Message-ID: <19990623180740.H28851@mindspring.com> Date: Wed, 23 Jun 1999 18:07:40 -0500 From: Anderson MacKay To: pgcc AT delorie DOT com Subject: Re: performance issue References: <19990621182836 DOT O28893 AT io DOT txc DOT com> <19990623150531 DOT A9441 AT cerebro DOT laendle> <14193 DOT 18042 DOT 784347 DOT 242833 AT lrz DOT de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.93i In-Reply-To: <14193.18042.784347.242833@lrz.de>; from Eugene Leitl on Wed, Jun 23, 1999 at 01:46:20PM -0700 Reply-To: pgcc AT delorie DOT com X-Mailing-List: pgcc AT delorie DOT com X-Unsubscribes-To: listserv AT delorie DOT com Precedence: bulk On Wed, Jun 23, 1999 at 01:46:20PM -0700, Eugene Leitl wrote: > > Apropos another CPU you don't have: do you think that K7 will be as > quirky as K6 in respect to compiler options? Assuming both egcs/gcc > and pgcc. > Probably not. The K7 has very deep scheduling buffers, which should make it run quickly on generally well-optimized code, regardless of the asm scheduling. It even lacks the restriction of the P2/3 series of needing to intersperse FADD and FMUL to achieve full floating-point bandwidth. -- Anderson MacKay