From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Tue, 11 Jul 2000 19:54:31 +0200 (CEST) From: Andrea Arcangeli Subject: Re: [PATCH] 2.2.17pre7 VM enhancement Re: I/O performance on 2.4.0-test2 In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-mm@kvack.org Return-Path: To: Rik van Riel Cc: "Juan J. Quintela" , "Stephen C. Tweedie" , Marcelo Tosatti , Jens Axboe , Alan Cox , Derek Martin , Linux Kernel , linux-mm@kvack.org, "David S. Miller" List-ID: On Tue, 11 Jul 2000, Rik van Riel wrote: >This is why LRU is wrong and we need page aging (which >approximates both LRU and NFU). > >The idea is to remove those pages from memory which will >not be used again for the longest time, regardless of in >which 'state' they live in main memory. > >(and proper page aging is a good approximation to this) It will still drop _all_ VM mappings from memory if you left "cp /dev/zero ." in background for say 2 hours. This in turn mean that during streming I/O you'll have _much_ more than the current swapin/swapout troubles. If I download a dozen of CD images with a gigabit ethernet I don't want _anything_ to be unmapped from main RAM, and yes I may have 8giga of RAM so I don't want to use O_DIRECT for the downloads. Andrea -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux.eu.org/Linux-MM/