From: Tejun Heo <tj@kernel.org>
To: Kenny Ho <y2kenny@gmail.com>
Cc: "Kenny Ho" <Kenny.Ho@amd.com>,
"Kuehling, Felix" <felix.kuehling@amd.com>,
jsparks@cray.com, "amd-gfx list" <amd-gfx@lists.freedesktop.org>,
lkaplan@cray.com, dri-devel <dri-devel@lists.freedesktop.org>,
"Greathouse, Joseph" <joseph.greathouse@amd.com>,
"Alex Deucher" <alexander.deucher@amd.com>,
cgroups@vger.kernel.org,
"Christian König" <christian.koenig@amd.com>
Subject: Re: [PATCH v2 00/11] new cgroup controller for gpu/drm subsystem
Date: Mon, 13 Apr 2020 17:53:30 -0400 [thread overview]
Message-ID: <20200413215330.GN60335@mtj.duckdns.org> (raw)
In-Reply-To: <CAOWid-fvmxSXtGUtQSZ4Ow1fK+wR8hbnUe5PcsM56EZMOMwb6g@mail.gmail.com>
Hello,
On Mon, Apr 13, 2020 at 05:40:32PM -0400, Kenny Ho wrote:
> By lack of consense, do you mean Intel's assertion that a standard is
> not a standard until Intel implements it? (That was in the context of
> OpenCL language standard with the concept of SubDevice.) I thought
> the discussion so far has established that the concept of a compute
> unit, while named differently (AMD's CUs, ARM's SCs, Intel's EUs,
> Nvidia's SMs, Qualcomm's SPs), is cross vendor. While an AMD CU is
> not the same as an Intel EU or Nvidia SM, the same can be said for CPU
> cores. If cpuset is acceptable for a diversity of CPU core designs
> and arrangements, I don't understand why an interface derived from GPU
> SubDevice is considered premature.
CPUs are a lot more uniform across vendors than GPUs and have way higher user
observability and awareness. And, even then, it's something which has limited
usefulness because the configuration is inherently more complex involving
topology details and the end result is not work-conserving.
cpuset is there partly due to historical reasons and its features can often be
trivially replicated with some scripting around taskset. If that's all you're
trying to add, I don't see why it needs to be in cgroup at all. Just implement
a tool similar to taskset and build sufficient tooling around it. Given how
hardware specific it can become, that is likely the better direction anyway.
Thanks.
--
tejun
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
WARNING: multiple messages have this Message-ID (diff)
From: Tejun Heo <tj@kernel.org>
To: Kenny Ho <y2kenny@gmail.com>
Cc: "Kenny Ho" <Kenny.Ho@amd.com>,
"Kuehling, Felix" <felix.kuehling@amd.com>,
jsparks@cray.com, "amd-gfx list" <amd-gfx@lists.freedesktop.org>,
lkaplan@cray.com, dri-devel <dri-devel@lists.freedesktop.org>,
"Greathouse, Joseph" <joseph.greathouse@amd.com>,
"Alex Deucher" <alexander.deucher@amd.com>,
cgroups@vger.kernel.org,
"Christian König" <christian.koenig@amd.com>
Subject: Re: [PATCH v2 00/11] new cgroup controller for gpu/drm subsystem
Date: Mon, 13 Apr 2020 17:53:30 -0400 [thread overview]
Message-ID: <20200413215330.GN60335@mtj.duckdns.org> (raw)
In-Reply-To: <CAOWid-fvmxSXtGUtQSZ4Ow1fK+wR8hbnUe5PcsM56EZMOMwb6g@mail.gmail.com>
Hello,
On Mon, Apr 13, 2020 at 05:40:32PM -0400, Kenny Ho wrote:
> By lack of consense, do you mean Intel's assertion that a standard is
> not a standard until Intel implements it? (That was in the context of
> OpenCL language standard with the concept of SubDevice.) I thought
> the discussion so far has established that the concept of a compute
> unit, while named differently (AMD's CUs, ARM's SCs, Intel's EUs,
> Nvidia's SMs, Qualcomm's SPs), is cross vendor. While an AMD CU is
> not the same as an Intel EU or Nvidia SM, the same can be said for CPU
> cores. If cpuset is acceptable for a diversity of CPU core designs
> and arrangements, I don't understand why an interface derived from GPU
> SubDevice is considered premature.
CPUs are a lot more uniform across vendors than GPUs and have way higher user
observability and awareness. And, even then, it's something which has limited
usefulness because the configuration is inherently more complex involving
topology details and the end result is not work-conserving.
cpuset is there partly due to historical reasons and its features can often be
trivially replicated with some scripting around taskset. If that's all you're
trying to add, I don't see why it needs to be in cgroup at all. Just implement
a tool similar to taskset and build sufficient tooling around it. Given how
hardware specific it can become, that is likely the better direction anyway.
Thanks.
--
tejun
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
WARNING: multiple messages have this Message-ID (diff)
From: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
To: Kenny Ho <y2kenny-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Cc: "Kenny Ho" <Kenny.Ho-5C7GfCeVMHo@public.gmane.org>,
"Kuehling, Felix" <felix.kuehling-5C7GfCeVMHo@public.gmane.org>,
jsparks-WVYJKLFxKCc@public.gmane.org,
"amd-gfx list"
<amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org>,
lkaplan-WVYJKLFxKCc@public.gmane.org,
dri-devel
<dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org>,
"Greathouse,
Joseph" <joseph.greathouse-5C7GfCeVMHo@public.gmane.org>,
"Alex Deucher" <alexander.deucher-5C7GfCeVMHo@public.gmane.org>,
cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
"Christian König" <christian.koenig-5C7GfCeVMHo@public.gmane.org>
Subject: Re: [PATCH v2 00/11] new cgroup controller for gpu/drm subsystem
Date: Mon, 13 Apr 2020 17:53:30 -0400 [thread overview]
Message-ID: <20200413215330.GN60335@mtj.duckdns.org> (raw)
In-Reply-To: <CAOWid-fvmxSXtGUtQSZ4Ow1fK+wR8hbnUe5PcsM56EZMOMwb6g-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
Hello,
On Mon, Apr 13, 2020 at 05:40:32PM -0400, Kenny Ho wrote:
> By lack of consense, do you mean Intel's assertion that a standard is
> not a standard until Intel implements it? (That was in the context of
> OpenCL language standard with the concept of SubDevice.) I thought
> the discussion so far has established that the concept of a compute
> unit, while named differently (AMD's CUs, ARM's SCs, Intel's EUs,
> Nvidia's SMs, Qualcomm's SPs), is cross vendor. While an AMD CU is
> not the same as an Intel EU or Nvidia SM, the same can be said for CPU
> cores. If cpuset is acceptable for a diversity of CPU core designs
> and arrangements, I don't understand why an interface derived from GPU
> SubDevice is considered premature.
CPUs are a lot more uniform across vendors than GPUs and have way higher user
observability and awareness. And, even then, it's something which has limited
usefulness because the configuration is inherently more complex involving
topology details and the end result is not work-conserving.
cpuset is there partly due to historical reasons and its features can often be
trivially replicated with some scripting around taskset. If that's all you're
trying to add, I don't see why it needs to be in cgroup at all. Just implement
a tool similar to taskset and build sufficient tooling around it. Given how
hardware specific it can become, that is likely the better direction anyway.
Thanks.
--
tejun
next prev parent reply other threads:[~2020-04-13 21:53 UTC|newest]
Thread overview: 90+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <lkaplan@cray.com; daniel@ffwll.ch; nirmoy.das@amd.com; damon.mcdougall@amd.com; juan.zuniga-anaya@amd.com; hannes@cmpxchg.org>
2020-02-26 19:01 ` [PATCH v2 00/11] new cgroup controller for gpu/drm subsystem Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` [PATCH v2 01/11] cgroup: Introduce cgroup for drm subsystem Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` [PATCH v2 02/11] drm, cgroup: Bind drm and cgroup subsystem Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` [PATCH v2 03/11] drm, cgroup: Initialize drmcg properties Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` [PATCH v2 04/11] drm, cgroup: Add total GEM buffer allocation stats Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` [PATCH v2 05/11] drm, cgroup: Add peak " Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` [PATCH v2 06/11] drm, cgroup: Add GEM buffer allocation count stats Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` [PATCH v2 07/11] drm, cgroup: Add total GEM buffer allocation limit Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` [PATCH v2 08/11] drm, cgroup: Add peak " Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` [PATCH v2 09/11] drm, cgroup: Add compute as gpu cgroup resource Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` [PATCH v2 10/11] drm, cgroup: add update trigger after limit change Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` [PATCH v2 11/11] drm/amdgpu: Integrate with DRM cgroup Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-02-26 19:01 ` Kenny Ho
2020-03-17 16:03 ` [PATCH v2 00/11] new cgroup controller for gpu/drm subsystem Kenny Ho
2020-03-17 16:03 ` Kenny Ho
2020-03-17 16:03 ` Kenny Ho
2020-03-24 18:46 ` Tejun Heo
2020-03-24 18:46 ` Tejun Heo
2020-03-24 18:46 ` Tejun Heo
2020-03-24 18:49 ` Kenny Ho
2020-03-24 18:49 ` Kenny Ho
2020-03-24 18:49 ` Kenny Ho
2020-04-13 19:11 ` Tejun Heo
2020-04-13 19:11 ` Tejun Heo
2020-04-13 19:11 ` Tejun Heo
2020-04-13 20:12 ` Ho, Kenny
2020-04-13 20:12 ` Ho, Kenny
2020-04-13 20:12 ` Ho, Kenny
2020-04-13 20:17 ` Kenny Ho
2020-04-13 20:17 ` Kenny Ho
2020-04-13 20:17 ` Kenny Ho
2020-04-13 20:54 ` Tejun Heo
2020-04-13 20:54 ` Tejun Heo
2020-04-13 20:54 ` Tejun Heo
2020-04-13 21:40 ` Kenny Ho
2020-04-13 21:40 ` Kenny Ho
2020-04-13 21:40 ` Kenny Ho
2020-04-13 21:53 ` Tejun Heo [this message]
2020-04-13 21:53 ` Tejun Heo
2020-04-13 21:53 ` Tejun Heo
2020-04-14 12:20 ` Daniel Vetter
2020-04-14 12:20 ` Daniel Vetter
2020-04-14 12:20 ` Daniel Vetter
2020-04-14 12:47 ` Kenny Ho
2020-04-14 12:47 ` Kenny Ho
2020-04-14 12:47 ` Kenny Ho
2020-04-14 12:52 ` Daniel Vetter
2020-04-14 12:52 ` Daniel Vetter
2020-04-14 12:52 ` Daniel Vetter
2020-04-14 13:14 ` Kenny Ho
2020-04-14 13:14 ` Kenny Ho
2020-04-14 13:14 ` Kenny Ho
2020-04-14 13:26 ` Daniel Vetter
2020-04-14 13:26 ` Daniel Vetter
2020-04-14 13:26 ` Daniel Vetter
2020-04-14 13:50 ` Kenny Ho
2020-04-14 13:50 ` Kenny Ho
2020-04-14 13:50 ` Kenny Ho
2020-04-14 14:04 ` Daniel Vetter
2020-04-14 14:04 ` Daniel Vetter
2020-04-14 14:04 ` Daniel Vetter
2020-04-14 14:29 ` Kenny Ho
2020-04-14 14:29 ` Kenny Ho
2020-04-14 14:29 ` Kenny Ho
2020-04-14 15:01 ` Daniel Vetter
2020-04-14 15:01 ` Daniel Vetter
2020-04-14 15:01 ` Daniel Vetter
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=20200413215330.GN60335@mtj.duckdns.org \
--to=tj@kernel.org \
--cc=Kenny.Ho@amd.com \
--cc=alexander.deucher@amd.com \
--cc=amd-gfx@lists.freedesktop.org \
--cc=cgroups@vger.kernel.org \
--cc=christian.koenig@amd.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=felix.kuehling@amd.com \
--cc=joseph.greathouse@amd.com \
--cc=jsparks@cray.com \
--cc=lkaplan@cray.com \
--cc=y2kenny@gmail.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.