみたいと思います。
追記 System Benchmarks Index Scoreではなく、個別のテストごとに比較すべきか。
Comments