ECSのLIVA-B3-2G-32GでUNIX Benchmarks PART2 (カスタムカーネル編)
LIVA用に作ったカスタムカーネルなCentOS7でUNIX Benchmarksをやり直したらディスクアクセス速度が凄まじく上昇した。
前回の記事にてLIVAにインストールしたカスタムカーネルなCentOS7ではeMMCの認識に必要なドライバをカーネルに組み込んでいました。
カーネルモジュールとして利用するよりもパフォーマンスがちょっと上昇したりしてないかなと思いUnix Benchmarksを取り直してみたところ、
以前Ubuntuで実施した時と比べて638.2→844.2の大幅な上昇で、特にファイルコピーの速度とパイプのスループット2倍弱に上昇しており、
ストレージ関連のドライバをカーネルへ組み込むのはかなり効果が大きいことが分かりました。
======================================================================== BYTE UNIX Benchmarks (Version 5.1.3) System: localhost.localdomain: GNU/Linux OS: GNU/Linux -- 3.16.2 -- #1 SMP Sat Sep 20 21:50:15 JST 2014 Machine: x86_64 (x86_64) Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8") CPU 0: Intel(R) Celeron(R) CPU N2806 @ 1.60GHz (3196.8 bogomips) Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization CPU 1: Intel(R) Celeron(R) CPU N2806 @ 1.60GHz (3196.8 bogomips) Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization 13:22:20 up 8:36, 1 user, load average: 0.23, 0.22, 0.13; runlevel 3 ------------------------------------------------------------------------ Benchmark Run: 日 9月 21 2014 13:22:20 - 13:50:17 2 CPUs in system; running 1 parallel copy of tests Dhrystone 2 using register variables 9847127.3 lps (10.0 s, 7 samples) Double-Precision Whetstone 1644.8 MWIPS (10.1 s, 7 samples) Execl Throughput 1406.5 lps (30.0 s, 2 samples) File Copy 1024 bufsize 2000 maxblocks 314179.0 KBps (30.0 s, 2 samples) File Copy 256 bufsize 500 maxblocks 89449.4 KBps (30.0 s, 2 samples) File Copy 4096 bufsize 8000 maxblocks 712974.1 KBps (30.0 s, 2 samples) Pipe Throughput 514748.3 lps (10.0 s, 7 samples) Pipe-based Context Switching 16936.3 lps (10.0 s, 7 samples) Process Creation 3903.9 lps (30.0 s, 2 samples) Shell Scripts (1 concurrent) 2112.5 lpm (60.0 s, 2 samples) Shell Scripts (8 concurrent) 444.0 lpm (60.1 s, 2 samples) System Call Overhead 898242.2 lps (10.0 s, 7 samples) System Benchmarks Index Values BASELINE RESULT INDEX Dhrystone 2 using register variables 116700.0 9847127.3 843.8 Double-Precision Whetstone 55.0 1644.8 299.1 Execl Throughput 43.0 1406.5 327.1 File Copy 1024 bufsize 2000 maxblocks 3960.0 314179.0 793.4 File Copy 256 bufsize 500 maxblocks 1655.0 89449.4 540.5 File Copy 4096 bufsize 8000 maxblocks 5800.0 712974.1 1229.3 Pipe Throughput 12440.0 514748.3 413.8 Pipe-based Context Switching 4000.0 16936.3 42.3 Process Creation 126.0 3903.9 309.8 Shell Scripts (1 concurrent) 42.4 2112.5 498.2 Shell Scripts (8 concurrent) 6.0 444.0 739.9 System Call Overhead 15000.0 898242.2 598.8 ======== System Benchmarks Index Score 439.6 ------------------------------------------------------------------------ Benchmark Run: 日 9月 21 2014 13:50:17 - 14:18:16 2 CPUs in system; running 2 parallel copies of tests Dhrystone 2 using register variables 19450016.5 lps (10.0 s, 7 samples) Double-Precision Whetstone 3251.8 MWIPS (10.1 s, 7 samples) Execl Throughput 2862.8 lps (30.0 s, 2 samples) File Copy 1024 bufsize 2000 maxblocks 506379.9 KBps (30.0 s, 2 samples) File Copy 256 bufsize 500 maxblocks 152404.5 KBps (30.0 s, 2 samples) File Copy 4096 bufsize 8000 maxblocks 1049099.5 KBps (30.0 s, 2 samples) Pipe Throughput 1024932.7 lps (10.0 s, 7 samples) Pipe-based Context Switching 124047.7 lps (10.0 s, 7 samples) Process Creation 8042.0 lps (30.0 s, 2 samples) Shell Scripts (1 concurrent) 3139.1 lpm (60.0 s, 2 samples) Shell Scripts (8 concurrent) 465.6 lpm (60.2 s, 2 samples) System Call Overhead 1501137.5 lps (10.0 s, 7 samples) System Benchmarks Index Values BASELINE RESULT INDEX Dhrystone 2 using register variables 116700.0 19450016.5 1666.7 Double-Precision Whetstone 55.0 3251.8 591.2 Execl Throughput 43.0 2862.8 665.8 File Copy 1024 bufsize 2000 maxblocks 3960.0 506379.9 1278.7 File Copy 256 bufsize 500 maxblocks 1655.0 152404.5 920.9 File Copy 4096 bufsize 8000 maxblocks 5800.0 1049099.5 1808.8 Pipe Throughput 12440.0 1024932.7 823.9 Pipe-based Context Switching 4000.0 124047.7 310.1 Process Creation 126.0 8042.0 638.3 Shell Scripts (1 concurrent) 42.4 3139.1 740.4 Shell Scripts (8 concurrent) 6.0 465.6 776.0 System Call Overhead 15000.0 1501137.5 1000.8 ======== System Benchmarks Index Score 844.2
広告
