Return-Path: X-Original-To: apmail-hadoop-common-issues-archive@minotaur.apache.org Delivered-To: apmail-hadoop-common-issues-archive@minotaur.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 7960817F02 for ; Mon, 29 Sep 2014 19:24:36 +0000 (UTC) Received: (qmail 18407 invoked by uid 500); 29 Sep 2014 19:24:35 -0000 Delivered-To: apmail-hadoop-common-issues-archive@hadoop.apache.org Received: (qmail 18362 invoked by uid 500); 29 Sep 2014 19:24:35 -0000 Mailing-List: contact common-issues-help@hadoop.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: common-issues@hadoop.apache.org Delivered-To: mailing list common-issues@hadoop.apache.org Received: (qmail 18348 invoked by uid 99); 29 Sep 2014 19:24:35 -0000 Received: from arcas.apache.org (HELO arcas.apache.org) (140.211.11.28) by apache.org (qpsmtpd/0.29) with ESMTP; Mon, 29 Sep 2014 19:24:35 +0000 Date: Mon, 29 Sep 2014 19:24:34 +0000 (UTC) From: "Jeff Buell (JIRA)" To: common-issues@hadoop.apache.org Message-ID: In-Reply-To: References: Subject: [jira] [Commented] (HADOOP-11152) Better random number generator MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-JIRA-FingerPrint: 30527f35849b9dde25b450d4833f0394 [ https://issues.apache.org/jira/browse/HADOOP-11152?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14152122#comment-14152122 ] Jeff Buell commented on HADOOP-11152: ------------------------------------- Others have investigated skew due to random selection, i.e. http://cs.brown.edu/~rfonseca/pubs/ferguson-atc10-poster.pdf. They got good results from a round-robin scheme. Need to make sure such a scheme works well in a multi-threaded environment. In any case, what is needed is a scheme that actively tries to uniformly spread the copies rather than relying on a completely random process. > Better random number generator > ------------------------------ > > Key: HADOOP-11152 > URL: https://issues.apache.org/jira/browse/HADOOP-11152 > Project: Hadoop Common > Issue Type: Improvement > Reporter: Luke Lu > Labels: newbie++ > > HDFS-7122 showed that naive ThreadLocal usage of simple LCG based j.u.Random creates unacceptable distribution of random numbers for block placement. Similarly, ThreadLocalRandom in java 7 (same static thread local with synchronized methods overridden) has the same problem. > "Better" is defined as better quality and faster than j.u.Random (which is already much faster (20x) than SecureRandom). > People (e.g. Numerical Recipes) have shown that by combining LCG and XORShift we can have a better fast RNG. It'd be worthwhile to investigate a thread local version of these "better" RNG. -- This message was sent by Atlassian JIRA (v6.3.4#6332)