Google Cloud の Compute Engine を使う機会があったのでメモ。
マシンタイプ
n1-standard-2(vCPU x 2、メモリ 7.5 GB)
30GB SSD
CentOS Linux release 7.7.1908 (Core)
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)
System: xxx-vm: GNU/Linux
OS: GNU/Linux -- 3.10.0-1062.12.1.el7.x86_64 -- #1 SMP Tue Feb 4 23:02:59 UTC 2020
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Xeon(R) CPU @ 2.00GHz (4000.3 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Xeon(R) CPU @ 2.00GHz (4000.3 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
18:57:23 up 17 days, 22:15, 2 users, load average: 0.00, 0.01, 0.05; runlevel 3
------------------------------------------------------------------------
Benchmark Run: 金 3月 13 2020 18:57:23 - 19:25:28
2 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 32986641.0 lps (10.0 s, 7 samples)
Double-Precision Whetstone 3914.6 MWIPS (9.8 s, 7 samples)
Execl Throughput 1949.5 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 342471.9 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 91571.0 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1050008.3 KBps (30.0 s, 2 samples)
Pipe Throughput 433570.6 lps (10.0 s, 7 samples)
Pipe-based Context Switching 47364.9 lps (10.0 s, 7 samples)
Process Creation 5337.6 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 3961.7 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 641.7 lpm (60.1 s, 2 samples)
System Call Overhead 325793.5 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 32986641.0 2826.6
Double-Precision Whetstone 55.0 3914.6 711.8
Execl Throughput 43.0 1949.5 453.4
File Copy 1024 bufsize 2000 maxblocks 3960.0 342471.9 864.8
File Copy 256 bufsize 500 maxblocks 1655.0 91571.0 553.3
File Copy 4096 bufsize 8000 maxblocks 5800.0 1050008.3 1810.4
Pipe Throughput 12440.0 433570.6 348.5
Pipe-based Context Switching 4000.0 47364.9 118.4
Process Creation 126.0 5337.6 423.6
Shell Scripts (1 concurrent) 42.4 3961.7 934.4
Shell Scripts (8 concurrent) 6.0 641.7 1069.4
System Call Overhead 15000.0 325793.5 217.2
========
System Benchmarks Index Score 616.2
------------------------------------------------------------------------
Benchmark Run: 金 3月 13 2020 19:25:28 - 19:53:37
2 CPUs in system; running 2 parallel copies of tests
Dhrystone 2 using register variables 35009933.2 lps (10.0 s, 7 samples)
Double-Precision Whetstone 6775.5 MWIPS (10.0 s, 7 samples)
Execl Throughput 2934.5 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 434373.3 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 114990.4 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1382207.8 KBps (30.0 s, 2 samples)
Pipe Throughput 554561.5 lps (10.0 s, 7 samples)
Pipe-based Context Switching 135627.6 lps (10.0 s, 7 samples)
Process Creation 10330.6 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 4781.7 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 644.4 lpm (60.1 s, 2 samples)
System Call Overhead 389906.7 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 35009933.2 3000.0
Double-Precision Whetstone 55.0 6775.5 1231.9
Execl Throughput 43.0 2934.5 682.5
File Copy 1024 bufsize 2000 maxblocks 3960.0 434373.3 1096.9
File Copy 256 bufsize 500 maxblocks 1655.0 114990.4 694.8
File Copy 4096 bufsize 8000 maxblocks 5800.0 1382207.8 2383.1
Pipe Throughput 12440.0 554561.5 445.8
Pipe-based Context Switching 4000.0 135627.6 339.1
Process Creation 126.0 10330.6 819.9
Shell Scripts (1 concurrent) 42.4 4781.7 1127.8
Shell Scripts (8 concurrent) 6.0 644.4 1074.0
System Call Overhead 15000.0 389906.7 259.9
========
System Benchmarks Index Score 866.3
最初に、GCPは¥32,579の無料クレジットがあるんで最初の2、3ヶ月くらいはタダでいけそうですね。
866.3 ってことは意外に速くはないんですね。
以前(2018/01)に、g1-small(vCPU x 1、メモリ 1.7 GB)を計測したことがあります。そん時は、683でした。
GCP f1-micro の UnixBench
https://hack.gpl.jp/2020/03/13/gcp-unixbench-n1standard2/
2Core にして、シングルの1.4倍ということですか。
こんな構成のときは、もう少しスコアが出てましたね。
独自ドメインで WordPress を動作させる為、エンジニアが選んだ選択
::
score 1146
SSD のn1-standard-1(vCPU x 1、メモリ 3.75 GB)
自宅サーバにするときは、1000〜2000くらいのスペックが出るものを選ぶようにしますか。あまり高機能だと電気代がかかりますからね。
ヤフオクに転がってる、i5 (4世代くらい)で十分ですかね。