From owner-freebsd-java@FreeBSD.ORG Tue May 27 13:50:57 2003 Return-Path: Delivered-To: freebsd-java@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 878CB37B401 for ; Tue, 27 May 2003 13:50:57 -0700 (PDT) Received: from aromo.spock.cl (aromo.spock.cl [200.27.125.98]) by mx1.FreeBSD.org (Postfix) with ESMTP id 091B143FB1 for ; Tue, 27 May 2003 13:50:54 -0700 (PDT) (envelope-from roberto@spock.cl) Received: from spock.cl (multia.spock.cl [200.27.125.107]) by aromo.spock.cl (8.12.9/8.12.8) with ESMTP id h4RKosqd062354 for ; Tue, 27 May 2003 16:50:55 -0400 (CLT) Message-ID: <3ED3CFBE.5050805@spock.cl> Date: Tue, 27 May 2003 16:51:10 -0400 From: Roberto de Iriarte User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.3) Gecko/20030418 X-Accept-Language: en-us, en MIME-Version: 1.0 To: freebsd-java@freebsd.org Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Subject: jdk on alpha X-BeenThere: freebsd-java@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Porting Java to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 27 May 2003 20:50:57 -0000 Two comments on the Java on alpha thing. a) Java is slow on the Alpha, with the current implementations (i.e Linux-Java for AXP) a 500 Mhz 21164 is about as fast as a 167 Mhz UltraSparc I. b) One could think about completing some work on an advanced JIT for the alpha based either on the kaffe JIT work or on the CACAO research project (see http://www.complang.tuwien.ac.at/java/cacao/) I emailed Andreas Krall a while ago, he can make the sources of his work availible, i recevied a copy of them but they would require a LOT of work to be made useful. Some personal thoughts on this subject. When i first became interested in the alpha platform some five years ago, it was not in the state of neglect that it is today. Back then, you could (with a thich enough wallet) purchase very interesting systems that had very little equal performance-wise (see, per example http://www.specbench.org/cpu95/results/res99q1/ , back then Alphas would outperform everything else in SpecINT by at least one third, and by a factor of two on SpecFP ...) Nowadays, i could not justify the purchase of a new alpha unless i had some application that cannot be ported to another platform. (Yes, i know old alphas are very nice machines of tremendous historical relevance, but the supply of newer used stock is getting scarce) BTW, i am no stranger to Alpha, i own a small collection: AS2100 4/200, AS2100A 5/375, AS1000 4/200, AS800, Digital Server 3305, Alphastation XP1000 6/500, Alphastation 250/4-300, Alphastation 200/4-166, and several Multias, just to show my appreciation of the architecture, and yet i am typing this on a PeeCee. IMHO, a project to develop a 64bit JVM that exploits the AMD64 platform would have a much brighter future. Even from this far (Chile) i have already placed an order for the building blocks to assembly such a system. (Yes, i know the amd64 port is not complete yet....) Comments? Regards, and thanks to all for the excellent work! Roberto de Iriarte Santiago, Chile