From owner-freebsd-current Mon Jul 27 08:21:33 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id IAA03802 for freebsd-current-outgoing; Mon, 27 Jul 1998 08:21:33 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from dyson.iquest.net (dyson.iquest.net [198.70.144.127]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA03794 for ; Mon, 27 Jul 1998 08:21:24 -0700 (PDT) (envelope-from toor@dyson.iquest.net) Received: (from root@localhost) by dyson.iquest.net (8.8.8/8.8.8) id KAA01416; Mon, 27 Jul 1998 10:20:13 -0500 (EST) (envelope-from toor) Message-Id: <199807271520.KAA01416@dyson.iquest.net> Subject: Re: New LINT options: what is VM coloring? In-Reply-To: from Doug Rabson at "Jul 27, 98 09:51:47 am" To: dfr@nlsystems.com (Doug Rabson) Date: Mon, 27 Jul 1998 10:20:13 -0500 (EST) Cc: dyson@iquest.net, tlambert@primenet.com, roberto@keltia.freenix.fr, freebsd-current@FreeBSD.ORG From: "John S. Dyson" Reply-To: dyson@iquest.net X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Doug Rabson said: > On Sun, 26 Jul 1998, John S. Dyson wrote: > > > > Actually, the L1 cache is where it's most important (see paper references > > > in my previous posting). Also, the Alpha can significantly benefit from > > > this, per Digital UNIX: > > > > > Is the L1 cache on the Alpha is direct mapped??? On the X86, it isn't. When > > actually running tests, it doesn't seem to make ANY differences on the X86, > > due to the very small number of pages, and the mapping scheme. > > I'm not too sure what the layout of the alpha caches are. I expect they > are different on different processor generations. I really must start > reading the processor manuals for these things... > For two level caches, the coloring should work fine. However, it is problematical with 3-level alpha caches. My guess is that one would want to page-color for the 2nd and 3rd level caches, but of course, you might want to experiment. It is critical to color for the larger cache. To color for all three caches, you could just expand the coloring scheme in FreeBSD. I don't know if it is the right approach though. One thing: I experimented with cache-line coloring for data structures, and found absolutely no performance improvement, and perhaps a significant performance decrease on PPro, and so did not add it. You might want to investigate adding the data structure coloring if it is helpful for Alphas. -- John | Never try to teach a pig to sing, dyson@iquest.net | it makes one look stupid jdyson@nc.com | and it irritates the pig. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message