From owner-freebsd-smp Mon Jun 21 13: 9:22 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ares.cs.Virginia.EDU (ares.cs.Virginia.EDU [128.143.136.41]) by hub.freebsd.org (Postfix) with ESMTP id 4D57914F99 for ; Mon, 21 Jun 1999 13:09:18 -0700 (PDT) (envelope-from chapin@cs.virginia.edu) Received: from cs.virginia.edu (mamba.cs.Virginia.EDU [128.143.137.15]) by ares.cs.Virginia.EDU (8.9.2/8.9.2/UVACS-1999030200) with ESMTP id QAA21524; Mon, 21 Jun 1999 16:08:35 -0400 (EDT) Message-Id: <199906212008.QAA21524@ares.cs.Virginia.EDU> Reply-To: chapin@cs.virginia.edu From: chapin@cs.virginia.edu (Steve Chapin) To: Richard Cownie Cc: Kedar Rajadnya , freebsd-smp@freebsd.org Zevon-of-the-day: Stand in the Fire Subject: Re: SMP, 4GB RAM, 4x CPU In-reply-to: Your message of Mon, 21 Jun 1999 12:37:47 EDT. <99062113061000.18239@par28.ma.ikos.com> Date: Mon, 21 Jun 1999 16:08:34 -0400 Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > Also there are rumours that Alpha performance is critically > dependent on the compiler used - i.e. the DEC compiler on Tru64 Unix > (is that this week's name ?) might give you 30% more performance > than gcc. This is third-hand information, so may be false (or it > may be true for Alpha 21164 but not 21264, which I think does more > dynamic scheduling ?). I can confirm this for our case. We run a mixed cluster of Alpha machines (533 MHz 21664LX) and Intel machines (mostly dual Pentium II 450MHz), and the use of the "correct" compiler (especially for some of our Fortran applications) ais critical to performance on the Alphas. This isn't too surprising, as a hard-core RISC arch like the Alpha is going to be much more sensitive to proper code optimization (or lack thereof) than will the Pentium. sc -- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message