LKML Archive on lore.kernel.org
 help / color / Atom feed
From: Lauro Venancio <lvenanci@redhat.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: linux-kernel@vger.kernel.org, lwang@redhat.com, riel@redhat.com,
	Mike Galbraith <efault@gmx.de>,
	Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@kernel.org>
Subject: Re: [RFC 2/3] sched/topology: fix sched groups on NUMA machines with mesh topology
Date: Thu, 13 Apr 2017 18:06:57 -0300
Message-ID: <dea71fe0-9e39-43c8-2710-13767b14ce63@redhat.com> (raw)
In-Reply-To: <5166d6ba-c8e6-c60e-61af-d32124234bb9@redhat.com>

On 04/13/2017 05:21 PM, Lauro Venancio wrote:
> On 04/13/2017 12:48 PM, Peter Zijlstra wrote:
>> On Thu, Apr 13, 2017 at 10:56:08AM -0300, Lauro Ramos Venancio wrote:
>>> Currently, on a 4 nodes NUMA machine with ring topology, two sched
>>> groups are generated for the last NUMA sched domain. One group has the
>>> CPUs from NUMA nodes 3, 0 and 1; the other group has the CPUs from nodes
>>> 1, 2 and 3. As CPUs from nodes 1 and 3 belongs to both groups, the
>>> scheduler is unable to directly move tasks between these nodes. In the
>>> worst scenario, when a set of tasks are bound to nodes 1 and 3, the
>>> performance is severely impacted because just one node is used while the
>>> other node remains idle.
>> I feel a picture would be ever so much clearer.
>>
>>> This patch constructs the sched groups from each CPU perspective. So, on
>>> a 4 nodes machine with ring topology, while nodes 0 and 2 keep the same
>>> groups as before [(3, 0, 1)(1, 2, 3)], nodes 1 and 3 have new groups
>>> [(0, 1, 2)(2, 3, 0)]. This allows moving tasks between any node 2-hops
>>> apart.
>> So I still have no idea what specifically goes wrong and how this fixes
>> it. Changelog is impenetrable.
> On a 4 nodes machine with ring topology, the last sched domain level
> contains groups with 3 numa nodes each. So we have four possible groups:
> (0, 1, 2) (1, 2, 3) (2, 3, 0)(3, 0, 1). As we need just two groups to
> fill the sched domain, currently, the groups (3, 0, 1) and (1, 2, 3) are
> used for all CPUs. The problem with it is that nodes 1 and 3 belongs to
> both groups, becoming impossible to move tasks between these two nodes.
>
> This patch uses different groups depending on the CPU they are
> installed. So nodes 0 and 2 CPUs keep the same group as before: (3, 0,
> 1) and (1, 2, 3). Nodes 1 and 3 CPUs use the new groups: (0, 1, 2) and
> (2, 3, 0). So the first pair of groups allows movement between nodes 0
> and 2; and the second pair of groups allows movement between nodes 1 and 3.
>
> I will improve the changelog.
>
>> "From each CPU's persepective" doesn't really help, there already is a
>> for_each_cpu() in.
> The for_each_cpu() is used to iterate across all sched domain cpus. It
> doesn't consider the CPU where the groups are being installed (parameter
> cpu in build_overlap_sched_groups()). Currently, the parameter cpu is
> used just for memory allocation and for ordering the groups, it doesn't
> change the groups that are chosen. This patch uses the parameter cpu to
> choose the first group, changing also, as consequence, the second group.
>> Also, since I'm not sure what happend to the 4 node system, I cannot
>> begin to imagine what would happen on the 8 node one.

Just for clarification, I am sending the nodes distance table for the
two most common typologies affected by this issue.

4 nodes, ring topology
node distances:
node   0   1   2   3
  0:  10  20  30  20
  1:  20  10  20  30
  2:  30  20  10  20
  3:  20  30  20  10

8 node, mesh topology
node distances:
node   0   1   2   3   4   5   6   7
  0:  10  16  16  22  16  22  16  22
  1:  16  10  16  22  22  16  22  16
  2:  16  16  10  16  16  16  16  22
  3:  22  22  16  10  16  16  22  16
  4:  16  22  16  16  10  16  16  16
  5:  22  16  16  16  16  10  22  22
  6:  16  22  16  22  16  22  10  16
  7:  22  16  22  16  16  22  16  10

  reply index

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-04-13 13:56 [RFC 0/3] " Lauro Ramos Venancio
2017-04-13 13:56 ` [RFC 1/3] sched/topology: Refactor function build_overlap_sched_groups() Lauro Ramos Venancio
2017-04-13 14:50   ` Rik van Riel
2017-05-15  9:02   ` [tip:sched/core] " tip-bot for Lauro Ramos Venancio
2017-04-13 13:56 ` [RFC 2/3] sched/topology: fix sched groups on NUMA machines with mesh topology Lauro Ramos Venancio
2017-04-13 15:16   ` Rik van Riel
2017-04-13 15:48   ` Peter Zijlstra
2017-04-13 20:21     ` Lauro Venancio
2017-04-13 21:06       ` Lauro Venancio [this message]
2017-04-13 23:38         ` Rik van Riel
2017-04-14 10:48           ` Peter Zijlstra
2017-04-14 11:38   ` Peter Zijlstra
2017-04-14 12:20     ` Peter Zijlstra
2017-05-15  9:03       ` [tip:sched/core] sched/fair, cpumask: Export for_each_cpu_wrap() tip-bot for Peter Zijlstra
2017-05-17 10:53         ` hackbench vs select_idle_sibling; was: " Peter Zijlstra
2017-05-17 12:46           ` Matt Fleming
2017-05-17 14:49           ` Chris Mason
2017-05-19 15:00           ` Matt Fleming
2017-06-05 13:00             ` Matt Fleming
2017-06-06  9:21               ` Peter Zijlstra
2017-06-09 17:52                 ` Chris Mason
2017-06-08  9:22           ` [tip:sched/core] sched/core: Implement new approach to scale select_idle_cpu() tip-bot for Peter Zijlstra
2017-04-14 16:58     ` [RFC 2/3] sched/topology: fix sched groups on NUMA machines with mesh topology Peter Zijlstra
2017-04-17 14:40       ` Lauro Venancio
2017-04-13 13:56 ` [RFC 3/3] sched/topology: Different sched groups must not have the same balance cpu Lauro Ramos Venancio
2017-04-13 15:27   ` Rik van Riel
2017-04-14 16:49   ` Peter Zijlstra
2017-04-17 15:34     ` Lauro Venancio
2017-04-18 12:32       ` Peter Zijlstra

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=dea71fe0-9e39-43c8-2710-13767b14ce63@redhat.com \
    --to=lvenanci@redhat.com \
    --cc=efault@gmx.de \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lwang@redhat.com \
    --cc=mingo@kernel.org \
    --cc=peterz@infradead.org \
    --cc=riel@redhat.com \
    --cc=tglx@linutronix.de \
    /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

LKML Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/lkml/0 lkml/git/0.git
	git clone --mirror https://lore.kernel.org/lkml/1 lkml/git/1.git
	git clone --mirror https://lore.kernel.org/lkml/2 lkml/git/2.git
	git clone --mirror https://lore.kernel.org/lkml/3 lkml/git/3.git
	git clone --mirror https://lore.kernel.org/lkml/4 lkml/git/4.git
	git clone --mirror https://lore.kernel.org/lkml/5 lkml/git/5.git
	git clone --mirror https://lore.kernel.org/lkml/6 lkml/git/6.git
	git clone --mirror https://lore.kernel.org/lkml/7 lkml/git/7.git
	git clone --mirror https://lore.kernel.org/lkml/8 lkml/git/8.git
	git clone --mirror https://lore.kernel.org/lkml/9 lkml/git/9.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 lkml lkml/ https://lore.kernel.org/lkml \
		linux-kernel@vger.kernel.org
	public-inbox-index lkml

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-kernel


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git