public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Steven Rostedt <rostedt@goodmis.org>
To: LKML <linux-kernel@vger.kernel.org>
Cc: Ingo Molnar <mingo@elte.hu>,
	Gregory Haskins <ghaskins@novell.com>,
	Peter Zijlstra <a.p.zijlstra@chello.nl>,
	Christoph Lameter <clameter@sgi.com>
Subject: [PATCH v4 18/20] Optimize cpu search with hamming weight
Date: Tue, 20 Nov 2007 20:01:12 -0500	[thread overview]
Message-ID: <20071121011251.746498886@goodmis.org> (raw)
In-Reply-To: 20071121010054.663842380@goodmis.org

[-- Attachment #1: rt-balance-find-cpu-count.patch --]
[-- Type: text/plain, Size: 2331 bytes --]

From: Gregory Haskins <ghaskins@novell.com>

We can cheaply track the number of bits set in the cpumask for the lowest
priority CPUs.  Therefore, compute the mask's weight and use it to skip
the optimal domain search logic when there is only one CPU available.

Signed-off-by: Gregory Haskins <ghaskins@novell.com>
---

 kernel/sched_rt.c |   25 ++++++++++++++++++-------
 1 file changed, 18 insertions(+), 7 deletions(-)

Index: linux-compile.git/kernel/sched_rt.c
===================================================================
--- linux-compile.git.orig/kernel/sched_rt.c	2007-11-20 19:53:13.000000000 -0500
+++ linux-compile.git/kernel/sched_rt.c	2007-11-20 19:53:15.000000000 -0500
@@ -300,7 +300,7 @@ static int find_lowest_cpus(struct task_
 	int       cpu;
 	cpumask_t *valid_mask = &__get_cpu_var(valid_cpu_mask);
 	int       lowest_prio = -1;
-	int       ret         = 0;
+	int       count       = 0;
 
 	cpus_clear(*lowest_mask);
 	cpus_and(*valid_mask, cpu_online_map, task->cpus_allowed);
@@ -313,7 +313,7 @@ static int find_lowest_cpus(struct task_
 
 		/* We look for lowest RT prio or non-rt CPU */
 		if (rq->rt.highest_prio >= MAX_RT_PRIO) {
-			if (ret)
+			if (count)
 				cpus_clear(*lowest_mask);
 			cpu_set(rq->cpu, *lowest_mask);
 			return 1;
@@ -325,14 +325,17 @@ static int find_lowest_cpus(struct task_
 			if (rq->rt.highest_prio > lowest_prio) {
 				/* new low - clear old data */
 				lowest_prio = rq->rt.highest_prio;
-				cpus_clear(*lowest_mask);
+				if (count) {
+					cpus_clear(*lowest_mask);
+					count = 0;
+				}
 			}
 			cpu_set(rq->cpu, *lowest_mask);
-			ret = 1;
+			count++;
 		}
 	}
 
-	return ret;
+	return count;
 }
 
 static inline int pick_optimal_cpu(int this_cpu, cpumask_t *mask)
@@ -356,9 +359,17 @@ static int find_lowest_rq(struct task_st
 	cpumask_t *lowest_mask = &__get_cpu_var(local_cpu_mask);
 	int this_cpu = smp_processor_id();
 	int cpu      = task_cpu(task);
+	int count    = find_lowest_cpus(task, lowest_mask);
 
-	if (!find_lowest_cpus(task, lowest_mask))
-		return -1;
+	if (!count)
+		return -1; /* No targets found */
+
+	/*
+	 * There is no sense in performing an optimal search if only one
+	 * target is found.
+	 */
+	if (count == 1)
+		return first_cpu(*lowest_mask);
 
 	/*
 	 * At this point we have built a mask of cpus representing the

-- 

  parent reply	other threads:[~2007-11-21  1:13 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-11-21  1:00 [PATCH v4 00/20] New RT Balancing version 4 Steven Rostedt
2007-11-21  1:00 ` [PATCH v4 01/20] Add rt_nr_running accounting Steven Rostedt
2007-11-21  1:00 ` [PATCH v4 02/20] track highest prio queued on runqueue Steven Rostedt
2007-11-21  1:00 ` [PATCH v4 03/20] push RT tasks Steven Rostedt
2007-11-21  1:00 ` [PATCH v4 04/20] RT overloaded runqueues accounting Steven Rostedt
2007-11-21  1:00 ` [PATCH v4 05/20] pull RT tasks Steven Rostedt
2007-11-21  1:01 ` [PATCH v4 06/20] wake up balance RT Steven Rostedt
2007-11-21  1:01 ` [PATCH v4 07/20] disable CFS RT load balancing Steven Rostedt
2007-11-21  1:01 ` [PATCH v4 08/20] Cache cpus_allowed weight for optimizing migration Steven Rostedt
2007-11-21  1:01 ` [PATCH v4 09/20] RT: Consistency cleanup for this_rq usage Steven Rostedt
2007-11-21  1:01 ` [PATCH v4 10/20] RT: Remove some CFS specific code from the wakeup path of RT tasks Steven Rostedt
2007-11-21  1:01 ` [PATCH v4 11/20] RT: Break out the search function Steven Rostedt
2007-11-21  1:01 ` [PATCH v4 12/20] RT: Allow current_cpu to be included in search Steven Rostedt
2007-11-21  1:01 ` [PATCH v4 13/20] RT: Pre-route RT tasks on wakeup Steven Rostedt
2007-11-21  1:01 ` [PATCH v4 14/20] RT: Optimize our cpu selection based on topology Steven Rostedt
2007-11-21  1:01 ` [PATCH v4 15/20] RT: Optimize rebalancing Steven Rostedt
2007-11-21  1:01 ` [PATCH v4 16/20] Avoid overload Steven Rostedt
2007-11-21  1:01 ` [PATCH v4 17/20] RT: restore the migratable conditional Steven Rostedt
2007-11-21  1:01 ` Steven Rostedt [this message]
2007-11-21  1:01 ` [PATCH v4 19/20] Optimize out cpu_clears Steven Rostedt
2007-11-21  2:10   ` Steven Rostedt
2007-11-21  3:10     ` [PATCH] Fix optimized search Gregory Haskins
2007-11-21  4:15       ` Steven Rostedt
2007-11-21  4:26         ` Steven Rostedt
2007-11-21  5:14           ` Gregory Haskins
2007-11-21  1:01 ` [PATCH v4 20/20] balance RT tasks no new wake up Steven Rostedt
2007-11-21  4:44 ` [PATCH 0/4] more RT balancing enhancements Gregory Haskins
2007-11-21  4:44   ` [PATCH 1/4] Fix optimized search Gregory Haskins
2007-11-21  4:44   ` [PATCH 2/4] RT: Add sched-domain roots Gregory Haskins
2007-11-21  4:44   ` [PATCH 3/4] RT: Only balance our RT tasks within our root-domain Gregory Haskins
2007-11-21  4:44   ` [PATCH 4/4] RT: Use a 2-d bitmap for searching lowest-pri CPU Gregory Haskins
2007-11-21 19:51   ` [PATCH 0/4] more RT balancing enhancements v6a Gregory Haskins
2007-11-21 19:52     ` [PATCH 1/4] SCHED: Add sched-domain roots Gregory Haskins
2007-11-21 19:52     ` [PATCH 2/4] SCHED: Track online cpus in the root-domain Gregory Haskins
2007-11-21 19:52     ` [PATCH 3/4] SCHED: Only balance our RT tasks within our root-domain Gregory Haskins
2007-11-21 19:52     ` [PATCH 4/4] SCHED: Use a 2-d bitmap for searching lowest-pri CPU Gregory Haskins

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20071121011251.746498886@goodmis.org \
    --to=rostedt@goodmis.org \
    --cc=a.p.zijlstra@chello.nl \
    --cc=clameter@sgi.com \
    --cc=ghaskins@novell.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox