From: Matt Fleming <matt@codeblueprint.co.uk>
To: Peter Zijlstra <peterz@infradead.org>
Cc: linux-kernel@vger.kernel.org, Tony Luck <tony.luck@intel.com>,
Rik van Riel <riel@surriel.com>,
Suravee.Suthikulpanit@amd.com, Borislav Petkov <bp@alien8.de>,
Thomas.Lendacky@amd.com, Mel Gorman <mgorman@techsingularity.net>,
Matt Fleming <matt@codeblueprint.co.uk>
Subject: [PATCH v4 2/2] sched/topology: Improve load balancing on AMD EPYC
Date: Thu, 8 Aug 2019 20:53:01 +0100 [thread overview]
Message-ID: <20190808195301.13222-3-matt@codeblueprint.co.uk> (raw)
In-Reply-To: <20190808195301.13222-1-matt@codeblueprint.co.uk>
SD_BALANCE_{FORK,EXEC} and SD_WAKE_AFFINE are stripped in sd_init()
for any sched domains with a NUMA distance greater than 2 hops
(RECLAIM_DISTANCE). The idea being that it's expensive to balance
across domains that far apart.
However, as is rather unfortunately explained in
commit 32e45ff43eaf ("mm: increase RECLAIM_DISTANCE to 30")
the value for RECLAIM_DISTANCE is based on node distance tables from
2011-era hardware.
Current AMD EPYC machines have the following NUMA node distances:
node distances:
node 0 1 2 3 4 5 6 7
0: 10 16 16 16 32 32 32 32
1: 16 10 16 16 32 32 32 32
2: 16 16 10 16 32 32 32 32
3: 16 16 16 10 32 32 32 32
4: 32 32 32 32 10 16 16 16
5: 32 32 32 32 16 10 16 16
6: 32 32 32 32 16 16 10 16
7: 32 32 32 32 16 16 16 10
where 2 hops is 32.
The result is that the scheduler fails to load balance properly across
NUMA nodes on different sockets -- 2 hops apart.
For example, pinning 16 busy threads to NUMA nodes 0 (CPUs 0-7) and 4
(CPUs 32-39) like so,
$ numactl -C 0-7,32-39 ./spinner 16
causes all threads to fork and remain on node 0 until the active
balancer kicks in after a few seconds and forcibly moves some threads
to node 4.
Override node_reclaim_distance for AMD Zen.
Signed-off-by: Matt Fleming <matt@codeblueprint.co.uk>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: Suravee.Suthikulpanit@amd.com
Cc: Borislav Petkov <bp@alien8.de>
Cc: Thomas.Lendacky@amd.com
---
arch/x86/kernel/cpu/amd.c | 5 +++++
include/linux/topology.h | 14 ++++++++++++++
kernel/sched/topology.c | 3 ++-
mm/khugepaged.c | 2 +-
mm/page_alloc.c | 2 +-
5 files changed, 23 insertions(+), 3 deletions(-)
diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index 8d4e50428b68..ceeb8afc7cf3 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -8,6 +8,7 @@
#include <linux/sched.h>
#include <linux/sched/clock.h>
#include <linux/random.h>
+#include <linux/topology.h>
#include <asm/processor.h>
#include <asm/apic.h>
#include <asm/cacheinfo.h>
@@ -824,6 +825,10 @@ static void init_amd_zn(struct cpuinfo_x86 *c)
{
set_cpu_cap(c, X86_FEATURE_ZEN);
+#ifdef CONFIG_NUMA
+ node_reclaim_distance = 32;
+#endif
+
/*
* Fix erratum 1076: CPB feature bit not being set in CPUID.
* Always set it, except when running under a hypervisor.
diff --git a/include/linux/topology.h b/include/linux/topology.h
index 47a3e3c08036..579522ec446c 100644
--- a/include/linux/topology.h
+++ b/include/linux/topology.h
@@ -59,6 +59,20 @@ int arch_update_cpu_topology(void);
*/
#define RECLAIM_DISTANCE 30
#endif
+
+/*
+ * The following tunable allows platforms to override the default node
+ * reclaim distance (RECLAIM_DISTANCE) if remote memory accesses are
+ * sufficiently fast that the default value actually hurts
+ * performance.
+ *
+ * AMD EPYC machines use this because even though the 2-hop distance
+ * is 32 (3.2x slower than a local memory access) performance actually
+ * *improves* if allowed to reclaim memory and load balance tasks
+ * between NUMA nodes 2-hops apart.
+ */
+extern int __read_mostly node_reclaim_distance;
+
#ifndef PENALTY_FOR_NODE_WITH_CPUS
#define PENALTY_FOR_NODE_WITH_CPUS (1)
#endif
diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
index 8f83e8e3ea9a..b5667a273bf6 100644
--- a/kernel/sched/topology.c
+++ b/kernel/sched/topology.c
@@ -1284,6 +1284,7 @@ static int sched_domains_curr_level;
int sched_max_numa_distance;
static int *sched_domains_numa_distance;
static struct cpumask ***sched_domains_numa_masks;
+int __read_mostly node_reclaim_distance = RECLAIM_DISTANCE;
#endif
/*
@@ -1402,7 +1403,7 @@ sd_init(struct sched_domain_topology_level *tl,
sd->flags &= ~SD_PREFER_SIBLING;
sd->flags |= SD_SERIALIZE;
- if (sched_domains_numa_distance[tl->numa_level] > RECLAIM_DISTANCE) {
+ if (sched_domains_numa_distance[tl->numa_level] > node_reclaim_distance) {
sd->flags &= ~(SD_BALANCE_EXEC |
SD_BALANCE_FORK |
SD_WAKE_AFFINE);
diff --git a/mm/khugepaged.c b/mm/khugepaged.c
index eaaa21b23215..ccede2425c3f 100644
--- a/mm/khugepaged.c
+++ b/mm/khugepaged.c
@@ -710,7 +710,7 @@ static bool khugepaged_scan_abort(int nid)
for (i = 0; i < MAX_NUMNODES; i++) {
if (!khugepaged_node_load[i])
continue;
- if (node_distance(nid, i) > RECLAIM_DISTANCE)
+ if (node_distance(nid, i) > node_reclaim_distance)
return true;
}
return false;
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 272c6de1bf4e..0d54cd2c43a4 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -3522,7 +3522,7 @@ bool zone_watermark_ok_safe(struct zone *z, unsigned int order,
static bool zone_allows_reclaim(struct zone *local_zone, struct zone *zone)
{
return node_distance(zone_to_nid(local_zone), zone_to_nid(zone)) <=
- RECLAIM_DISTANCE;
+ node_reclaim_distance;
}
#else /* CONFIG_NUMA */
static bool zone_allows_reclaim(struct zone *local_zone, struct zone *zone)
--
2.13.7
next prev parent reply other threads:[~2019-08-08 19:53 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-08 19:52 [PATCH v4 0/2] sched: Improve load balancing on AMD EPYC Matt Fleming
2019-08-08 19:53 ` [PATCH 1/2] ia64: Make NUMA select SMP Matt Fleming
2019-09-03 8:31 ` [tip: sched/core] arch, " tip-bot2 for Matt Fleming
2019-08-08 19:53 ` Matt Fleming [this message]
2019-09-03 8:31 ` [tip: sched/core] sched/topology: Improve load balancing on AMD EPYC systems tip-bot2 for Matt Fleming
2019-10-07 15:28 ` [PATCH v4 2/2] sched/topology: Improve load balancing on AMD EPYC Guenter Roeck
2019-10-09 12:04 ` Matt Fleming
2019-10-09 12:39 ` Guenter Roeck
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=20190808195301.13222-3-matt@codeblueprint.co.uk \
--to=matt@codeblueprint.co.uk \
--cc=Suravee.Suthikulpanit@amd.com \
--cc=Thomas.Lendacky@amd.com \
--cc=bp@alien8.de \
--cc=linux-kernel@vger.kernel.org \
--cc=mgorman@techsingularity.net \
--cc=peterz@infradead.org \
--cc=riel@surriel.com \
--cc=tony.luck@intel.com \
/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;
as well as URLs for NNTP newsgroup(s).