From baae6f425182a30bad2c148c2636a331f6c83b6f Mon Sep 17 00:00:00 2001 From: gferg <> Date: Thu, 26 Oct 2000 14:36:02 +0000 Subject: [PATCH] updated --- LDP/howto/linuxdoc/SMP-HOWTO.sgml | 49 +++++++++++++++++++++++++++++-- 1 file changed, 46 insertions(+), 3 deletions(-) diff --git a/LDP/howto/linuxdoc/SMP-HOWTO.sgml b/LDP/howto/linuxdoc/SMP-HOWTO.sgml index 4b5841ef..952e369e 100644 --- a/LDP/howto/linuxdoc/SMP-HOWTO.sgml +++ b/LDP/howto/linuxdoc/SMP-HOWTO.sgml @@ -4,7 +4,7 @@ Linux SMP HOWTO <author>David Mentré, <tt/David.Mentre@irisa.fr/ -<date>v1.11, 8 october 2000 +<date>v1.12.1, 25 october 2000 <abstract> This HOWTO reviews main issues (and I hope solutions) related to SMP @@ -261,6 +261,13 @@ If you want to gauge the performance of your SMP system, you can run some tests Cameron MacKinnon and available at <htmlurl name="http://www.phy.duke.edu/brahma/benchmarks.smp" url="http://www.phy.duke.edu/brahma/benchmarks.smp">. +<p> +Also have a look at this article by Bryant, Hartner, Qi and +Venkitachalam that compares 2.2 and 2.3/2.4 UP and SMP kernels : <url +url="http://www.usenix.org/publications/library/proceedings/als2000/bryantscale.html" +name="SMP Scalability Comparisons of Linux¨ Kernels 2.2.14 and 2.3.99"> +(<bf>Ray Bryant</bf>) (You'll find also a copy <url name="here" +url="bryantscale.pdf">) </enum> @@ -325,6 +332,29 @@ name="http://www.ima.umn.edu/~klee/linux/xosview-1.6.1-5a1.tgz"> By the way, you can't monitor processor scheduling precisely with xosview, as xosview itself causes a scheduling perturbation. (<bf>H. Peter Anvin</bf>) +<p> +And <bf>Rik van Riel</bf> tell us why: +<quote> +The answer is pretty simple. Basically there are 3 +processes involved: +<enum> +<item> the cpu hog (low scheduling priority because it eats CPU) +<item> xosview +<item> X +</enum> +<p> +The CPU hog is running on one CPU. Then xosview wakes up +(on the other CPU) and starts sending commands to X, which +wakes up as well. +<p> +Since both X and xosview have a much higher priority than +the CPU hog, xosview will run on one CPU and X on the other. +<p> +Then xosview stops running and we have an idle CPU --> Linux +moves the CPU hog over to the newly idle CPU (X is still +running on the CPU our hog was running on just before). +</quote> + <item> <bf>How can I enable more than 1 process for my kernel compile?</bf> <p> @@ -949,13 +979,13 @@ page (only SMP systems): <item> UltraSPARC SBUS based large servers: Enterprise 3000, 4000, 5000, 6000, 10000 <item> UltraSPARC PCI based servers: Enterprise 250, 450 <item> SPARC sun4m SMP machines (<bf>Anton Blanchard</bf>) -<item> <url name="StarFire E10000" +<item> <url name="Starfire E10000" url="http://linuxcare.com.au/anton/e10000/"> </itemize> UltraLinux has ran on a 14 CPUs machine (see the <url name="dmesg output" url="http://lwn.net/1998/1210/a/dm-sparc.html">) and on a -StarFire E10000 with 24 CPUs (see the <url name="dmesg +Starfire E10000 with 24 CPUs (see the <url name="dmesg output" url="http://linuxcare.com.au/anton/e10000/dmesg_24.shtml">). <sect1>Specific problem related to Sparc SMP support @@ -1290,6 +1320,17 @@ of locks. :-) <p> <descrip> +<tag/v1.12.1, 25 october 2000 +<itemize> +<item> Put all authors in Bryant, Hartner, Qi and Venkitachalam paper +</itemize> + +<tag/v1.12, 22 october 2000 +<itemize> +<item> Explanation on why not trust Xosview on scheduling (<bf>Rik van Riel</bf>) +<item> A pointer to an article that compares 2.2 and 2.4 kernels (<bf>Ray Bryant</bf>) +</itemize> + <tag/v1.11, 8 october 2000 <itemize> <item> Linux boots on a Sun E1000 with 24 CPUs @@ -1616,6 +1657,7 @@ Many thanks to those who help me to maintain this HOWTO: <item> Anton Blanchard <item> Emil Briggs <item> Robert G. Brown +<item> Ray Bryant <item> Alexandre Charbey <item> Michael Elizabeth Chastain <item> Samuel S. Chessman @@ -1665,6 +1707,7 @@ Many thanks to those who help me to maintain this HOWTO: <item> Jean-Michel Rouet <item> Volker Reichelt <item> Sean Reifschneider +<item> Rik van Riel <item> Sumit Roy <item> Thomas Schenk <item> Matthias Schniedermeyer