From owner-freebsd-sparc Thu Dec 5 21:16:18 2002 Delivered-To: freebsd-sparc@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 4942537B401 for ; Thu, 5 Dec 2002 21:16:17 -0800 (PST) Received: from corbulon.video-collage.com (corbulon.video-collage.com [64.35.99.179]) by mx1.FreeBSD.org (Postfix) with ESMTP id A735E43E9C for ; Thu, 5 Dec 2002 21:16:16 -0800 (PST) (envelope-from mi@corbulon.video-collage.com) Received: from corbulon.video-collage.com (localhost.video-collage.com [127.0.0.1]) by corbulon.video-collage.com (8.12.6/8.12.6) with ESMTP id gB65G3jJ047317 (version=TLSv1/SSLv3 cipher=EDH-RSA-DES-CBC3-SHA bits=168 verify=NO) for ; Fri, 6 Dec 2002 00:16:04 -0500 (EST) (envelope-from mi@corbulon.video-collage.com) Received: (from mi@localhost) by corbulon.video-collage.com (8.12.6/8.12.6/Submit) id gB65G3qK047316 for sparc@FreeBSD.org; Fri, 6 Dec 2002 00:16:03 -0500 (EST) (envelope-from mi) From: Mikhail Teterin Message-Id: <200212060516.gB65G3qK047316@corbulon.video-collage.com> Subject: why would Sparc be soo sloow? To: sparc@FreeBSD.org Date: Fri, 6 Dec 2002 00:16:02 -0500 (EST) X-Mailer: ELM [version 2.4ME+ PL92b (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII X-Scanned-By: MIMEDefang 2.21 (www . roaringpenguin . com / mimedefang) Sender: owner-freebsd-sparc@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Hello! This is only marginally FreeBSD related... Sorry... I'm puzzled by the poor performance of our SparcIII @900MHz. On our particular, heavily number crunching little application, the best (using Sun's own compiler with -xarch=v9b -xO5, etc.) it can show off is being about 7.5 times slower than my desktop Pentium4 @2GHz. The Sparc machine is running a recent Solaris (the box purchased about 9 months ago). My desktop runs FreeBSD-current. First observed with pvm, the slowness was confirmed by a simple standalone application, that read the input (through scanf()) one line at a time from the text file. The SpecBench score of this machine (Sunfire 280) is about 20-30% lower than that of mine (Dell Dimension)... I'd accept the factor of 3 or 4, but 7.5 is casting doubts over my measurements, yet the results are quite repeatable... Another mistery is a high "system" component of the time (builtin(1)) output -- on Solaris over half the time is spent in "kernel". Can it be trusted? On FreeBSD, very little is spent in kernel -- as expected. (FWIW, the older Pentium2 @450MHz is only 2.5 times slower than the P4 @2GHz. Still over 3 times faster than the SparcIII) Any hints? Thanks! -mi To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-sparc" in the body of the message