From mboxrd@z Thu Jan 1 00:00:00 1970 From: kanoj@google.engr.sgi.com (Kanoj Sarcar) Message-Id: <199906141717.KAA31065@google.engr.sgi.com> Subject: Re: process selection Date: Mon, 14 Jun 1999 10:17:27 -0700 (PDT) In-Reply-To: from "Rik van Riel" at Jun 12, 99 10:00:30 pm MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org Return-Path: To: Rik van Riel Cc: linux-mm@kvack.org List-ID: > > Could it be an idea to take the 'sleeping time' of each > process into account when selecting which process to swap > out? Due to extreme lack of free time, I'm asking what > you folks think of it before testing it myself... > You are right, sleep time is a good heuristic to determine the "swappability" of a process. Hmm, I wonder if this is what happended in your case: setiathome probably had a big rss, but netscape and X probably had larger rss and got selected for stealing. These are just a couple of things probably worth trying out: 1. The stealing algorithm can be upgraded to steal more than just SWAP_CLUSTER_MAX, for all the work it does. 2. Also, in swap_out, it might make sense to steal more than a single page from a victim process, to balance the overhead of scanning all the processes. Kanoj -- To unsubscribe, send a message with 'unsubscribe linux-mm my@address' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://humbolt.geo.uu.nl/Linux-MM/