From: Kayra Cizmeci <kayracizmeci@gmail.com>
To: vincent.guittot@linaro.org
Cc: arighi@nvidia.com, bsegall@google.com, changwoo@igalia.com,
christian.loehle@arm.com, dietmar.eggemann@arm.com,
juri.lelli@redhat.com, kprateek.nayak@amd.com,
linux-kernel@vger.kernel.org, linux-pm@vger.kernel.org,
lukasz.luba@arm.com, mgorman@suse.de, mingo@redhat.com,
peterz@infradead.org, pierre.gondois@arm.com,
qyousef@layalina.io, rafael@kernel.org, rostedt@goodmis.org,
sched-ext@lists.linux.dev, sshegde@linux.ibm.com, tj@kernel.org,
void@manifault.com, vschneid@redhat.com,
Kayra Cizmeci <kayracizmeci@gmail.com>
Subject: Re: [PATCH 00/18 v2] Improving latency of short slice tasks
Date: Thu, 8 Oct 2026 22:12:36 +0300 [thread overview]
Message-ID: <20261008191236.152916-1-kayracizmeci@gmail.com> (raw)
In-Reply-To: <20261002154415.2270586-1-vincent.guittot@linaro.org>
Hi All,
I wanted to ran some tests for this patch series. I couldn't really
apply it without some issues, but I think I managed to apply
it cleanly this time... Tho I may have been doing something wrong.
Please let me know if so.
base: 4a3b51aab6e25244d97936aa65e6d5425adf98e1 (on tip/sched/core)
Command Used Each Run:
hackbench -g 4 -l 100000 > /dev/null & sudo chrt -o -T 300000 0 ~/rt-tests/cyclictest --policy=other -t 1 -i 1000 -D 20 -q -m; wait
Environment:
Zen 3 with 8 Cores and 16 Threads. And Fedora.
THE RESULTS (COOKED):
On Base in 5 runs, average values are: Min: 3.8, Act: 5.8, Avg: 8.8 and
Max: 2511.6 (I need tea.)
On Base + Patches in 5 runs, average values bla bla you get it:
Min 4, Act: 5.8, Avg: 7.2 and Max: 1781.6
Or, in the shape of art made with love and passion:
+------BASE-----+--PATCHED--+
| MIN 3.8 | 4 |
| ACT 5.8 | 5.8 |
| AVG 8.8 | 7.2 |
| MAX 2511.6 | 1781.6 |
+--------------+------------+
Hmhp... Hope I'm not missing something. I need water. Bah.
Tested-by: Kayra Cizmeci <kayracizmeci@gmail.com>
Dancing donkey chases a car while shouting "Linux is the best, my dancing is the coolest. Cry."
Thanks,
Kayra
I'll add other not too important things below, these are pretty raw.
But if you wanna have a look you can.
THE RESULTS (RAW):
BASE 7.3.0-rc1+ #2:
RUN 1:
[1] 1950
# /dev/cpu_dma_latency set to 0us
T: 0 ( 2115) P: 0 I:1000 C: 19990 Min: 4 Act: 6 Avg: 8 Max: 1885
[1]+ Done hackbench -g 4 -l 100000 > /dev/null
RUN 2:
[1] 2132
# /dev/cpu_dma_latency set to 0us
T: 0 ( 2297) P: 0 I:1000 C: 19990 Min: 4 Act: 6 Avg: 9 Max: 2695
[1]+ Done hackbench -g 4 -l 100000 > /dev/null
linuxfucks@micro-softly:~$ (Opsies. Who's hostname is this? It's so cool. Wow. 'linuxfucks')
RUN 3:
[1] 2307
# /dev/cpu_dma_latency set to 0us
T: 0 ( 2472) P: 0 I:1000 C: 19988 Min: 4 Act: 5 Avg: 8 Max: 3246
[1]+ Done hackbench -g 4 -l 100000 > /dev/null
RUN 4:
[1] 2493
# /dev/cpu_dma_latency set to 0us
T: 0 ( 2658) P: 0 I:1000 C: 19995 Min: 3 Act: 6 Avg: 9 Max: 1991
[1]+ Done hackbench -g 4 -l 100000 > /dev/null
RUN 5:
[1] 2671
# /dev/cpu_dma_latency set to 0us
T: 0 ( 2836) P: 0 I:1000 C: 19992 Min: 4 Act: 6 Avg: 10 Max: 2741
[1]+ Done hackbench -g 4 -l 100000 > /dev/null
BASE + PATCHES 7.3.0-rc1+ #3:
RUN 1:
[1] 1490
# /dev/cpu_dma_latency set to 0us
T: 0 ( 1657) P: 0 I:1000 C: 19993 Min: 4 Act: 5 Avg: 7 Max: 3091
[1]+ Done hackbench -g 4 -l 100000 > /dev/null
RUN 2:
[1] 1671
# /dev/cpu_dma_latency set to 0us
T: 0 ( 1836) P: 0 I:1000 C: 19999 Min: 4 Act: 6 Avg: 7 Max: 1529
[1]+ Done hackbench -g 4 -l 100000 > /dev/null
RUN 3:
[1] 1848
# /dev/cpu_dma_latency set to 0us
T: 0 ( 2013) P: 0 I:1000 C: 19998 Min: 4 Act: 6 Avg: 7 Max: 1455
[1]+ Done hackbench -g 4 -l 100000 > /dev/null
RUN 4:
[1] 2023
# /dev/cpu_dma_latency set to 0us
T: 0 ( 2188) P: 0 I:1000 C: 19995 Min: 4 Act: 6 Avg: 7 Max: 1524
[1]+ Done hackbench -g 4 -l 100000 > /dev/null
RUN 5:
[1] 2200
# /dev/cpu_dma_latency set to 0us
T: 0 ( 2365) P: 0 I:1000 C: 19989 Min: 4 Act: 6 Avg: 8 Max: 1309
[1]+ Done hackbench -g 4 -l 100000 > /dev/null
next prev parent reply other threads:[~2026-10-08 19:13 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-10-02 15:43 [PATCH 00/18 v2] Improving latency of short slice tasks Vincent Guittot
2026-10-02 15:43 ` [PATCH 01/18 v2] sched/eevdf: Decay positive lag of sleeping entities Vincent Guittot
2026-10-02 15:43 ` [PATCH 02/18] sched/eevdf: Reset lag when waking up on idle cpu Vincent Guittot
2026-10-04 17:30 ` Kayra Cizmeci
2026-10-09 13:35 ` Vincent Guittot
2026-10-02 15:44 ` [PATCH 03/18 v2] sched/eevdf: Add per cpu cached min_slice Vincent Guittot
2026-10-02 15:44 ` [PATCH 04/18 v2] sched/eevdf: Compare min slice during wake_affine Vincent Guittot
2026-10-05 15:58 ` Kayra Cizmeci
2026-10-02 15:44 ` [PATCH 05/18 v2] sched/eevdf: Add min slice check when selecting CPU Vincent Guittot
2026-10-04 19:18 ` Kayra Cizmeci
2026-10-02 15:44 ` [PATCH 06/18 v2] sched/fair: Prepare select_task_rq_fair() to be called for new cases Vincent Guittot
2026-10-06 22:25 ` Tim Chen
2026-10-09 13:36 ` Vincent Guittot
2026-10-02 15:44 ` [PATCH 07/18] sched/fair: Add push task mechanism for fair Vincent Guittot
2026-10-07 2:45 ` Chen Yu
2026-10-09 13:38 ` Vincent Guittot
2026-10-09 14:31 ` Vincent Guittot
2026-10-02 15:44 ` [PATCH 08/18] sched/fair: Optimize " Vincent Guittot
2026-10-02 15:44 ` [PATCH 09/18 v2] sched/core: Add rq flag to tick parameters Vincent Guittot
2026-10-02 15:44 ` [PATCH 10/18 v2] sched/fair: Add force push task mechanism for fair Vincent Guittot
2026-10-06 19:24 ` Kayra Cizmeci
2026-10-09 14:01 ` Vincent Guittot
2026-10-09 5:53 ` Kayra Cizmeci
2026-10-09 14:04 ` Vincent Guittot
2026-10-09 14:47 ` Vincent Guittot
2026-10-02 15:44 ` [PATCH 11/18 v2] sched/fair: Support not wakeup case in select_idle_sibling Vincent Guittot
2026-10-07 17:55 ` Tim Chen
2026-10-09 14:06 ` Vincent Guittot
2026-10-02 15:44 ` [PATCH 12/18 v2] sched/eevdf: Try to push short slice task on a better CPU Vincent Guittot
2026-10-02 15:44 ` [PATCH 13/18 v2] sched/eevdf: Push short slice task that are not picked Vincent Guittot
2026-10-02 15:44 ` [PATCH 14/18 v2] sched/fair: Enable push task for preempt short Vincent Guittot
2026-10-09 11:16 ` Kayra Cizmeci
2026-10-09 14:07 ` Vincent Guittot
2026-10-02 15:44 ` [PATCH 15/18 v2] energy model: Add a get previous state function Vincent Guittot
2026-10-02 15:44 ` [PATCH 16/18 v2] sched/fair: Rework feec() to use cost instead of spare capacity Vincent Guittot
2026-10-02 15:44 ` [PATCH 17/18 v2] energy model: Remove unused em_cpu_energy() Vincent Guittot
2026-10-02 15:44 ` [PATCH 18/18 v2] sched/fair: Take into account slice in EAS Vincent Guittot
2026-10-07 14:36 ` Kayra Cizmeci
2026-10-08 22:53 ` Tim Chen
2026-10-09 6:31 ` Kayra Cizmeci
2026-10-09 14:17 ` Vincent Guittot
2026-10-08 19:12 ` Kayra Cizmeci [this message]
2026-10-09 14:56 ` [PATCH 00/18 v2] Improving latency of short slice tasks Vincent Guittot
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=20261008191236.152916-1-kayracizmeci@gmail.com \
--to=kayracizmeci@gmail.com \
--cc=arighi@nvidia.com \
--cc=bsegall@google.com \
--cc=changwoo@igalia.com \
--cc=christian.loehle@arm.com \
--cc=dietmar.eggemann@arm.com \
--cc=juri.lelli@redhat.com \
--cc=kprateek.nayak@amd.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=lukasz.luba@arm.com \
--cc=mgorman@suse.de \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=pierre.gondois@arm.com \
--cc=qyousef@layalina.io \
--cc=rafael@kernel.org \
--cc=rostedt@goodmis.org \
--cc=sched-ext@lists.linux.dev \
--cc=sshegde@linux.ibm.com \
--cc=tj@kernel.org \
--cc=vincent.guittot@linaro.org \
--cc=void@manifault.com \
--cc=vschneid@redhat.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