Recent Comments
Archives
- August 2017 (30)
- July 2017 (41)
- June 2017 (37)
- May 2017 (59)
- April 2017 (55)
- March 2017 (55)
- February 2017 (35)
- January 2017 (42)
- December 2016 (48)
- November 2016 (32)
- October 2016 (35)
- September 2016 (78)
- August 2016 (69)
- July 2016 (19)
- June 2016 (42)
- May 2016 (61)
- April 2016 (51)
- March 2016 (74)
- February 2016 (87)
- January 2016 (31)
- December 2015 (36)
- November 2015 (61)
- October 2015 (72)
- September 2015 (53)
- August 2015 (42)
- July 2015 (38)
- June 2015 (30)
- May 2015 (18)
- April 2015 (57)
- March 2015 (41)
- February 2015 (50)
- January 2015 (35)
- December 2014 (50)
- November 2014 (56)
- October 2014 (41)
- September 2014 (37)
- August 2014 (37)
- July 2014 (28)
- June 2014 (50)
- May 2014 (32)
- April 2014 (46)
- March 2014 (38)
- February 2014 (29)
- January 2014 (52)
- December 2013 (50)
- November 2013 (45)
- October 2013 (40)
- September 2013 (48)
- August 2013 (22)
- July 2013 (25)
- June 2013 (13)
- May 2013 (16)
- April 2013 (28)
- March 2013 (37)
- February 2013 (36)
- January 2013 (57)
- December 2012 (44)
- November 2012 (10)
- October 2012 (12)
- September 2012 (21)
- August 2012 (21)
- July 2012 (25)
- June 2012 (8)
- May 2012 (10)
- April 2012 (11)
- March 2012 (10)
- February 2012 (11)
- January 2012 (5)
- December 2011 (13)
- November 2011 (12)
- October 2011 (10)
- September 2011 (7)
- August 2011 (5)
- July 2011 (11)
- June 2011 (21)
- May 2011 (22)
- April 2011 (36)
- March 2011 (43)
- February 2011 (23)
- January 2011 (24)
- December 2010 (34)
- November 2010 (19)
- October 2010 (16)
- September 2010 (15)
- August 2010 (10)
- July 2010 (12)
- June 2010 (3)
- May 2010 (3)
- April 2010 (4)
- March 2010 (8)
- February 2010 (14)
- January 2010 (13)
- December 2009 (16)
- November 2009 (28)
- October 2009 (24)
- September 2009 (12)
- August 2009 (7)
- July 2009 (10)
- June 2009 (11)
- May 2009 (22)
- April 2009 (21)
- March 2009 (18)
- February 2009 (7)
- January 2009 (32)
- December 2008 (19)
- November 2008 (12)
- October 2008 (15)
- September 2008 (14)
- August 2008 (15)
- July 2008 (18)
- June 2008 (20)
- May 2008 (19)
- April 2008 (27)
- March 2008 (22)
- February 2008 (21)
- January 2008 (15)
- December 2007 (22)
- November 2007 (17)
- October 2007 (29)
- September 2007 (31)
- August 2007 (34)
- July 2007 (31)
- June 2007 (36)
- May 2007 (23)
- April 2007 (22)
- March 2007 (30)
- February 2007 (50)
- January 2007 (75)
- December 2006 (48)
- November 2006 (59)
- October 2006 (89)
- September 2006 (29)
- August 2006 (48)
- July 2006 (14)
- June 2006 (35)
- May 2006 (62)
- April 2006 (63)
- March 2006 (72)
- February 2006 (83)
- January 2006 (56)
- December 2005 (46)
- November 2005 (60)
- October 2005 (27)
- September 2005 (54)
- August 2005 (83)
Tags
Categories
- Anime (29)
- AWS (426)
- BBS (22)
- Blog (249)
- Book (31)
- Bridge (1)
- Browser (527)
- Cassandra (2)
- CDN (184)
- Cloud (569)
- CMS (58)
- Comic (19)
- Computer (4,465)
- Computer and Network Center (33)
- CSS (69)
- Database (421)
- DNS (128)
- Editor (25)
- Financial (89)
- Firefox (232)
- Food (14)
- FreeBSD (139)
- FTP (3)
- Game (61)
- GCP (4)
- Go (1)
- GoogleChrome (167)
- Hardware (385)
- IE (98)
- Joke (159)
- Lab (4)
- Library (4)
- Linux (175)
- MacOS (30)
- Mail (120)
- MariaDB (19)
- Movie (35)
- Murmuring (4,605)
- Music (49)
- MySQL (295)
- NCTU (65)
- NetBSD (7)
- Network (3,258)
- OpenBSD (8)
- Opera (26)
- OS (391)
- P2P (132)
- Photo (72)
- Political (124)
- PostgreSQL (41)
- Privacy (9)
- Programming (769)
- Recreation (534)
- RSS (79)
- Safari (38)
- Science (99)
- Search Engine (181)
- Security (1,053)
- Service (49)
- SMS (10)
- Social (218)
- Software (2,362)
- Spam (108)
- Sport (10)
- Telephone (117)
- Television (62)
- Usenet (15)
- Vim (14)
- VPN (19)
- Wiki (48)
- Windows (74)
- WWW (1,626)
Blogroll
Meta
Tag Archives: performance
Quotient filter
之前有提過「Cuckoo Filter：比 Bloom Filter 多了 Delete」，最近在「A general purpose counting filter: making every bit count」這邊看到 Quotient filter，也是類似 Bloom filter 的資料結構，但想要解決更多問題。 一般的 Bloom filter (BF) 會有這些問題： The inability to delete items Poor scaling out of RAM The inability to resize dynamically The … Continue reading
DigitalOcean 推出 High CPU Droplets
DigitalOcean 推出了高 CPU 使用量的 VPS 服務：「Introducing High CPU Droplets」。 價位從 USD$40/month 開始。相較於 Standard Droplet，CPU 的速度提升到 2.5 倍速 (平均)： The current CPUs powering High CPU Droplets are the Intel Broadwell 2697Av4 with a clock speed of 2.6Ghz, and the Intel Skylake … Continue reading
Posted in Computer, Hardware, Murmuring, Network, Service
Tagged cpu, digitalocean, droplet, high, hosting, performance, speed, vps
Leave a comment
小台機器上的 innodb_purge_threads 對效能的影響
雖然「MyISAM, small servers and sysbench at low concurrency」這篇標題是在講 MySQL 上的 MyISAM，但還是有提到一些 InnoDB 的東西... 其中提到了 innodb_purge_threads 對效能的影響： the default value for innodb_purge_threads, which is 4, can cause too much mutex contention and a loss in QPS on small servers. For sysbench … Continue reading
在 Production 上惡搞 PostgreSQL，建立一千萬筆 Index...
在「Running 10 Million PostgreSQL Indexes In Production (And Counting)」這邊看到 Heap 他們在 PostgreSQL 上惡搞的方式。 為了效能，所以在上面建立了大量的 Partial index。像是這樣的條件 (所以其實每個都不算太大)： CREATE INDEX ON events (time) WHERE type = ‘click’ AND text = ‘login’ 要注意的是 MySQL 的 Partial index 其實是 prefix index，用以減少 index 的空間要求，這在 … Continue reading
Posted in Computer, Database, Murmuring, PostgreSQL, Software
Tagged database, db, index, mysql, partial, performance, postgresql, rdbms, speed
Leave a comment
用 Machine Learning 調校資料庫
AWS AI Blog 在月初上放出來的消息：「Tuning Your DBMS Automatically with Machine Learning」。 Carnegie Mellon Database Group 做的研究，除了預設值以外，另外跟四種不同的參數做比較，分別是 OtterTune (也就是這次的研究)、Tuning script (對於不熟資料庫的人，常用的 open source 工具)、DBA 手動調整，以及 RDS： MySQL： PostgreSQL： 比較明顯的結論是： Default 值在所有的 case 下都是最差的 (無論是 MySQL 與 PostgreSQL 平台，以及包括 99% 的 Latency 與 … Continue reading
Posted in AWS, Cloud, Computer, Database, Murmuring, MySQL, Network, PostgreSQL, Programming, Service, Software
Tagged amazon, aws, carnegie, case, cloud, cmu, database, db, dba, dbms, default, latency, learning, machine, mellon, mysql, ottertune, performance, postgresql, qps, rdbms, rds, research, script, test, testing, throughput, tuning
Leave a comment
MySQL 5.6/5.7/5.8 在 single thread 時的效能比較
作者 Mark Callaghan 在討論 MySQL 在量小的時候，新版可能反而比舊版慢：「The history of low-concurrency performance regressions in MySQL 5.6, 5.7 and 5.8」。 One example of the regression occurs for in-memory sysbench where MySQL 5.7 and MySQL 8 get 60% to 70% of the QPS compared … Continue reading
performance_schema 的簡易用法
Mark Callaghan 寫了篇關於 performance_schema 的用法 (很短)，讓大家先把這個參數開習慣，雖是入門推廣班：「Short guide on using performance_schema for user & table stats」。 他推薦的兩個資訊是： select * from table_io_waits_summary_by_table select * from events_statements_summary_by_account_by_event_name 當使用 5.7+ 時，可以考慮這兩個： SELECT * FROM sys.schema_table_statistics SELECT * FROM sys.user_summary 簡單到不行，但卻可以幫不少忙... 很棒的入門推廣班 XDDD
Symfony 4 將放棄 HHVM
PHP 7.x 的效能已經趕上 HHVM (甚至在某些項目超越，參考下面的連結)，這使得後來大家為了相容性與擴充性的考量，HHVM 的社群一直沒有成長 (參考「PHP Versions Stats - 2017.1 Edition」這邊，作者從 packagist.org 上得到的數據)： PTS: PHP 7.1 vs. PHP 7.0 vs. HHVM Benchmarks HHVM vs PHP 7 – The Competition Gets Closer! 這使得 Symfony 決定在 Twitter 上蒐集意見，而後決定下一個 major version (4) … Continue reading
iOS 上 Spotlight Search 如果很慢...
看到「Disabling Slack Indexing Seems to Improve Spotlight Performance on iOS」這篇講 iOS 的 Spotlight Search 很慢的問題，作者後來找到與 Slack 有關？到這邊關掉就正常了： Settings > General > Spotlight Search > toggle Slack off 我是沒遇到... 不過先寫起來，如果以後遇到時應該會有幫助。
Posted in Computer, Murmuring, Network, Search Engine, Software
Tagged apple, index, ios, iphone, performance, search, slack, speed, spotlight
Leave a comment
zetcd：用 etcd 跑 ZooKeeper 架構
在「zetcd: running ZooKeeper apps without ZooKeeper」這邊介紹了用 etcd 當作 ZooKeeper 伺服器。程式碼在「Serve the Apache Zookeeper API but back it with an etcd cluster」這邊可以看到。 不過可以看到有不少 overhead： 但 etcd 用 Go 寫 (省下 JVM tuning？)，可能是個不錯的誘因...