[B! performance] clairvyã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

clairvy id:clairvy

performanceã«é–¢ã™ã‚‹clairvyã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (11)

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

${{author_name}}$
{{author_name}}{{created}}
{{ #comment }}{{ comment }}{{ /comment }}
- {{ label }}

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

High performance servers without the event loop
High performance servers without the event loop OSCON, Portland 21 July 2015 Dave Cheney whoami(1) Sysadmin before Devops Transitioned to Go 5 years ago Work for Canonical contributing to the Go project 2 Who is this presentation for ? The horror! As an admin in a past life, the most stressful times in my career were defined by unsatisfying performance. I got into Go because of its potential to bu
clairvy 2015/07/24
golang

performance
ãƒªãƒ³ã‚¯
MySQLã‚¤ãƒ³ãƒ‡ãƒƒã‚¯ã‚¹ã®åŸºç¤Ž : ã²ã¨ã¤ã®ãƒ†ãƒ¼ãƒ–ãƒ«ã«å¯¾ã™ã‚‹ã‚¯ã‚¨ãƒªã®æœ€é©åŒ–æ³• | Yakst
MySQLã®ã‚¤ãƒ³ãƒ‡ãƒƒã‚¯ã‚¹ã‚’åŠ¹æžœçš„ã«ä½¿ã†ã«ã¯ã©ã†ã—ãŸã‚‰ã„ã„ã®ã‹ã«ã¤ã„ã¦ã®åˆ†ã‹ã‚Šã‚„ã™ã„è§£èª¬ã€‚ãã‚‚ãã‚‚ã‚¤ãƒ³ãƒ‡ãƒƒã‚¯ã‚¹ã®å½¹å‰²ã¯ã¨ã¯ä½•ã‹ã€ãã—ã¦ã©ã†ã™ã‚Œã°ãã®å½¹å‰²ã‚’æžœãŸã—ã¦ãã‚Œã‚‹ã®ã‹ã‚’èª¬æ˜Žã™ã‚‹ã€‚ ãŸã¨ãˆ1ã¤ã®ãƒ†ãƒ¼ãƒ–ãƒ«ã ã‘ã«å¯¾ã—ã¦å®Ÿè¡Œã•ã‚Œã‚‹ã‚¯ã‚¨ãƒªã§ã‚‚ã€ãƒ‘ãƒ•ã‚©ãƒ¼ãƒžãƒ³ã‚¹ãŒæ‚ªã„ã¨ã„ã†ã®ã¯ã‚ˆãã‚ã‚‹ã“ã¨ã§ã™ã€‚ãã®ç†ç”±ã¯ç°¡å˜ã§ã€ã‚¤ãƒ³ãƒ‡ãƒƒã‚¯ã‚¹ã®ä½œã‚Šæ–¹ãŒã¾ãšã„ãŸã‚ã€å®Ÿè¡Œè¨ˆç”»ãŒãŠã‹ã—ããªã£ã¦ã—ã¾ã†ã®ã§ã™ã€‚ã“ã“ã§ã¯ã€1ã¤ã®ãƒ†ãƒ¼ãƒ–ãƒ«ã®ã¿ã«å¯¾ã™ã‚‹è‰²ã€…ãªã‚¯ã‚¨ãƒªã‚’æœ€é©åŒ–ã™ã‚‹ãŸã‚ã®ã‚¬ã‚¤ãƒ‰ãƒ©ã‚¤ãƒ³ã‚’æŒ™ã’ã¦ã¿ãŸã„ã¨æ€ã„ã¾ã™ã€‚ ãŠã“ã¨ã‚ã‚Š : ã‚ã‚‰ã‚†ã‚‹çŠ¶æ³ã‚’ã‚«ãƒãƒ¼ã—ã‚ˆã†ã¨ã¯ã›ãšã€ä¸€èˆ¬çš„ãªã‚¬ã‚¤ãƒ‰ãƒ©ã‚¤ãƒ³ã‚’æç¤ºã™ã‚‹ã«ç•™ã‚ã‚‹ã¤ã‚‚ã‚Šã§ã™ã€‚ã“ã“ã§æŒ™ã’ãŸã‚‚ã®ãŒã†ã¾ãé©ç”¨ã§ããªã„ä¾‹ã‚’ç°¡å˜ã«è¦‹ã¤ã‘ã‚‹ã“ã¨ãŒã§ãã‚‹ã®ã¯é–“é•ã„ãªã„ã§ã—ã‚‡ã†ãŒã€ã»ã¨ã‚“ã©ã®å ´åˆã¯ã“ã“ã«æ›¸ã„ãŸã“ã¨ãŒååˆ†ãªã®ã‚‚äº‹å®Ÿã§ã™ã€‚ã¾ãŸã€MySQL 5.6ä»¥ä¸Šã«ã‚ã‚‹Index Condition Pu
clairvy 2015/05/26
mysql

performance

Index
ãƒªãƒ³ã‚¯
Linuxã§ãƒãƒ¼ãƒ‰ãƒãƒ©ãƒ³ã‚µã‚„ã‚ãƒ£ãƒƒã‚·ãƒ¥ã‚µãƒ¼ãƒã‚’ãƒžãƒ«ãƒã‚³ã‚¢ã‚¹ã‚±ãƒ¼ãƒ«ã•ã›ã‚‹ãŸã‚ã®ã‚«ãƒ¼ãƒãƒ«ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚° - ã‚†ã†ã†ããƒ–ãƒã‚°
æœ¬è¨˜äº‹ã®å…¬é–‹å¾Œã®2016å¹´7æœˆã«ã¯ã¦ãªã«ãŠã‘ã‚‹ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°äº‹ä¾‹ã‚’ç´¹ä»‹ã—ãŸã€‚ ã¯ã¦ãªã«ãŠã‘ã‚‹Linuxãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚¹ã‚¿ãƒƒã‚¯ãƒ‘ãƒ•ã‚©ãƒ¼ãƒžãƒ³ã‚¹æ”¹å–„ / Linux network performance improvement at hatena - Speaker Deck HAProxy ã‚„ nginx ãªã©ã®ã‚½ãƒ•ãƒˆã‚¦ã‚§ã‚¢ãƒãƒ¼ãƒ‰ãƒãƒ©ãƒ³ã‚µã‚„ãƒªãƒãƒ¼ã‚¹ãƒ—ãƒã‚ã‚·ã€memcached ãªã©ã® KVS ã®ã‚ˆã†ãªé«˜ãƒ‘ã‚±ãƒƒãƒˆãƒ¬ãƒ¼ãƒˆã«ãªã‚Šã‚„ã™ã„ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³ã«ãŠã„ã¦ã€å˜ä¸€ã® CPU ã‚³ã‚¢ã«è² è·ãŒåã‚Šã€ãƒžãƒ«ãƒã‚³ã‚¢ã‚¹ã‚±ãƒ¼ãƒ«ã—ãªã„ã“ã¨ãŒã‚ã‚Šã¾ã™ã€‚ ä»Šå›žã¯ã€ã“ã®ã‚ˆã†ãªãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³ã«ãŠã„ã¦ CPU è² è·ãŒãƒžãƒ«ãƒã‚³ã‚¢ã‚¹ã‚±ãƒ¼ãƒ«ã—ãªã„ç†ç”±ã¨ã€ãƒžãƒ«ãƒã‚³ã‚¢ã‚¹ã‚±ãƒ¼ãƒ«ã•ã›ã‚‹ãŸã‚ã® Linux ã‚«ãƒ¼ãƒãƒ«ã®ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚¹ã‚¿ãƒƒã‚¯ã®ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°æ‰‹æ³•ã¨ã—ã¦ RFS (Receive Flow Steering) ã‚’
clairvy 2015/03/31
performance

kernel

network
ãƒªãƒ³ã‚¯
Performance Tuning Guide | Red Hat Enterprise Linux | 7 | Red Hat Documentation
clairvy 2015/02/14
performance

tuning

Linux

rhel
ãƒªãƒ³ã‚¯
JMeterä½¿ã†ã®ã ã‚‹ã„ãªãƒ¼ã¨æ€ã£ã¦ãŸã‚‰ruby-jmeterã¨ã„ã†Rubyã§ãƒ†ã‚¹ãƒˆãƒ—ãƒ©ãƒ³ã‚’æ›¸ã‘ã‚‹ãƒ„ãƒ¼ãƒ«ãŒã‚ã£ãŸ | mah365
JMeterä½¿ã†ã®ã ã‚‹ã„ãªãƒ¼ã¨æ€ã£ã¦ãŸã‚‰ruby-jmeterã¨ã„ã†Rubyã§ãƒ†ã‚¹ãƒˆãƒ—ãƒ©ãƒ³ã‚’æ›¸ã‘ã‚‹ãƒ„ãƒ¼ãƒ«ãŒã‚ã£ãŸã€‚çŸ¥ã‚‰ãªã‹ã£ãŸï¼ˆè¿«çœŸï¼‰ã€‚ å…¸åž‹çš„ãªRailsã‚¢ãƒ—ãƒªã®ãƒ†ã‚¹ãƒˆãƒ—ãƒ©ãƒ³ ãã†ã„ã†è¨³ã§å…¸åž‹çš„ãªRailsã‚¢ãƒ—ãƒªã®ãƒ†ã‚¹ãƒˆãƒ—ãƒ©ãƒ³ã‚’æ›¸ã„ã¦ã¿ãŸã®ãŒã“ã¡ã‚‰ã€‚ ãƒ¦ãƒ¼ã‚¶ãƒ¼ãƒã‚°ã‚¤ãƒ³ãƒšãƒ¼ã‚¸ã§CSRFãƒˆãƒ¼ã‚¯ãƒ³ã‚’å–å¾—ã—ã€å¸¸ã«HTTPãƒ˜ãƒƒãƒ€ã«ã¤ã‘ã‚‹ã‚ˆã†ã«ã™ã‚‹ ãƒ¦ãƒ¼ã‚¶ãƒ¼ãƒã‚°ã‚¤ãƒ³æƒ…å ±ã‚’ã‚¯ãƒƒã‚ãƒ¼ã«ä¿å˜ ã¨ã„ã£ãŸå…¸åž‹çš„ãªå‡¦ç†ã‚’ç››ã‚Šè¾¼ã‚“ã§ã„ã¾ã™ã€‚ã‚ã¨ã¯README.mdã‚’èªã‚“ã§ã‚‚ã‚‰ãˆã‚Œã°å¤§ä½“ã®æ›¸ãæ–¹ãŒæŠŠæ¡ã§ãã‚‹ã‹ã¨æ€ã„ã¾ã™ã€‚ ã¡ãªã¿ã«ã€# Debugã¨ã„ã†ã‚³ãƒ¡ãƒ³ãƒˆã®ä¸‹2è¡Œã‚’ã‚³ãƒ¡ãƒ³ãƒˆã‚¢ã‚¦ãƒˆã—ã¦ã‚‚ã‚‰ã†ã¨ã€JMeterä¸Šã§ãƒ‡ãƒãƒƒã‚°ç”¨ã®å‡ºåŠ›ã‚’è¡¨ç¤ºã™ã‚‹ã“ã¨ãŒã§ãã¾ã™ã€‚ãƒ†ã‚¹ãƒˆãƒ—ãƒ©ãƒ³ãŒä¸Šæ‰‹ãå‹•ã‹ãªã„ã¨ãã«ã€ãƒªã‚¯ã‚¨ã‚¹ãƒˆãƒ˜ãƒƒãƒ€ã‚„ãƒ¬ã‚¹ãƒãƒ³ã‚¹ã‚’ç¢ºèªã™ã‚‹ã®ã«ä¾¿åˆ©ã§ã™ã€‚ ã§ã€ã“ã‚Œã‚’ã‚³ãƒžãƒ³ãƒ‰ãƒ©ã‚¤ãƒ³ã§ ruby sample.jmx.rb && j
clairvy 2014/11/23
Ruby

jmeter

performance
ãƒªãƒ³ã‚¯
ã€ŒRaptorã¯ã©ã®ã‚ˆã†ã«ã—ã¦Unicornã®4å€ã€Puma, Torqueboxã®2å€ã®é€Ÿåº¦ã‚’é”æˆã—ãŸã®ã‹ã€ã‚’èªã‚“ã§ã¾ã¨ã‚ã¦ã¿ãŸ - Qiita
Help us understand the probl em. What is going on with this article? ã€ŒRaptorã¯ã©ã®ã‚ˆã†ã«ã—ã¦Unicornã®4å€ã€Puma, Torqueboxã®2å€ã®é€Ÿåº¦ã‚’é”æˆã—ãŸã®ã‹ã€ã‚’èªã‚“ã§ã¾ã¨ã‚ã¦ã¿ã¾ã—ãŸã€‚ åŽŸæ–‡ã¯ã“ã¡ã‚‰ã§ã™ã€‚ç´¹ä»‹ã«ã¤ã„ã¦ã¯è¨±å¯ã‚’è²°ã£ã¦ã„ã¾ã™ã€‚ How we've made Raptor up to 4x faster than Unicorn, up to 2x faster than Puma, Torquebox ã¨ã¦ã‚‚èªã¿ã‚„ã™ã„è‹±èªžã§ã™ã®ã§æ˜¯éžåŽŸæ–‡ã‚‚èªã‚“ã§ã¿ã¦ãã ã•ã„ã€‚ How Ruby app servers work Rackã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³ã®æ§‹æˆã«ã¤ã„ã¦ã®ç´¹ä»‹ã¨ã€ã‚³ãƒã‚¯ã‚·ãƒ§ãƒ³ã‚’ã©ã®ã‚ˆã†ã«æ‰±ã†ã®ã‹ã«ã¤ã„ã¦ã€‚ prefork/threadingã‚„Blocking I/OãŠã‚ˆã³Event I/Oã®çµ„ã¿
clairvy 2014/11/11
Ruby

performance

Raptor
ãƒªãƒ³ã‚¯
Dockerã¯é€Ÿã„ã®ã‹ï¼ŸDockerã®ãƒ‘ãƒ•ã‚©ãƒ¼ãƒžãƒ³ã‚¹ã«ã¤ã„ã¦é‡è¦ãªã“ã¨ã¯ä½•ã‹ï¼Ÿ - ã‚†ã†ã†ããƒ–ãƒã‚°
ã ã„ã¶å‰ã‹ã‚‰Dockerï¼ˆLinuxã‚³ãƒ³ãƒ†ãƒŠï¼‰ã®ãƒ‘ãƒ•ã‚©ãƒ¼ãƒžãƒ³ã‚¹ã«ã¤ã„ã¦ã€é€Ÿã„ã“ã¨ã¯é€Ÿã„ã ã‚ã†ã‘ã©ã©ã®ç¨‹åº¦é€Ÿã„ã®ã‹ã€ã‚‚ã—é…ã„ã“ã¨ãŒã‚ã‚‹ãªã‚‰ä½•ãŒãƒ‘ãƒ•ã‚©ãƒ¼ãƒžãƒ³ã‚¹ã«ã¨ã£ã¦é‡è¦ãªã®ã‹ï¼ˆAUFSãŒé…ã„ã¨ã‹ãã†ã„ã†ã®ï¼‰ãŒæ°—ã«ãªã£ã¦ã„ãŸã®ã§ã€ä»Šå›žã¯ ã§ç´¹ä»‹ã•ã‚Œã¦ã„ãŸ Docker ã®ãƒ‘ãƒ•ã‚©ãƒ¼ãƒžãƒ³ã‚¹æ¤œè¨¼ã«é–¢ã™ã‚‹ IBM ã® Research Report ã‚’èªã‚“ã ã€‚Report ã®å†…å®¹ã‚’ãƒ™ãƒ¼ã‚¹ã«ã€Docker ã®ãƒ‘ãƒ•ã‚©ãƒ¼ãƒžãƒ³ã‚¹ã®å‹˜æ‰€ãªã©ã‚’ã¾ã¨ã‚ã¦ã¿ãŸã€‚ Report ã®ã‚¿ã‚¤ãƒˆãƒ«ã¯ An Updated Performance Comparison of Virtual Machines and Linux Containers ã€‚ GitHub ã«ãƒ™ãƒ³ãƒãƒžãƒ¼ã‚¯ã‚³ãƒ¼ãƒ‰ã¨å®Ÿé¨“ãƒ‡ãƒ¼ã‚¿ãŒç½®ã„ã¦ã‚ã£ã¦ã¡ã‚ƒã‚“ã¨ã—ã¦ã‚‹ã€‚ å‰æ ã¾ãšã€VMã¨ã‚³ãƒ³ãƒ†ãƒŠã®æ´å²ã‚’æŒ¯ã‚Šè¿”ã‚‹ã®ã«çŸ¥ã‚‰ã¬ã¯ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ã®æ¥ã€‚ä»Šã•ã‚‰èžã‘ãªã„ã€ã‚³ãƒ³ãƒ†ãƒŠ/ä»®æƒ³
clairvy 2014/11/10
docker

performance
ãƒªãƒ³ã‚¯
Linux Performance
static, benchmarking, tuning: sar, perf-tools, bcc/BPF: bpftrace, BPF book: Images license: creative commons Attribution-ShareAlike 4.0. This page links to various Linux performance material I've created, including the tools maps on the right. These use a large font size to suit slide decks. You can also print them out for your office wall. They show: Linux observability tools, Linux static perfor
clairvy 2014/07/27
Linux

performance
ãƒªãƒ³ã‚¯
High Performance Rails (long edition)
Which I talked at Rubykaigi 2013 is: https://speakerdeck.com/mirakui/high-performance-rails
clairvy 2013/06/01
Rails

é–‹ç™º

performance
ãƒªãƒ³ã‚¯
Node Templates
Overall Compile and Render Benchmark Scores (bigger is better). Run on a 2.93 GHz Intel Core i7 iMac with 12GB RAM. Engine (npm name) Average (mean) Scores (bigger is better) swig
clairvy 2013/05/23
node.js

template

performance
ãƒªãƒ³ã‚¯
JSã§i++ã¨++iã©ã£ã¡ãŒé€Ÿã„ï¼Ÿ - Islands in the byte stream (legacy)
çµè«–ã‹ã‚‰è¨€ã†ã¨ã€ç¾åœ¨ã®Chromeã®ã¿ã‚’ã‚¿ãƒ¼ã‚²ãƒƒãƒˆã«ã—ã¦æœ€é©åŒ–ã™ã‚‹ã¨ã„ã†ç‰¹æ®Šãªã‚±ãƒ¼ã‚¹ã‚’é™¤ã*1ã€ã©ã¡ã‚‰ã§ã‚‚å¤‰ã‚ã‚‰ãªã„ã¨ã„ãˆã¾ã™ã€‚ [è¿½è¨˜] æŒ‡æ‘˜ã‚’å—ã‘ã¦å†è€ƒã—ã¦ã¿ã¾ã—ãŸã€‚ãã‚‚ãã‚‚jsperfã§ã¯åˆæœŸåŒ–ã‚³ãƒ¼ãƒ‰ï¼ˆä»Šå›žã¯dataãªã©ã®åˆæœŸåŒ–ã«ä½¿ç”¨ï¼‰ã¯ä¸€åº¦ã—ã‹èµ°ã‚‰ãªã„ã«ã‚‚ã‹ã‹ã‚ã‚‰ãšã€ã“ã®ãƒ™ãƒ³ãƒãƒžãƒ¼ã‚¯ã‚³ãƒ¼ãƒ‰ã§ã¯dataã®ä¸èº«ã‚’æ›¸ãæ›ãˆã¦ã„ã¾ã™ã€‚ã“ã‚ŒãŒã‚¹ã‚³ã‚¢ã«å½±éŸ¿ã‚’ä¸Žãˆã¦ã„ãŸã‚ˆã†ã§ã™ã€‚ data[index] = data[index] * 2ã‚’data[index] = index * c ã«ã—ãŸçµæžœã¯Chromeã§ã‚‚å®‰å®šã—ã¦åŒæ–¹æœ‰æ„å·®ãªã—ã¨ã„ã†çµæžœã«ãªã‚Šã¾ã—ãŸã€‚ http://jsperf.com/postfix-or-pretfix-increment/4 JSã®ãƒ™ãƒ³ãƒãƒžãƒ¼ã‚¯ã®é›£ã—ã•ã‚’æ€ã„çŸ¥ã‚Šã¾ã—ãŸã€‚ [/è¿½è¨˜] http://jsperf.com/prefix-or-postfix-incre
clairvy 2012/03/18
javascript

performance
ãƒªãƒ³ã‚¯
1