From: Nate Eldredge <neldredge AT hmc DOT edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <14545.19777.671368.908549@mercury.st.hmc.edu>
Date: Thu, 16 Mar 2000 13:08:17 -0800 (PST)
To: djgpp-workers AT delorie DOT com
Subject: Re: Fastest integer type
In-Reply-To: <200003161405.TAA00904@midpec.com>
References: <Pine DOT SUN DOT 3 DOT 91 DOT 1000316115430 DOT 3117J-100000 AT is>
	<200003161405 DOT TAA00904 AT midpec DOT com>
X-Mailer: VM 6.75 under Emacs 20.5.1
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: dj-admin AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

Prashant TR writes:
 > Eli Zaretskii <eliz AT is DOT elta DOT co DOT il> writes:
 > 
 > > > On my Pentium, 8-bits seem to be faster than the 32-bit
 > > 
 > > By how much?
 > 
 > The ratio of 8-bit:32-bit regs for the addition operation is
 > something like 140.97:102.87 on my Pentium.

Any idea what the benchmark it uses is like?  For instance, if it does
a lot of operations on a whole block of numbers, it may be important
that more 8-bit ints fit into the cache than 32-bits, even if they
aren't inherently faster.

 > Seems like there's quite a bit of difference here. IMHO 8-bit
 > (signed char) can be used.

-- 

Nate Eldredge
neldredge AT hmc DOT edu