From: K Prateek Nayak <kprateek.nayak@amd.com>
To: "Chen, Yu C" <yu.c.chen@intel.com>,
Tim Chen <tim.c.chen@linux.intel.com>,
Peter Zijlstra <peterz@infradead.org>
Cc: Pan Deng <pan.deng@intel.com>, <mingo@kernel.org>,
<linux-kernel@vger.kernel.org>, <tianyou.li@intel.com>
Subject: Re: [PATCH v2 4/4] sched/rt: Split cpupri_vec->cpumask to per NUMA node to reduce contention
Date: Wed, 8 Apr 2026 21:22:04 +0530 [thread overview]
Message-ID: <2dcbf93d-030d-466c-8b1c-8387513e9eb9@amd.com> (raw)
In-Reply-To: <14eda829-fc6b-407a-93a3-0794ab521177@intel.com>
Hello Chenyu,
On 4/8/2026 5:05 PM, Chen, Yu C wrote:
> We haven't tried breaking it down further. One possible approach
> is to partition it at L2 scope, the benefit of which may depend on
> the workload.
I fear at that point we'll have too many cachelines and too much
cache pollution when the CPU starts reading this at tick to schedule
a newidle balance.
A 128 core system would bring in 128 * 64B = 8kB worth of data to
traverse the mask and at that point it becomes a trade off between
how fast you want reads vs writes and does it even speed up writes
after a certain point?
Sorry I got distracted by some other stuff today but I'll share the
results from my experiments tomorrow.
--
Thanks and Regards,
Prateek
next prev parent reply other threads:[~2026-04-08 15:52 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-07-21 6:10 [PATCH v2 0/4] sched/rt: mitigate root_domain cache line contention Pan Deng
2025-07-21 6:10 ` [PATCH v2 1/4] sched/rt: Optimize cpupri_vec layout to mitigate " Pan Deng
2026-03-20 10:09 ` Peter Zijlstra
2026-03-24 9:36 ` Deng, Pan
2026-03-24 12:11 ` Peter Zijlstra
2026-03-27 10:17 ` Deng, Pan
2026-04-02 10:37 ` Deng, Pan
2026-04-02 10:43 ` Peter Zijlstra
2026-04-08 10:16 ` Chen, Yu C
2026-04-09 11:47 ` Deng, Pan
2025-07-21 6:10 ` [PATCH v2 2/4] sched/rt: Restructure root_domain to reduce cacheline contention Pan Deng
2026-03-20 10:18 ` Peter Zijlstra
2025-07-21 6:10 ` [PATCH v2 3/4] sched/rt: Split root_domain->rto_count to per-NUMA-node counters Pan Deng
2026-03-20 10:24 ` Peter Zijlstra
2026-03-23 18:09 ` Tim Chen
2026-03-24 12:16 ` Peter Zijlstra
2026-03-24 22:40 ` Tim Chen
2025-07-21 6:10 ` [PATCH v2 4/4] sched/rt: Split cpupri_vec->cpumask to per NUMA node to reduce contention Pan Deng
2026-03-20 12:40 ` Peter Zijlstra
2026-03-23 18:45 ` Tim Chen
2026-03-24 12:00 ` Peter Zijlstra
2026-03-31 5:37 ` Chen, Yu C
2026-03-31 10:19 ` K Prateek Nayak
2026-04-02 3:15 ` Chen, Yu C
2026-04-02 4:41 ` K Prateek Nayak
2026-04-02 10:55 ` Peter Zijlstra
2026-04-02 11:06 ` K Prateek Nayak
2026-04-03 5:46 ` Chen, Yu C
2026-04-03 8:13 ` K Prateek Nayak
2026-04-07 20:35 ` Tim Chen
2026-04-08 3:06 ` K Prateek Nayak
2026-04-08 11:35 ` Chen, Yu C
2026-04-08 15:52 ` K Prateek Nayak [this message]
2026-04-09 5:17 ` K Prateek Nayak
2026-04-09 23:09 ` Tim Chen
2026-04-10 5:51 ` Chen, Yu C
2026-04-10 6:02 ` K Prateek Nayak
2026-04-08 9:25 ` Chen, Yu C
2026-04-08 16:47 ` Tim Chen
2026-03-20 9:59 ` [PATCH v2 0/4] sched/rt: mitigate root_domain cache line contention Peter Zijlstra
2026-03-20 12:50 ` 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=2dcbf93d-030d-466c-8b1c-8387513e9eb9@amd.com \
--to=kprateek.nayak@amd.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=pan.deng@intel.com \
--cc=peterz@infradead.org \
--cc=tianyou.li@intel.com \
--cc=tim.c.chen@linux.intel.com \
--cc=yu.c.chen@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