Amazon

2013年8月31日土曜日

「ベンチマーク]UNIX Benchmarks



VurtialBox上で稼働するLinuxが増えてきたのでベンチマークすることにしました。


手始めにホストOSのUbuntu DesktopとゲストOSのUbuntu Serverです。


先月初めに実行したのですが、結果の読み方が分かってから書こうと思っているうちに一ヶ月が過ぎてしまいました。


忘れてしまう前に、とりあえず結果だけアップしておきます。


CPUがCore2Duoなので、マルチコアに対応していないVirtualBox上ではどの程度性能が落ちるか比較して見たいところです。




CPU



$ cat /proc/cpuinfo
model name : Intel(R) Pentium(R) D CPU 3.20GHz
stepping : 2
cpu MHz : 3200.186
cache size : 2048 KB

ホストOS


Ubuntu Desktopです。



$ cat /etc/lsb-release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=8.04
DISTRIB_CODENAME=hardy
DISTRIB_DESCRIPTION="Ubuntu 8.04"
$ cat report

BYTE UNIX Benchmarks (Version 4.0.1)
System -- Linux ubuntu10 2.6.24-18-generic #1 SMP Wed May 28 20:27:26 UTC 2008 i686 GNU/Linux
Start Benchmark Run: 2008年 6月 7日 土曜日 22:33:21 JST
2 interactive users.
22:33:21 up 2:32, 2 users, load average: 0.29, 0.13, 0.04
lrwxrwxrwx 1 root root 4 2008-05-01 19:58 /bin/sh -> dash
/bin/sh: symbolic link to `dash'
/dev/sda1 74328480 41138584 29443944 59% /
Dhrystone 2 using register variables 4335930.0 lps (10 secs, 10 samples)
Arithmetic Test (type = double) 701982.1 lps (10 secs, 10 samples)
System Call Overhead 1073058.2 lps (10 secs, 10 samples)
Pipe Throughput 546424.4 lps (10 secs, 10 samples)
Pipe-based Context Switching 109051.2 lps (10 secs, 10 samples)
Process Creation 7929.9 lps (30 secs, 3 samples)
Execl Throughput 2420.9 lps (30 secs, 3 samples)
File Read 1024 bufsize 2000 maxblocks 887197.0 KBps (30 secs, 3 samples)
File Write 1024 bufsize 2000 maxblocks 411296.0 KBps (30 secs, 3 samples)
File Copy 1024 bufsize 2000 maxblocks 278376.0 KBps (30 secs, 3 samples)
File Read 256 bufsize 500 maxblocks 262372.0 KBps (30 secs, 3 samples)
File Write 256 bufsize 500 maxblocks 113266.0 KBps (30 secs, 3 samples)
File Copy 256 bufsize 500 maxblocks 78798.0 KBps (30 secs, 3 samples)
File Read 4096 bufsize 8000 maxblocks 2169279.0 KBps (30 secs, 3 samples)
File Write 4096 bufsize 8000 maxblocks 1248707.0 KBps (30 secs, 3 samples)
File Copy 4096 bufsize 8000 maxblocks 767117.0 KBps (30 secs, 3 samples)
Shell Scripts (1 concurrent) 3651.3 lpm (60 secs, 3 samples)
Shell Scripts (8 concurrent) 701.3 lpm (60 secs, 3 samples)
Shell Scripts (16 concurrent) 359.7 lpm (60 secs, 3 samples)
Arithmetic Test (type = short) 908474.7 lps (10 secs, 3 samples)
Arithmetic Test (type = int) 915412.1 lps (10 secs, 3 samples)
Arithmetic Test (type = long) 916176.0 lps (10 secs, 3 samples)
Arithmetic Test (type = float) 701548.2 lps (10 secs, 3 samples)
Arithoh nan lps (10 secs, 3 samples)
C Compiler Throughput inf lpm (40 secs, 3 samples)
Dc: sqrt(2) to 99 decimal places 92840.7 lpm (30 secs, 3 samples)
Recursion Test--Tower of Hanoi 120612.5 lps (20 secs, 3 samples)


INDEX VALUES
TEST BASELINE RESULT INDEX

Arithmetic Test (type = double) 29820.0 701982.1 235.4
Dhrystone 2 using register variables 116700.0 4335930.0 371.5
Execl Throughput 43.0 2420.9 563.0
File Copy 1024 bufsize 2000 maxblocks 3960.0 278376.0 703.0
File Copy 256 bufsize 500 maxblocks 1655.0 78798.0 476.1
File Copy 4096 bufsize 8000 maxblocks 5800.0 767117.0 1322.6
Pipe Throughput 12440.0 546424.4 439.2
Pipe-based Context Switching 4000.0 109051.2 272.6
Process Creation 126.0 7929.9 629.4
Shell Scripts (8 concurrent) 6.0 701.3 1168.8
System Call Overhead 15000.0 1073058.2 715.4
=========
FINAL SCORE 549.3
$

VirtualBox


VirtualBoxのバージョンは記録し忘れました。


今は、このバージョンです。ベンチマーク当時から、おそらく変わっていないと思います。



$ dpkg -l | grep virtualbox
ii virtualbox-ose 1.5.6-dfsg-6ubuntu1 x86 virtualization solution - binaries
ii virtualbox-ose-modules-2.6.24-16-generic 24 virtualbox-ose module for linux-image-2.6.24
ii virtualbox-ose-modules-2.6.24-17-generic 24.0.1 virtualbox-ose module for linux-image-2.6.24
ii virtualbox-ose-modules-2.6.24-18-generic 24.0.3 virtualbox-ose module for linux-image-2.6.24
ii virtualbox-ose-modules-2.6.24-19-generic 24.0.4 virtualbox-ose module for linux-image-2.6.24
ii virtualbox-ose-modules-generic 24.0.4 virtualbox-ose module for linux-image-generi
ii virtualbox-ose-source 1.5.6-dfsg-6ubuntu1 x86 virtualization solution - kernel module
$

ゲストOS


Ubuntu Serverです。



$ cat /etc/lsb-release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=8.04
DISTRIB_CODENAME=hardy
DISTRIB_DESCRIPTION="Ubuntu 8.04"
$ cat report804

BYTE UNIX Benchmarks (Version 4.0.1)
System -- Linux ubuntu804s 2.6.24-16-generic #1 SMP Thu Apr 10 13:23:42 UTC 2008 i686 GNU/Linux
Start Benchmark Run: Sun Jun 8 22:38:41 JST 2008
2 interactive users.
22:38:41 up 46 min, 2 users, load average: 0.08, 0.02, 0.01
lrwxrwxrwx 1 root root 4 May 10 23:51 /bin/sh -> dash
/bin/sh: symbolic link to `dash'
/dev/sda1 7913216 1060164 6454240 15% /
Dhrystone 2 using register variables 4388019.5 lps (10 secs, 10 samples)
Arithmetic Test (type = double) 700191.0 lps (10 secs, 10 samples)
System Call Overhead 1224515.1 lps (10 secs, 10 samples)
Pipe Throughput 830865.8 lps (10 secs, 10 samples)
Pipe-based Context Switching 9499.7 lps (10 secs, 10 samples)
Process Creation 672.3 lps (30 secs, 3 samples)
Execl Throughput 323.1 lps (29 secs, 3 samples)
File Read 1024 bufsize 2000 maxblocks 1282152.0 KBps (30 secs, 3 samples)
File Write 1024 bufsize 2000 maxblocks 569791.0 KBps (30 secs, 3 samples)
File Copy 1024 bufsize 2000 maxblocks 370647.0 KBps (30 secs, 3 samples)
File Read 256 bufsize 500 maxblocks 397187.0 KBps (30 secs, 3 samples)
File Write 256 bufsize 500 maxblocks 169311.0 KBps (30 secs, 3 samples)
File Copy 256 bufsize 500 maxblocks 113124.0 KBps (30 secs, 3 samples)
File Read 4096 bufsize 8000 maxblocks 2641597.0 KBps (30 secs, 3 samples)
File Write 4096 bufsize 8000 maxblocks 1502841.0 KBps (30 secs, 3 samples)
File Copy 4096 bufsize 8000 maxblocks 895282.0 KBps (30 secs, 3 samples)
Shell Scripts (1 concurrent) 877.0 lpm (60 secs, 3 samples)
Shell Scripts (8 concurrent) 117.7 lpm (60 secs, 3 samples)
Shell Scripts (16 concurrent) 59.0 lpm (60 secs, 3 samples)
Arithmetic Test (type = short) 935525.4 lps (10 secs, 3 samples)
Arithmetic Test (type = int) 915998.5 lps (10 secs, 3 samples)
Arithmetic Test (type = long) 941366.4 lps (10 secs, 3 samples)
Arithmetic Test (type = float) 711156.6 lps (10 secs, 3 samples)
Arithoh nan lps (10 secs, 3 samples)
C Compiler Throughput 424.6 lpm (60 secs, 3 samples)
Dc: sqrt(2) to 99 decimal places 35642.1 lpm (30 secs, 3 samples)
Recursion Test--Tower of Hanoi 121187.3 lps (20 secs, 3 samples)


INDEX VALUES
TEST BASELINE RESULT INDEX

Arithmetic Test (type = double) 29820.0 700191.0 234.8
Dhrystone 2 using register variables 116700.0 4388019.5 376.0
Execl Throughput 43.0 323.1 75.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 370647.0 936.0
File Copy 256 bufsize 500 maxblocks 1655.0 113124.0 683.5
File Copy 4096 bufsize 8000 maxblocks 5800.0 895282.0 1543.6
Pipe Throughput 12440.0 830865.8 667.9
Pipe-based Context Switching 4000.0 9499.7 23.7
Process Creation 126.0 672.3 53.4
Shell Scripts (8 concurrent) 6.0 117.7 196.2
System Call Overhead 15000.0 1224515.1 816.3
=========
FINAL SCORE 281.7
$

結論?


やっぱり、結果の見方が分からないので、どこに差があるのか説明できません。


lps(lines per second)は、1秒当たりの実行回数、lpmは、1分当たりの実行回数


INDEX VALUESは、何かのベースラインと比べた比率となっているようですが...





0 件のコメント: