ErrorOops something went wrong! Either the site is offline or an unhandled error occurred. We apologize and have logged the error. Please try your request again or if you know who your site administrator is let them know too.
memory-configuration-scalability-blog-series
frankdenneman Follow Frank Denneman is the Chief Technologist for AI at VMware by Broadcom. He is an author of the vSphere host and clustering deep dive series, as well as a podcast host for the Unexplored Territory podcast. You can follow him on Twitter @frankdenneman Processor speed and core counts are important factors when designing a new server platform. However with virtualization platforms,
How to receive a million packets per second
Last week during a casual conversation I overheard a colleague saying: "The Linux network stack is slow! You can't expect it to do more than 50 thousand packets per second per core!" That got me thinking. While I agree that 50kpps per core is probably the limit for any practical application, what is the Linux networking stack capable of? Let's rephrase that to make it more fun: On Linux, how hard
JNuma Library
1. Silk Weaver: A Scalable Data Processing Platform JNuma library ï¬ Java/Scalaã§NUMA aware アクセスを実ç¾ã™ã‚‹ãŸã‚ã®ãƒ©ã‚¤ãƒ–ラリ ï¬ 2012å¹´11月ã«ä½œæˆãƒ»å…¬é–‹æ¸ˆã¿ ï¬ https://github.com/xerial/jnuma ï¬ Javaã‹ã‚‰JNI経由ã§NUMA APIを呼ã³å‡ºã™ ï¬ æ©Ÿèƒ½ ï¬ ç‰¹å®šã®NUMA node （メモリ）ã§ãƒãƒƒãƒ•ã‚¡ã‚’ç¢ºä¿ ï¬ GCã®ç®¡ç†å¯¾è±¡å¤–é ˜åŸŸãªã®ã§ã€Numa.freeを実行ã—ãŸçž¬é–“ã«è§£æ”¾ã•ã‚Œã‚‹ (JVMãŒãƒ¡ãƒ¢ãƒªã‚’食ã„å°½ãã™ã®ã‚’防ã） ï¬ Threadを特定ã®CPU(s)ã«å›ºå®šã™ã‚‹ ï¬ Javaã«ã¯ãªã„機能 1 2. Silk Weaver: A Scalable Data Processing Platform Distance and CPU affinity o
Riak, Elasticsearch and numad walk into a Red Hat
What starts like a joke is really a serious post about large distributed systems. This is what we run at meltwater - and while there is a lot of fun designing and administering these systems, we run into odd problems from time to time. In this case: some well-meant Red Hat memory optimisations interfering with Riak performance. There is a lot of talk around “Big Data†in the industry - and here at
Intel，Haswell世代ã®ã‚µãƒ¼ãƒãƒ¼å‘ã‘æ–°CPU「Xeon E5 v3ã€ãƒ•ã‚¡ãƒŸãƒªãƒ¼ã‚’発表。最大CPUコア数ã¯18基ã«
Intel，Haswell世代ã®ã‚µãƒ¼ãƒãƒ¼å‘ã‘æ–°CPU「Xeon E5 v3ã€ãƒ•ã‚¡ãƒŸãƒªãƒ¼ã‚’発表。最大CPUコア数ã¯18基㫠ライター：塩田紳二 Intelã®è³‡æ–™ã«ã‚ã£ãŸXeon E5 v3ã®ãƒ€ã‚¤å†™çœŸ 2014å¹´9月9æ—¥1：00，Intelã¯ï¼ŒHaswell世代ã®ãƒžã‚¤ã‚¯ãƒã‚¢ãƒ¼ã‚テクãƒãƒ£ã‚’採用ã™ã‚‹ã‚µãƒ¼ãƒãƒ¼ï¼†ãƒ¯ãƒ¼ã‚¯ã‚¹ãƒ†ãƒ¼ã‚·ãƒ§ãƒ³å‘ã‘CPU「Xeon E5 v3ã€ãƒ•ã‚¡ãƒŸãƒªãƒ¼è¨ˆ32製å“を発表ã—ãŸã€‚開発コードãƒãƒ¼ãƒ 「Haswell-EPã€ã¨ã‚‚呼ã°ã‚Œã¦ã„ãŸã‚‚ã®ã§ï¼Œå†…蔵ã™ã‚‹CPUコア数ãŒæœ€å¤§18基ã«é”ã™ã‚‹ã»ã‹ï¼ŒDDR4メモリã«å¯¾å¿œã™ã‚‹ã¨ã„ã£ãŸç‰¹å¾´ã‚’å‚™ãˆã‚‹ã€‚ Xeonブランドã®CPUã¯ï¼Œã‚²ãƒ¼ãƒžãƒ¼ãŒç©æ¥µçš„ã«é¸ã¶ã‚ˆã†ãªè£½å“ã§ã¯ãªã„ãŒï¼Œ2014å¹´8月30æ—¥ã«ç™ºè¡¨ã•ã‚ŒãŸãƒã‚¤ã‚¨ãƒ³ãƒ‰ãƒ‡ã‚¹ã‚¯ãƒˆãƒƒãƒ—PC用CPU「Core i7-5000番å°ã€ï¼ˆé–‹ç™ºã‚³ãƒ¼ãƒ‰ãƒãƒ¼ãƒ Haswell-E，以下åŒï¼‰ã¨å…±é€šã™ã‚‹éƒ¨åˆ†ã‚‚ã‚り，ã“ã“ã§ã¯æ¦‚è¦ã‚’ã¾
DMM inside
ãªã›ã‚™DMMã‹ã‚™web3ã«å‚å…¥ã—ãŸã®ã‹ã€‚Seamoon Protocolã‹ã‚™ç›®æŒ‡ã™æ–°ãŸãªã‚¨ãƒ³ã‚¿ãƒ¡ä½“験ã®æœªæ¥ã¨ã¯
Production Notes for Self-Managed Deployments - MongoDB Manual v8.0
This page details system configurations that affect MongoDB, especially when running in production. MongoDB Atlas is a cloud-hosted database-as-a-service. MongoDB Cloud Manager, a hosted service, and Ops Manager, an on-premise solution, provide monitoring, backup, and automation of MongoDB instances. For documentation, see Atlas documentation, the MongoDB Cloud Manager documentation and Ops Manage
Kvm performance optimization for ubuntu
This document discusses various techniques for optimizing KVM performance on Linux systems. It covers CPU and memory optimization through techniques like vCPU pinning, NUMA affinity, transparent huge pages, KSM, and virtio_balloon. For networking, it discusses vhost-net, interrupt handling using MSI/MSI-X, and NAPI. It also covers block device optimization through I/O scheduling, cache mode, and a
Linux packet-forwarding
2. ã¯ã˜ã‚ã¾ã—ã¦ï¼Ÿ • 新潟県三æ¡å¸‚ã«ã‚る地域 ISP 有é™ä¼šç¤¾éŠ€åº§å ‚ã§ã‚µãƒ¼ãƒç®¡ç†ã‚„ãƒãƒƒãƒˆ ワーク管ç†ã‚’ã—ã¦ã„ã¾ã™ • 弊社ã§ã¯ä¸Šæµ ISP(OCN ã•ã‚“ã¨æ–°æ½ŸçœŒå†…ã® ISP) ã‚„ Echigo-IX(新潟県 内ã®åœ°åŸŸ IX) ã¨ã®æŽ¥ç¶šã« Linux(Vyatta Core) を使ã£ã¦ BGP 接続を ã—ã¦ã„ã¾ã™ • 2010 å¹´é ƒã‹ã‚‰ Linux カーãƒãƒ«ã®ãƒ‘ケット転é€å‡¦ç†ã«ã¤ã„ã¦èª¿ã¹ InternetWeek ã§ç™ºè¡¨ã—ãŸã‚Šã—ã¦ã„ã¾ã—㟠• 趣味ã§ã¯ IETF ã§æ¨™æº–化ãŒé€²ã‚られã¦ã„ã‚‹ MAP(Mapping of Address and Port with Encapsulation) ã¨ã„ã†ãƒ—ãƒãƒˆã‚³ãƒ«ã‚’ Linux カーãƒãƒ«ã«å®Ÿè£…ã— Vyatta Core ã«çµ„ã¿è¾¼ã¿ ASAMAP ã¨ã„ã†åå‰ã§ オープンソースã§å…¬é–‹ã—ãŸã‚Šã—ã¦ã„ã¾ã™ 2 3. ASAMAP？ • IPv4 over
サーãƒãƒ¼ã•ã‚“ã«æœ¬æ°—を出ã—ã¦ã‚‚らã†ãŸã‚ã«æ†¶ãˆã¦ãŠããŸã„è¨å®šé …ç›®
cpuspeed ãŒã‚ªãƒ³ã ã¨.... — ã¯ã›ãŒãƒ¯ãƒ³ (@hasegaw) 2014, 5月 29 ミドルウェアã®ã‚¹ãƒ«ãƒ¼ãƒ—ットを測ã‚ã†ã¨æ€ã£ãŸã®ã§ã™ãŒ cpuspeed ãªã©ã®è¨å®šã‚’ãœã‚“ãœã‚“ã‚„ã£ã¦ã„ã¾ã›ã‚“ã§ã—ãŸã€‚。。 経験上ã€ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°éŽç¨‹ã§ã„ã˜ã‚ŠãŸããªã‚‹ã‚ˆã†ãªãƒ‘ラメータをæ€ã„出ã—ã¦ã¿ã¾ã™ã€‚ パワーマãƒã‚¸ãƒ¡ãƒ³ãƒˆã«é–¢ã™ã‚‹è¨å®šã¯ã‚ªãƒ•ã«ã™ã‚‹ UEFIã‚„BIOSã«ã¯ãƒ‘ワーマãƒã‚¸ãƒ¡ãƒ³ãƒˆè¨å®šãŒã‚ã‚Šã¾ã™ãŒã€ã“れらを無効ã«ã™ã‚‹ã¨ãƒ—ãƒã‚»ãƒƒã‚µãªã©ãŒç„¡æ¡ä»¶ã§å®šæ ¼ã‚¯ãƒãƒƒã‚¯ã§èµ°ã‚Šç¶šã‘ã¾ã™ã€‚ピーク性能を高ã‚ãŸã‚Šçž¬ç™ºåŠ›ã‚’上ã’ã‚‹ãŸã‚ã«ã¯ãƒ‘ワーマãƒã‚¸ãƒ¡ãƒ³ãƒˆã¯ã‚ªãƒ•ã«ã—ã¾ã™ã€‚当然ãªãŒã‚‰ãƒ™ãƒ¼ã‚¹ã®æ¶ˆè²»é›»åŠ›ã‚„ファンã®é¨’音ã¯å¢—ãˆã¾ã™ãŒã€ã‹ã‚ã‚Šã«ã„ãらã‹ãƒ”ーク性能ã®å‘上ãŒè¦‹è¾¼ã‚ã¾ã™ã€‚ Hyper Threading ã¯ãƒ¬ã‚¤ãƒ†ãƒ³ã‚·ãƒ¼ã¨ã‚¹ãƒ«ãƒ¼ãƒ—ットã®ãƒˆãƒ¬ãƒ¼ãƒ‰ã‚ªãƒ• Hyper Threadingã¯ã€ãŸã¶ã‚“ã€ã‚³ã‚¢å†…ã§ãƒ‘イプラインをå–ã‚Šåˆã†ã‹ã‚‰
NUMAアーã‚テクãƒãƒ£ã¨ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯æ€§èƒ½ï¼ˆï¼’）ï¼å®Ÿè·µç·¨ - Blog
å‰å›žã€NUMAアーã‚テクãƒãƒ£ãŒãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯æ€§èƒ½ã«ä¸Žãˆã‚‹å½±éŸ¿ã«ã¤ã„ã¦ã®æŠ€è¡“çš„ãªèƒŒæ™¯ã‚’説明ã—ãŸã€‚ (å‰å›žè¨˜äº‹) NUMAアーã‚テクãƒãƒ£ã¨ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯æ€§èƒ½ï¼ˆï¼‘）ï¼èª¬æ˜Žç·¨ - Interconnect Your Future - InfiniBand and RoCE 本稿ã§ã¯ã€å®Ÿéš›ã«Linux環境ã§ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯æ€§èƒ½ã‚’測定ã—ã€NUMAãŒä¸Žãˆã‚‹å½±éŸ¿ã«ã¤ã„ã¦ã®äº‹ä¾‹ã¨ãƒŽã‚¦ãƒã‚¦ã‚’紹介ã—ãŸã„。 本稿ã§ã¯ã€æœ€çµ‚çš„ã«ã¯ä¸‹è¨˜æ¸¬å®šä¾‹ã‚’ã¾ã¨ã‚ã¦ã„る。NUMA構æˆãŒç•°ãªã‚‹å ´åˆã€æ€§èƒ½ãŒå…¨ãç•°ãªã£ã¦ã—ã¾ã†ã€‚極端ãªè©±ã€40GbE NICãªã®ã«4-5Gbpsã—ã‹å‡ºãªã„ã“ã¨ã«ãªã£ã¦ã—ã¾ã†ã€‚実際ã«ã¯æ€§èƒ½å‡ºã¾ã™ã€‚注æ„。 NUMA構æˆä¸Šé ã„構æˆã§ã®iperf性能：1プãƒã‚»ã‚¹(1コア) 2.00Gbpsã€12プãƒã‚»ã‚¹(6コア) 4.84Gbps NUMA構æˆä¸Šè¿‘ã„構æˆã§ã®iperf性能：1プãƒã‚»ã‚¹(1コア) 14.7Gbpsã€12プãƒã‚»
NUMAアーã‚テクãƒãƒ£ã¨ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯æ€§èƒ½ï¼ˆï¼‘）ï¼èª¬æ˜Žç·¨ - Blog
NUMA(Non-Uniform Memory Access)ã¨ã¯ Non-uniform memory access - Wikipedia, the free encyclopedia NUMA - Wikipedia NUMA（Non-Uniform Memory Accessã€ãƒŒãƒžï¼‰ã¨ã¯ã€å…±æœ‰ãƒ¡ãƒ¢ãƒªåž‹ãƒžãƒ«ãƒãƒ—ãƒã‚»ãƒƒã‚µã‚³ãƒ³ãƒ”ュータシステムã®ã‚¢ãƒ¼ã‚テクãƒãƒ£ã®ã²ã¨ã¤ã§ã€è¤‡æ•°ãƒ—ãƒã‚»ãƒƒã‚µãŒå…±æœ‰ã™ã‚‹ãƒ¡ã‚¤ãƒ³ãƒ¡ãƒ¢ãƒªã¸ã®ã‚¢ã‚¯ã‚»ã‚¹ã‚³ã‚¹ãƒˆãŒã€ãƒ¡ãƒ¢ãƒªé ˜åŸŸã¨ãƒ—ãƒã‚»ãƒƒã‚µã«ä¾å˜ã—ã¦å‡ä¸€ã§ãªã„アーã‚テクãƒãƒ£ã§ã‚る。 複数CPUãŒã‚ã‚‹å ´åˆã€è¿‘ã„メモリã¨é ã„メモリãŒæ§‹æˆä¸Šå˜åœ¨ã—ã€ãƒ‡ãƒ¼ã‚¿ãŒç½®ã‹ã‚Œã‚‹å ´æ‰€ã«ã‚ˆã£ã¦ã€æ€§èƒ½ã«å·®ãŒå‡ºã¦ã—ã¾ã†ã‚¢ãƒ¼ã‚テクãƒãƒ£ã§ã‚る。 ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯æ€§èƒ½ã‚’測定ã™ã‚‹å ´åˆã‚‚ã€NUMA構æˆä¸Šä¸åˆ©ãªçŠ¶æ…‹ã§æ€§èƒ½æ¸¬å®šã‚’ã—ã¦ã—ã¾ã†ã¨ã€æ€ã†ã‚ˆã†ãªæ€§èƒ½ãŒå‡ºãªã„å ´åˆãŒã‚ã‚‹ãŸã‚ã€æ³¨æ„ãŒå¿…è¦ã§ã‚る。本稿ã§ã¯ã€NUMAåŠã³CPU
High Performance I/O with NUMA Systems in Linux
Fusion-io Confidential—Copyright © 2013 Fusion-io, Inc. All rights reserved. Fusion-io Confidential—Copyright © 2013 Fusion-io, Inc. All rights reserved. High Performance I/O with NUMA Systems in Linux Lance Shelton How prevalent is NUMA? ▸ All major vendors • HP, Dell, IBM, Cisco, SuperMicro, Hitachi, etc. ▸ As small as 1U ▸ 2, 4, and 8 socket systems ▸ 2 to 10 cores per socket ▸ Number of cores
qpstudy 2014.04 ãƒãƒ¼ãƒ‰ã‚¦ã‚§ã‚¢è¨è¨ˆã®å‹˜æ‰€
1. 第3セッション ãƒãƒ¼ãƒ‰ã‚¦ã‚§ã‚¢è¨è¨ˆã®å‹˜æ‰€ qpstudy 2014.04 -険ã—ã„é“ã®æ©ã方をã¡ã‚‡ã£ã¨ã ã‘æ•™ãˆã¦ã‚ã’ã‚‹- Apr 19, 2014 @ DWANGO セミナールームTakeshi HASEGAWA (@hasegaw) 2. @hasegaw is 誰 é•·è°·å·â€ƒçŒ› (HASEGAWA Takeshi) twitter: @hasegaw å‰è·æ™‚代 ・SEã¨ã—ã¦ã‚·ã‚¹ãƒ†ãƒ 構築ã€å®¢å…ˆã®ã‚·ã‚¹ãƒ†ãƒ é‹ç”¨ã€æ案 ・気付ã„ãŸã‚‰ãƒ—リセールス∼PM担当SE （ã–ã£ãりデザインã€å·¥æ•°ï¼å°Žå…¥ç‰©å“見ç©ã‚‚り〠構築プãƒã‚¸ã‚§ã‚¯ãƒˆã®ç®¡ç†ã€ä¿å®ˆç‰ã®å•ã„åˆã‚ã›å¯¾å¿œï¼‰ ç¾è· ・フラッシュを軸ã¨ã—ãŸã‚¢ãƒ—リケーション高速化を支æ´ã™ã‚‹ セールスエンジニア
Gc and-pagescan-attacks-by-linux
GC-stall and Page Scan Attacks by Linux The document discusses GC attacks and page scan attacks by the Linux kernel that can cause long GC pauses. It provides examples of GC attacks due to I/O starvation and memory starvation. Solutions proposed include configuring Linux to flush disk writes more frequently and protect JVM heap memory from swapping. Page scan attacks are caused by Transparent Huge
Dual Socket 㪠PC ã®å‰²ã‚Šè¾¼ã¿ã«ä»˜ã„ã¦
デュアルソケットã€ã¤ã¾ã‚ŠCPUを２個æ載ã™ã‚‹ãƒžã‚¶ãƒ¼ãƒœãƒ¼ãƒ‰é™å®šã§è‡ªä½œPCを楽ã—ã‚€ç†è€…ã®æ§˜åã‚’æ™’ã—ã¦ã‚†ãã¾ã™ã€‚ 今回ã¯ã€PCã®å‰²ã‚Šè¾¼ã¿ï¼ˆInterrupt）処ç†ã«ä»˜ã„ã¦ã€åŸºæœ¬çš„ãªæ‰€ã‹ã‚‰ Dual Socket ã§ã®å‰²ã‚Šè¾¼ã¿å‡¦ç†ã¾ã§ã‚’ç°¡å˜ã«ã¾ã¨ã‚ã¦ã¿ã¾ã™ã€‚（基本的ã«ã¯è‡ªåˆ†ç”¨ãƒ¡ãƒ¢ã§ã€ç¾åœ¨ç·¨é›†ä¸ã§ã™ï¼‰ ç†è€…ã®çŸ¥ã‚‹å‰²ã‚Šè¾¼ã¿å‡¦ç†ã«ã¯å¤§ãã分ã‘ã¦ï¼“タイプ（ソフト割り込ã¿ã€ãƒãƒ¼ãƒ‰å‰²ã‚Šè¾¼ã¿ã€NMI(強制割り込ã¿)）ãŒæœ‰ã‚Šã¾ã™ãŒã€ä»Šå›žã®è¨˜äº‹ã®å¯¾è±¡ã¯I/Oデãƒã‚¤ã‚¹ã®ãƒãƒ¼ãƒ‰å‰²ã‚Šè¾¼ã¿ã«ä»˜ã„ã¦ã§ã™ã€‚余談ã§ã™ãŒã‚½ãƒ•ãƒˆå‰²ã‚Šè¾¼ã¿ã¯ä¾‹å¤–処ç†ã¨ã‚‚言ã‚れ代表的ãªã‚‚ã®ã«NullPointerExceptionãªã©ãŒæœ‰ã‚Šã¾ã™ã€‚ 発端㯠先日ã®è¨˜äº‹ ã§ã™ã€‚ 例ãˆã°ä¸‹å›³ã®æ§˜ãªNUMA構æˆã§CPU1ã§å‹•ä½œã—ã¦ã„るアプリãŒCPU2ã®I/Oãƒã‚¹ã«æŽ¥ç¶šã•ã‚ŒãŸãƒ‡ãƒã‚¤ã‚¹ã‚’利用ã™ã‚‹å ´åˆã€å‰²ã‚Šè¾¼ã¿ã®çµŒè·¯ã‚„ドライãƒã¨ã®é–¢ä¿‚ã«ä»˜ã„ã¦è©³ã—ã調ã¹ã‚ˆã†ã¨æ€ã£ãŸè¨³ã§
リリースã€éšœå®³æƒ…å ±ãªã©ã®ã‚µãƒ¼ãƒ“スã®ãŠçŸ¥ã‚‰ã›
最新ã®äººæ°—エントリーã®é…ä¿¡
j次ã®ãƒ–ックマーク
kå‰ã®ãƒ–ックマーク
lã‚ã¨ã§èªã‚€
eコメント一覧を開ã
oページを開ã
Intel Xeon E5 Version 3: Up to 18 Haswell EP Cores
lanceshelton/irqstat · GitHub - A better way to watch /proc/interrupts, designed for NUMA systems with many processors.
» Of Penguins, Rabbits and Buses Brain Matters
{{#tags}}- {{label}}
{{/tags}}