LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
To: Ingo Molnar <mingo@kernel.org>, Peter Zijlstra <peterz@infradead.org>
Cc: LKML <linux-kernel@vger.kernel.org>,
	Mel Gorman <mgorman@techsingularity.net>,
	Rik van Riel <riel@surriel.com>,
	Srikar Dronamraju <srikar@linux.vnet.ibm.com>,
	Thomas Gleixner <tglx@linutronix.de>
Subject: [PATCH 12/19] sched:numa Remove numa_has_capacity
Date: Mon,  4 Jun 2018 15:30:21 +0530	[thread overview]
Message-ID: <1528106428-19992-13-git-send-email-srikar@linux.vnet.ibm.com> (raw)
In-Reply-To: <1528106428-19992-1-git-send-email-srikar@linux.vnet.ibm.com>

task_numa_find_cpu helps to find the cpu to swap/move the task.
Its guarded by numa_has_capacity(). However node not having capacity
shouldn't deter a task swapping if it helps numa placement.

Further load_too_imbalanced, which evaluates possibilities of move/swap,
provides similar checks as numa_has_capacity.

Hence remove numa_has_capacity() to enhance possibilities of task
swapping even if load is imbalanced.

Testcase       Time:         Min         Max         Avg      StdDev
numa01.sh      Real:      674.61      997.71      785.01      115.95
numa01.sh       Sys:      180.87      318.88      270.13       51.32
numa01.sh      User:    54001.30    71936.50    60495.48     6237.55
numa02.sh      Real:       60.62       62.30       61.46        0.62
numa02.sh       Sys:       15.01       33.63       24.38        6.81
numa02.sh      User:     5234.20     5325.60     5276.23       38.85
numa03.sh      Real:      827.62      946.85      914.48       44.58
numa03.sh       Sys:      135.55      172.40      158.46       12.75
numa03.sh      User:    64839.42    73195.44    70805.96     3061.20
numa04.sh      Real:      481.01      608.76      521.14       47.28
numa04.sh       Sys:      329.59      373.15      353.20       14.20
numa04.sh      User:    37649.09    40722.94    38806.32     1072.32
numa05.sh      Real:      399.21      415.38      409.88        5.54
numa05.sh       Sys:      319.46      418.57      363.31       37.62
numa05.sh      User:    33727.77    34732.68    34127.41      447.11

Testcase       Time:         Min         Max         Avg      StdDev 	 %Change
numa01.sh      Real:      435.67      707.28      527.49       97.85 	 48.81%
numa01.sh       Sys:       76.41      231.19      162.49       56.13 	 66.24%
numa01.sh      User:    38247.36    59033.52    45129.31     7642.69 	 34.04%
numa02.sh      Real:       60.35       62.09       61.09        0.69 	 0.605%
numa02.sh       Sys:       15.01       30.20       20.64        5.56 	 18.12%
numa02.sh      User:     5195.93     5294.82     5240.99       40.55 	 0.672%
numa03.sh      Real:      752.04      919.89      836.81       63.29 	 9.281%
numa03.sh       Sys:      115.10      133.35      125.46        7.78 	 26.30%
numa03.sh      User:    58736.44    70084.26    65103.67     4416.10 	 8.758%
numa04.sh      Real:      418.43      709.69      512.53      104.17 	 1.679%
numa04.sh       Sys:      242.99      370.47      297.39       42.20 	 18.76%
numa04.sh      User:    34916.14    48429.54    38955.65     4928.05 	 -0.38%
numa05.sh      Real:      379.27      434.05      403.70       17.79 	 1.530%
numa05.sh       Sys:      145.94      344.50      268.72       68.53 	 35.20%
numa05.sh      User:    32679.32    35449.75    33989.10      913.19 	 0.406%

Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
---
 kernel/sched/fair.c | 37 +++----------------------------------
 1 file changed, 3 insertions(+), 34 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 259c343..709c77c 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -1417,7 +1417,6 @@ struct numa_stats {
 	unsigned long compute_capacity;
 
 	unsigned int nr_running;
-	int has_free_capacity;
 };
 
 /*
@@ -1444,8 +1443,7 @@ static void update_numa_stats(struct numa_stats *ns, int nid)
 	 * the @ns structure is NULL'ed and task_numa_compare() will
 	 * not find this node attractive.
 	 *
-	 * We'll either bail at !has_free_capacity, or we'll detect a huge
-	 * imbalance and bail there.
+	 * We'll detect a huge imbalance and bail there.
 	 */
 	if (!cpus)
 		return;
@@ -1456,7 +1454,6 @@ static void update_numa_stats(struct numa_stats *ns, int nid)
 
 	capacity = min_t(unsigned, capacity,
 		DIV_ROUND_CLOSEST(ns->compute_capacity, SCHED_CAPACITY_SCALE));
-	ns->has_free_capacity = (ns->nr_running < capacity);
 }
 
 struct task_numa_env {
@@ -1672,7 +1669,6 @@ static void task_numa_find_cpu(struct task_numa_env *env,
 	 * than swapping tasks around, check if a move is possible.
 	 */
 	move = !load_too_imbalanced(src_load, dst_load, env);
-
 	for_each_cpu(cpu, cpumask_of_node(env->dst_nid)) {
 		/* Skip this CPU if the source task cannot migrate */
 		if (!cpumask_test_cpu(cpu, &env->p->cpus_allowed))
@@ -1683,31 +1679,6 @@ static void task_numa_find_cpu(struct task_numa_env *env,
 	}
 }
 
-/* Only move tasks to a NUMA node less busy than the current node. */
-static bool numa_has_capacity(struct task_numa_env *env)
-{
-	struct numa_stats *src = &env->src_stats;
-	struct numa_stats *dst = &env->dst_stats;
-
-	if (src->has_free_capacity && !dst->has_free_capacity)
-		return false;
-
-	/*
-	 * Only consider a task move if the source has a higher load
-	 * than the destination, corrected for CPU capacity on each node.
-	 *
-	 *      src->load                dst->load
-	 * --------------------- vs ---------------------
-	 * src->compute_capacity    dst->compute_capacity
-	 */
-	if (src->load * dst->compute_capacity * env->imbalance_pct >
-
-	    dst->load * src->compute_capacity * 100)
-		return true;
-
-	return false;
-}
-
 static int task_numa_migrate(struct task_struct *p)
 {
 	struct task_numa_env env = {
@@ -1764,8 +1735,7 @@ static int task_numa_migrate(struct task_struct *p)
 	update_numa_stats(&env.dst_stats, env.dst_nid);
 
 	/* Try to find a spot on the preferred nid. */
-	if (numa_has_capacity(&env))
-		task_numa_find_cpu(&env, taskimp, groupimp);
+	task_numa_find_cpu(&env, taskimp, groupimp);
 
 	/*
 	 * Look at other nodes in these cases:
@@ -1795,8 +1765,7 @@ static int task_numa_migrate(struct task_struct *p)
 			env.dist = dist;
 			env.dst_nid = nid;
 			update_numa_stats(&env.dst_stats, env.dst_nid);
-			if (numa_has_capacity(&env))
-				task_numa_find_cpu(&env, taskimp, groupimp);
+			task_numa_find_cpu(&env, taskimp, groupimp);
 		}
 	}
 
-- 
1.8.3.1

  parent reply	other threads:[~2018-06-04 10:01 UTC|newest]

Thread overview: 66+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-06-04 10:00 [PATCH 00/19] Fixes for sched/numa_balancing Srikar Dronamraju
2018-06-04 10:00 ` [PATCH 01/19] sched/numa: Remove redundant field Srikar Dronamraju
2018-06-04 14:53   ` Rik van Riel
2018-06-05  8:41   ` Mel Gorman
2018-06-04 10:00 ` [PATCH 02/19] sched/numa: Evaluate move once per node Srikar Dronamraju
2018-06-04 14:51   ` Rik van Riel
2018-06-04 15:45     ` Srikar Dronamraju
2018-06-04 10:00 ` [PATCH 03/19] sched/numa: Simplify load_too_imbalanced Srikar Dronamraju
2018-06-04 14:57   ` Rik van Riel
2018-06-05  8:46   ` Mel Gorman
2018-06-04 10:00 ` [PATCH 04/19] sched/numa: Set preferred_node based on best_cpu Srikar Dronamraju
2018-06-04 12:18   ` Peter Zijlstra
2018-06-04 12:53     ` Srikar Dronamraju
2018-06-04 12:23   ` Peter Zijlstra
2018-06-04 12:59     ` Srikar Dronamraju
2018-06-04 13:39       ` Peter Zijlstra
2018-06-04 13:48         ` Srikar Dronamraju
2018-06-04 14:37       ` Rik van Riel
2018-06-04 15:56         ` Srikar Dronamraju
2018-06-04 10:00 ` [PATCH 05/19] sched/numa: Use task faults only if numa_group is not yet setup Srikar Dronamraju
2018-06-04 12:24   ` Peter Zijlstra
2018-06-04 13:09     ` Srikar Dronamraju
2018-06-04 10:00 ` [PATCH 06/19] sched/debug: Reverse the order of printing faults Srikar Dronamraju
2018-06-04 16:28   ` Rik van Riel
2018-06-05  8:50   ` Mel Gorman
2018-06-04 10:00 ` [PATCH 07/19] sched/numa: Skip nodes that are at hoplimit Srikar Dronamraju
2018-06-04 16:27   ` Rik van Riel
2018-06-05  8:50   ` Mel Gorman
2018-06-04 10:00 ` [PATCH 08/19] sched/numa: Remove unused task_capacity from numa_stats Srikar Dronamraju
2018-06-04 16:28   ` Rik van Riel
2018-06-05  8:57   ` Mel Gorman
2018-06-04 10:00 ` [PATCH 09/19] sched/numa: Modify migrate_swap to accept additional params Srikar Dronamraju
2018-06-04 17:00   ` Rik van Riel
2018-06-05  8:58   ` Mel Gorman
2018-06-04 10:00 ` [PATCH 10/19] sched/numa: Stop multiple tasks from moving to the cpu at the same time Srikar Dronamraju
2018-06-04 17:57   ` Rik van Riel
2018-06-05  9:51   ` Mel Gorman
2018-06-04 10:00 ` [PATCH 11/19] sched/numa: Restrict migrating in parallel to the same node Srikar Dronamraju
2018-06-04 17:59   ` Rik van Riel
2018-06-05  9:53   ` Mel Gorman
2018-06-06 12:58     ` Srikar Dronamraju
2018-06-04 10:00 ` Srikar Dronamraju [this message]
2018-06-04 18:07   ` [PATCH 12/19] sched:numa Remove numa_has_capacity Rik van Riel
2018-06-04 10:00 ` [PATCH 13/19] mm/migrate: Use xchg instead of spinlock Srikar Dronamraju
2018-06-04 18:22   ` Rik van Riel
2018-06-04 19:28   ` Peter Zijlstra
2018-06-05  7:24     ` Srikar Dronamraju
2018-06-05  8:16       ` Peter Zijlstra
2018-06-04 10:00 ` [PATCH 14/19] sched/numa: Updation of scan period need not be in lock Srikar Dronamraju
2018-06-04 18:24   ` Rik van Riel
2018-06-04 10:00 ` [PATCH 15/19] sched/numa: Use group_weights to identify if migration degrades locality Srikar Dronamraju
2018-06-04 18:56   ` Rik van Riel
2018-06-04 10:00 ` [PATCH 16/19] sched/numa: Detect if node actively handling migration Srikar Dronamraju
2018-06-04 20:05   ` Rik van Riel
2018-06-05  3:56     ` Srikar Dronamraju
2018-06-05 13:07       ` Rik van Riel
2018-06-06 12:55         ` Srikar Dronamraju
2018-06-06 13:55           ` Rik van Riel
2018-06-06 15:32             ` Srikar Dronamraju
2018-06-06 17:06               ` Rik van Riel
2018-06-04 10:00 ` [PATCH 17/19] sched/numa: Pass destination cpu as a parameter to migrate_task_rq Srikar Dronamraju
2018-06-04 10:00 ` [PATCH 18/19] sched/numa: Reset scan rate whenever task moves across nodes Srikar Dronamraju
2018-06-04 20:08   ` Rik van Riel
2018-06-05  9:58   ` Mel Gorman
2018-06-06 13:47     ` Srikar Dronamraju
2018-06-04 10:00 ` [PATCH 19/19] sched/numa: Move task_placement closer to numa_migrate_preferred Srikar Dronamraju

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=1528106428-19992-13-git-send-email-srikar@linux.vnet.ibm.com \
    --to=srikar@linux.vnet.ibm.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mgorman@techsingularity.net \
    --cc=mingo@kernel.org \
    --cc=peterz@infradead.org \
    --cc=riel@surriel.com \
    --cc=tglx@linutronix.de \
    --subject='Re: [PATCH 12/19] sched:numa Remove numa_has_capacity' \
    /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

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).