Date: Mon, 21 Aug 2000 06:50:54 +1000 From: Peter Jeremy <peter.jeremy@alcatel.com.au> To: Wilko Bulte <wkb@freebie.demon.nl> Cc: alpha@FreeBSD.org Subject: Re: optimized mem* and str* functions Message-ID: <00Aug21.065104est.115216@border.alcanet.com.au> In-Reply-To: <20000819152032.A1483@freebie.demon.nl>; from wkb@freebie.demon.nl on Sat, Aug 19, 2000 at 03:20:32PM %2B0200 References: <39999750.C80F3B0B@imimic.com> <00Aug18.140950est.115233@border.alcanet.com.au> <20000819152032.A1483@freebie.demon.nl>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2000-Aug-19 15:20:32 +0200, Wilko Bulte <wkb@freebie.demon.nl> wrote: >The later alpha/AXP implementations have byte manipulation instructions >IIRC. Correct. > Are they used when available? DEC CC and (AFAIK) GCC don't generate them by default, but they can be enabled (with GCC) by specifying `-mbwx' or `-mcpu=ev56' (or later). Note that using the byte instructions isn't necessarily an optimisation. It's often (usually?) faster to load a full register and unroll the loop to process each byte in the register. Peter To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?00Aug21.065104est.115216>