Received-SPF: pass (google.com: domain of linux-kernel+bounces-174109-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Date: Thu, 9 May 2024 11:48:36 +0800
From: Chen Yu <yu.c.chen@intel.com>
To: Mike Galbraith <efault@gmx.de>
CC: Peter Zijlstra <peterz@infradead.org>, <mingo@redhat.com>,
	<juri.lelli@redhat.com>, <vincent.guittot@linaro.org>,
	<dietmar.eggemann@arm.com>, <rostedt@goodmis.org>, <bsegall@google.com>,
	<mgorman@suse.de>, <bristot@redhat.com>, <vschneid@redhat.com>,
	<linux-kernel@vger.kernel.org>, <kprateek.nayak@amd.com>,
	<wuyun.abel@bytedance.com>, <tglx@linutronix.de>, <tim.c.chen@intel.com>,
	<yu.c.chen.y@gmail.com>
Subject: Re: [RFC][PATCH 10/10] sched/eevdf: Use sched_attr::sched_runtime to
 set request/slice suggestion
Message-ID: <ZjxHlLQv1WuFq+SC@chenyu5-mobl2>
References: <20240405102754.435410987@infradead.org>
 <20240405110010.934104715@infradead.org>
 <ZjpFruUiBiNi6VSO@chenyu5-mobl2>
 <9c360c0d337b124c71095f06889d1c69279a7c06.camel@gmx.de>
Content-Type: text/plain; charset="us-ascii"
Content-Disposition: inline
In-Reply-To: <9c360c0d337b124c71095f06889d1c69279a7c06.camel@gmx.de>
Precedence: bulk
MIME-Version: 1.0
X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1
X-MS-Exchange-AntiSpam-MessageData-0: =?us-ascii?Q?Qc+B3n+noZLUUuY/TFCyFYFCUUgYmhN3fiJ6wV3b6fYJLRYsJ48ma0CYw3Os?=
 =?us-ascii?Q?tJ1vKoCQBeDsWmzsBhNUjPqyL7Ny8y3cExSnxQN6j4q6KYO+RYrajqtLoZoC?=
 =?us-ascii?Q?cxipYF6EOeU77rfUQalMH7u0F8P36JsaKYKI4T2yI3MSTCGludoJncK7PGoe?=
 =?us-ascii?Q?ShuskctSIh04CAEh4mxSraVDuzVjGouuzwI6pMGxSqRkopRtvQMAsack5mSR?=
 =?us-ascii?Q?xGXHoSWGbiGvnOxury+/9mhvIu8NsUvPR8yhb3AlD9xinIZw+7kVrDc6XZzz?=
 =?us-ascii?Q?0kW2u5EcagdP2OC2CPjbmuCkkBMvm+XxyX5ZNR85hxJiFkRMsrSJlzqsALcV?=
 =?us-ascii?Q?z/Ub/IrsD2GLEr62Sq7CcVBWhqUphqms38KrPgDA3VEqL3blvhtnqdYrwCOX?=
 =?us-ascii?Q?nRTZ+LGphozSW9kgHQsiXhqbjsZ29q05smQ7fbsK+hnPUJr5CK2r1BJmaZwB?=
 =?us-ascii?Q?cJJcNia1EAXB2hrhEwQA+cxgEcUK2l+yDJ6pzc6x/oD7ptGpaSbMU6qNT4Tt?=
 =?us-ascii?Q?crJ2WqGBqn8dnRlZ1qgY0gv2IUjoXUc8pxj1o4UK29gARoiqyNeAkFW4x+BO?=
 =?us-ascii?Q?vhpDayXeadoWs0AIBbrwZvZtiY8Kz/o0FTcwyksqs3mUQqZG4SGzPKXAwQii?=
 =?us-ascii?Q?yF5VHs6JbqJqtD0PLY5dBHlbsdFk3eoXjPgG1FCSlGexvIV7zqZLH+GM9bXr?=
 =?us-ascii?Q?ykCQOF5Tj23xyk/RMvlCBJvk+7ITmT/Z4p/vLp3TmaQFzBp7SyGaf3WoTI1d?=
 =?us-ascii?Q?AsxiY3X++QXwO46bDs85yPIxOz2Il2tM12fhGzT5cAHEuHQl8GMZIxmmFB7V?=
 =?us-ascii?Q?9ngFqyHwTimusUL5HOIIEtd/yNfGohjOQIL9XMJ571hfxe/42jCBDrzOfW7t?=
 =?us-ascii?Q?TKwZ+9yae0W5+yILN2ZvpAFPl9B6L8+G5n9WMXX0Mkishr1GV0sZZKLxK6T+?=
 =?us-ascii?Q?qTBfeSz6dm+UXU1GRKwdCLxlbiG261o1OJ/700QwJqCJiex8ulNdwyCzU4j2?=
 =?us-ascii?Q?2Ek+YTiDhKs8YzAjCUGNUM5+lDjFQC8HZpfrSRSU17FEAh3ucP8HSfP3fbe2?=
 =?us-ascii?Q?sWAgKCPxYn4ii8Gc9j1Ycwds3Qw7F8xu46TH2AsQ2gG9gnY7960xpJ4Cf/na?=
 =?us-ascii?Q?i3pYOFzmDkk0dWYazeGtmS5KR5QDEFY211ZygzUBCQikUzNmCvK/NtbBGiqQ?=
 =?us-ascii?Q?hEuaI0cmhAOKJa1rc8r+L9HMNjla1xc3AXCZ7X3zdD+dJOQ+YxPg3BWlPgS2?=
 =?us-ascii?Q?Gv3QNCA1apsBFqsixPTNw2iViftojrKVaXHh7egB/1K9T5uf8YjBGBPUnEzV?=
 =?us-ascii?Q?e0uHtcOt56IlEhAUlKz9f6LGNaO/Mhw7whYDs44ryfgB2gKD/+r8qCcanM/N?=
 =?us-ascii?Q?TfiJxRPhgoP7moQHfenewWXWWXGNtrNE+eS2FltgvTr923VSOdBq8nrcKKDm?=
 =?us-ascii?Q?R2KPdu0CZl7xRI8lCFwDnKdISa1HarCv/JKObuCuGmomIrr2krdgU23TF2Nb?=
 =?us-ascii?Q?1bDOxDuQohFiPjcK2xHAlMx2yMkzjWdNlG8FmtooCEt7dPsHRyHiOk/ob6YN?=
 =?us-ascii?Q?z6IrTgwCE9CVp3w+6CMIl7XeZ3gVGyIpgbUk6pOr?=
X-MS-Exchange-CrossTenant-Network-Message-Id: 9f01a0da-2692-4f9b-5617-08dc6fdaf009
X-MS-Exchange-CrossTenant-AuthSource: DM4PR11MB6020.namprd11.prod.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Internal
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 09 May 2024 03:48:50.1827
 (UTC)
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-CrossTenant-Id: 46c98d88-e344-4ed4-8496-4ed7712e255d
X-MS-Exchange-CrossTenant-MailboxType: HOSTED
X-MS-Exchange-CrossTenant-UserPrincipalName: UDFGRnJt9aocZ5cqYiq6NP61UdKqho/qsDKXtVj+e+561ZfHYeHG7HTqFg9oLLCH5zTCj/2C9PuH8iQ7X5FQog==
X-MS-Exchange-Transport-CrossTenantHeadersStamped: DS0PR11MB6446
X-OriginatorOrg: intel.com

On 2024-05-08 at 15:52:32 +0200, Mike Galbraith wrote:
> On Tue, 2024-05-07 at 23:15 +0800, Chen Yu wrote:
> > On 2024-04-05 at 12:28:04 +0200, Peter Zijlstra wrote:
> > > Allow applications to directly set a suggested request/slice length using
> > > sched_attr::sched_runtime.
> > >
> > > The implementation clamps the value to: 0.1[ms] <= slice <= 100[ms]
> > > which is 1/10 the size of HZ=1000 and 10 times the size of HZ=100.
> > >
> > > Applications should strive to use their periodic runtime at a high
> > > confidence interval (95%+) as the target slice. Using a smaller slice
> > > will introduce undue preemptions, while using a larger value will
> > > increase latency.
> > >
> > > Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> > >
> >
> > Is it possible to leverage this task slice to do better task wakeup placement?
> 
> Slice being unrelated to placement makes its use in a placement related
> knob look wrong.  Even the smallest possible slice is orders of
> magnitude larger than the cycle time of TCP_RR, making slice nearly
> irrelevant to the issue being demonstrating via TCP_RR.

Yes, I agree that there is no direct relationship between the slice and
the task placement. The idea is to use slice as an input hint from the user
to tell the kernel how much latency this user cares(but not expect the task
to last for that long).

> Even for that huge socket box, it won't take long as cycle time increases toward that
> smallest possible slice for the cost of needless wait to bury placement
> decision costs.
>

I see. The wake up latency is composed of:
1. The time to do placement decision.
2. The time wait in the runqueue to be picked.

Even if the 1 has been reduced, the 2 might take more time if the runqueue
is busy. We can mitigate this by checking if there is <= 1 short duration task
on that target CPU. 
 
> > The idea is that, the smaller the slice the wakee has, the less idle CPU it
> > should scan. This can reduce wake latency and inhibit costly task migration,
> > especially on large systems.
> 
> Sure, this is an age old issue that's scaled up to size extra ugly in
> that huge socket box.  Any solution needs to scale as well methinks, a
> simple fixed yardstick won't work, as the costs being mitigated vary
> wildly with platform size/shape.
>

Understand, this slice based placement was chosen for experimental purpose,
and seek for directions. And I agree we should take the platform size(such
as CPU number) into consideration. 

thanks,
Chenyu