先日 NAS 用のマシンに導入した FreeBSD 9.1-RC3 (amd64) で Unixbench を実行してみました。
| CPU | Intel CPU Core I3-3220T |
| メモリ | DDR3 PC3-12800 4GBx2 |
| ディスク | OCZ Agility4 64GB |
| OS | FreeBSD 9.1-RC3 (amd64) |
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)
System: FreeSMB.local: FreeBSD
OS: FreeBSD -- 9.1-RC3 -- FreeBSD 9.1-RC3 #0 r242324: Tue Oct 30 00:58:57 UTC 2012 root@farrell.cse.buffalo.edu:/usr/obj/usr/src/sys/GENERIC
Machine: amd64 (GENERIC)
Language: en_US.utf8 (charmap=, collate=)
CPU: no details available
11:30PM up 1 day, 2:11, 4 users, load averages: 2.34, 2.77, 3.02; runlevel
------------------------------------------------------------------------
Benchmark Run: 火 12月 11 2012 23:30:13 - 23:59:19
unknown CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 15595529.2 lps (10.0 s, 7 samples)
Double-Precision Whetstone 3715.8 MWIPS (9.8 s, 7 samples)
Execl Throughput 2998.8 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 74341.3 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 17112.7 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 269919.1 KBps (30.0 s, 2 samples)
Pipe Throughput 1325452.0 lps (10.0 s, 7 samples)
Pipe-based Context Switching 170155.4 lps (10.0 s, 7 samples)
Process Creation 8026.9 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 5786.2 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1151.5 lpm (60.0 s, 2 samples)
System Call Overhead 1012342.2 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 15595529.2 1336.4
Double-Precision Whetstone 55.0 3715.8 675.6
Execl Throughput 43.0 2998.8 697.4
File Copy 1024 bufsize 2000 maxblocks 3960.0 74341.3 187.7
File Copy 256 bufsize 500 maxblocks 1655.0 17112.7 103.4
File Copy 4096 bufsize 8000 maxblocks 5800.0 269919.1 465.4
Pipe Throughput 12440.0 1325452.0 1065.5
Pipe-based Context Switching 4000.0 170155.4 425.4
Process Creation 126.0 8026.9 637.1
Shell Scripts (1 concurrent) 42.4 5786.2 1364.7
Shell Scripts (8 concurrent) 6.0 1151.5 1919.2
System Call Overhead 15000.0 1012342.2 674.9
========
System Benchmarks Index Score 614.6
しかし、上の結果では “unknown CPUs in system” となり CPU 数を認識してくれません (CPUに関する情報を /proc/cpuinfo から取っているため)。そこで ここ で公開されているパッチを利用させていただき、再度実行してみました。
========================================================================
BYTE UNIX Benchmarks (Version 5.1.2-custom)
System: FreeSMB.local: FreeBSD
OS: FreeBSD -- 9.1-RC3 -- FreeBSD 9.1-RC3 #0 r242324: Tue Oct 30 00:58:57 UTC 2012 root@farrell.cse.buffalo.edu:/usr/obj/usr/src/sys/GENERIC
Machine: amd64 (GENERIC)
Language: en_US.utf8 (charmap=, collate=)
Compiler: gcc (GCC) 4.2.1 20070831 patched [FreeBSD]
CPU 0: hw.model: Intel(R) Core(TM) i3-3220T CPU @ 2.80GHz (0.0 bogomips)
CPU 1: hw.model: Intel(R) Core(TM) i3-3220T CPU @ 2.80GHz (0.0 bogomips)
CPU 2: hw.model: Intel(R) Core(TM) i3-3220T CPU @ 2.80GHz (0.0 bogomips)
CPU 3: hw.model: Intel(R) Core(TM) i3-3220T CPU @ 2.80GHz (0.0 bogomips)
2:23AM up 1 day, 5:04, 3 users, load averages: 2.90, 2.57, 2.72; runlevel
------------------------------------------------------------------------
Benchmark Run: 水 12月 12 2012 02:23:32 - 02:52:37
4 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 14333107.1 lps (10.0 s, 7 samples)
Double-Precision Whetstone 3631.5 MWIPS (9.8 s, 7 samples)
Execl Throughput 2685.6 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 67892.4 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 17257.6 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 273503.9 KBps (30.0 s, 2 samples)
Pipe Throughput 1334970.8 lps (10.0 s, 7 samples)
Pipe-based Context Switching 168712.2 lps (10.0 s, 7 samples)
Process Creation 8148.1 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 5857.9 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1170.0 lpm (60.0 s, 2 samples)
System Call Overhead 1029659.0 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 14333107.1 1228.2
Double-Precision Whetstone 55.0 3631.5 660.3
Execl Throughput 43.0 2685.6 624.6
File Copy 1024 bufsize 2000 maxblocks 3960.0 67892.4 171.4
File Copy 256 bufsize 500 maxblocks 1655.0 17257.6 104.3
File Copy 4096 bufsize 8000 maxblocks 5800.0 273503.9 471.6
Pipe Throughput 12440.0 1334970.8 1073.1
Pipe-based Context Switching 4000.0 168712.2 421.8
Process Creation 126.0 8148.1 646.7
Shell Scripts (1 concurrent) 42.4 5857.9 1381.6
Shell Scripts (8 concurrent) 6.0 1170.0 1950.0
System Call Overhead 15000.0 1029659.0 686.4
========
System Benchmarks Index Score 603.0
------------------------------------------------------------------------
Benchmark Run: 水 12月 12 2012 02:52:37 - 03:22:03
4 CPUs in system; running 4 parallel copies of tests
Dhrystone 2 using register variables 40456167.5 lps (10.0 s, 7 samples)
Double-Precision Whetstone 13538.1 MWIPS (9.8 s, 7 samples)
Execl Throughput 4898.4 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 67320.4 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 16861.1 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 259059.4 KBps (30.0 s, 2 samples)
Pipe Throughput 3849589.3 lps (10.0 s, 7 samples)
Pipe-based Context Switching 673858.5 lps (10.0 s, 7 samples)
Process Creation 12580.0 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 9850.6 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1561.1 lpm (60.1 s, 2 samples)
System Call Overhead 2987632.0 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 40456167.5 3466.7
Double-Precision Whetstone 55.0 13538.1 2461.5
Execl Throughput 43.0 4898.4 1139.2
File Copy 1024 bufsize 2000 maxblocks 3960.0 67320.4 170.0
File Copy 256 bufsize 500 maxblocks 1655.0 16861.1 101.9
File Copy 4096 bufsize 8000 maxblocks 5800.0 259059.4 446.7
Pipe Throughput 12440.0 3849589.3 3094.5
Pipe-based Context Switching 4000.0 673858.5 1684.6
Process Creation 126.0 12580.0 998.4
Shell Scripts (1 concurrent) 42.4 9850.6 2323.2
Shell Scripts (8 concurrent) 6.0 1561.1 2601.9
System Call Overhead 15000.0 2987632.0 1991.8
========
System Benchmarks Index Score 1137.9
ちなみに、このブロクで利用しているさくら VPS 1GB (Debian Squeeze) での Unixbench は次のようになりました。
| CPU | Intel(R) Xeon(R) CPU E5645 (仮想2core) |
| メモリ | 1GB |
| ディスク | 100GB |
| OS | Debian Squeeze, GNU/Linux 2.6.32-5-amd64 |
| リージョン | 大阪 |
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)
System: wwwxxxx: GNU/Linux
OS: GNU/Linux -- 2.6.32-5-amd64 -- #1 SMP Sun Sep 23 10:07:46 UTC 2012
Machine: x86_64 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Xeon(R) CPU E5645 (4800.2 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET
CPU 1: Intel(R) Xeon(R) CPU E5645 (4800.2 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET
23:13:59 up 39 days, 59 min, 1 user, load average: 0.07, 0.02, 0.00; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Tue Dec 11 2012 23:13:59 - 23:42:13
2 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 21026271.9 lps (10.0 s, 7 samples)
Double-Precision Whetstone 2808.6 MWIPS (9.8 s, 7 samples)
Execl Throughput 1836.4 lps (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 510326.9 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 186278.8 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1095003.6 KBps (30.0 s, 2 samples)
Pipe Throughput 1802119.5 lps (10.0 s, 7 samples)
Pipe-based Context Switching 18009.8 lps (10.0 s, 7 samples)
Process Creation 5080.0 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 4527.7 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1065.4 lpm (60.0 s, 2 samples)
System Call Overhead 3197935.7 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 21026271.9 1801.7
Double-Precision Whetstone 55.0 2808.6 510.7
Execl Throughput 43.0 1836.4 427.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 510326.9 1288.7
File Copy 256 bufsize 500 maxblocks 1655.0 186278.8 1125.6
File Copy 4096 bufsize 8000 maxblocks 5800.0 1095003.6 1887.9
Pipe Throughput 12440.0 1802119.5 1448.6
Pipe-based Context Switching 4000.0 18009.8 45.0
Process Creation 126.0 5080.0 403.2
Shell Scripts (1 concurrent) 42.4 4527.7 1067.9
Shell Scripts (8 concurrent) 6.0 1065.4 1775.6
System Call Overhead 15000.0 3197935.7 2132.0
========
System Benchmarks Index Score 834.7
------------------------------------------------------------------------
Benchmark Run: Tue Dec 11 2012 23:42:13 - 00:10:29
2 CPUs in system; running 2 parallel copies of tests
Dhrystone 2 using register variables 41156890.2 lps (10.0 s, 7 samples)
Double-Precision Whetstone 5566.1 MWIPS (9.9 s, 7 samples)
Execl Throughput 5449.9 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 583950.5 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 185449.7 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1390411.1 KBps (30.0 s, 2 samples)
Pipe Throughput 3439889.4 lps (10.0 s, 7 samples)
Pipe-based Context Switching 546665.8 lps (10.0 s, 7 samples)
Process Creation 20555.1 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 8796.4 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1160.9 lpm (60.1 s, 2 samples)
System Call Overhead 5057060.5 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 41156890.2 3526.7
Double-Precision Whetstone 55.0 5566.1 1012.0
Execl Throughput 43.0 5449.9 1267.4
File Copy 1024 bufsize 2000 maxblocks 3960.0 583950.5 1474.6
File Copy 256 bufsize 500 maxblocks 1655.0 185449.7 1120.5
File Copy 4096 bufsize 8000 maxblocks 5800.0 1390411.1 2397.3
Pipe Throughput 12440.0 3439889.4 2765.2
Pipe-based Context Switching 4000.0 546665.8 1366.7
Process Creation 126.0 20555.1 1631.4
Shell Scripts (1 concurrent) 42.4 8796.4 2074.6
Shell Scripts (8 concurrent) 6.0 1160.9 1934.8
System Call Overhead 15000.0 5057060.5 3371.4
========
System Benchmarks Index Score 1838.9
比較すると VPS のお得度がわかります。価格 (メモリ1G, 月々￥980) の割にはとても良い値が出ているようです。やはり VPS は利用しない手はないですね。