Date: Thu, 09 Nov 1995 09:40:03 +0100 From: Poul-Henning Kamp <phk@critter.tfs.com> To: Michael Smith <msmith@atrad.adelaide.edu.au> Cc: hackers@freebsd.org Subject: Re: Load/Store using FPU regs ... Message-ID: <6340.815906403@critter.tfs.com> In-Reply-To: Your message of "Thu, 09 Nov 1995 09:01:31 GMT." <199511090901.JAA09519@genesis.atrad.adelaide.edu.au>
next in thread | previous in thread | raw e-mail | index | archive | help
> koshy@blr.novell.com stands accused of saying: > > >>> L20: fldl (%ebx) fstpl (%ecx) ... > > >>> > > >>> The resulting program copies data at about 60 Megabytes per > > >>> second. > > > > > > Using the FPU registers for memmove/bitblt operations was a technique > > I first saw on an i860. We used to do a series of reads into FPU regs > > Wheras those of us with 68K backgrounds are rolling in the aisles about > this one 8) > > (For the uninitiated; the 68K can read/write arbitrary groups of registers > and increment/decrement the source/destination pointers at the same time. > Depending on coding technique, you can read or write as much as 56 > bytes at a time; the big win (microcoded processor, remember) being no > instruction fetches between reads. It's a pity that Motorola have axed > it as a mainstram family 8( ) > > Anyway, enough from the nostalgia corner - I'm too young for this! What you really want of course is a "outside the cache" bzero and bcopy function... -- Poul-Henning Kamp | phk@FreeBSD.ORG FreeBSD Core-team. http://www.freebsd.org/~phk | phk@login.dknet.dk Private mailbox. whois: [PHK] | phk@ref.tfs.com TRW Financial Systems, Inc. Future will arrive by its own means, progress not so.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?6340.815906403>