[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <2f73f1d9-ad0c-40db-9697-3070cd52f688@gmail.com>
Date: Thu, 31 Jul 2025 12:55:29 +0100
From: Usama Arif <usamaarif642@...il.com>
To: Casey Chen <cachen@...estorage.com>, akpm@...ux-foundation.org,
surenb@...gle.com
Cc: kent.overstreet@...ux.dev, corbet@....net, dennis@...nel.org,
tj@...nel.org, cl@...two.org, vbabka@...e.cz, mhocko@...e.com,
jackmanb@...gle.com, hannes@...xchg.org, ziy@...dia.com,
rientjes@...gle.com, roman.gushchin@...ux.dev, harry.yoo@...cle.com,
linux-mm@...ck.org, linux-kernel@...r.kernel.org, linux-doc@...r.kernel.org,
yzhong@...estorage.com, souravpanda@...gle.com, 00107082@....com,
pyy@...a.com
Subject: Re: [PATCH v3] alloc_tag: add per-NUMA node stats
On 11/07/2025 01:23, Casey Chen wrote:
> This patch adds per-NUMA node breakdown of memory allocation,
> enabling more precise visibility into memory usage patterns across nodes.
> It is particularly valuable in cloud environments,
> where tracking asymmetric memory usage and identifying NUMA imbalances
> down to the allocation caller helps optimize memory efficiency, avoid
> CPU stranding, and improve system responsiveness under memory pressure.
>
Hi Casey,
I was just curious how you used the output from per NUMA allocation profiling
to fix the issues above? (So that we can do that as well :)). We get a lot of
NUMA high level meminfo from /sys/devices/system/node/nodeX/vmstat.
Are there specific allocations that are more of an issue?
Do you monitor these in userspace and use maybe something like numactl/migratepages
to migrate resources to another node?
Thanks!
Usama
Powered by blists - more mailing lists