From: Alex Deucher <alexdeucher@gmail.com>
To: Tejun Heo <tj@kernel.org>
Cc: Song Liu <songliubraving@fb.com>,
Alexei Starovoitov <ast@kernel.org>,
DRI Development <dri-devel@lists.freedesktop.org>,
Alexei Starovoitov <alexei.starovoitov@gmail.com>,
Daniel Borkmann <daniel@iogearbox.net>,
Brian Welty <brian.welty@intel.com>,
John Fastabend <john.fastabend@gmail.com>,
amd-gfx list <amd-gfx@lists.freedesktop.org>,
Yonghong Song <yhs@fb.com>, Andrii Nakryiko <andriin@fb.com>,
Linux-Fsdevel <linux-fsdevel@vger.kernel.org>,
Kenny Ho <y2kenny@gmail.com>,
Alexander Viro <viro@zeniv.linux.org.uk>,
KP Singh <kpsingh@chromium.org>,
"open list:CONTROL GROUP \(CGROUP\)" <cgroups@vger.kernel.org>,
Kenny Ho <Kenny.Ho@amd.com>,
Network Development <netdev@vger.kernel.org>,
Alex Deucher <alexander.deucher@amd.com>,
bpf <bpf@vger.kernel.org>, Martin KaFai Lau <kafai@fb.com>
Subject: Re: [RFC] Add BPF_PROG_TYPE_CGROUP_IOCTL
Date: Fri, 7 May 2021 15:55:39 -0400 [thread overview]
Message-ID: <CADnq5_Mwd-xHZQ4pt34=FPk2Gq3ij1FNHWsEz1LdS7_Dyo00iQ@mail.gmail.com> (raw)
In-Reply-To: <YJWWByISHSPqF+aN@slm.duckdns.org>
On Fri, May 7, 2021 at 3:33 PM Tejun Heo <tj@kernel.org> wrote:
>
> Hello,
>
> On Fri, May 07, 2021 at 06:54:13PM +0200, Daniel Vetter wrote:
> > All I meant is that for the container/cgroups world starting out with
> > time-sharing feels like the best fit, least because your SRIOV designers
> > also seem to think that's the best first cut for cloud-y computing.
> > Whether it's virtualized or containerized is a distinction that's getting
> > ever more blurry, with virtualization become a lot more dynamic and
> > container runtimes als possibly using hw virtualization underneath.
>
> FWIW, I'm completely on the same boat. There are two fundamental issues with
> hardware-mask based control - control granularity and work conservation.
> Combined, they make it a significantly more difficult interface to use which
> requires hardware-specific tuning rather than simply being able to say "I
> wanna prioritize this job twice over that one".
>
> My knoweldge of gpus is really limited but my understanding is also that the
> gpu cores and threads aren't as homogeneous as the CPU counterparts across
> the vendors, product generations and possibly even within a single chip,
> which makes the problem even worse.
>
> Given that GPUs are time-shareable to begin with, the most universal
> solution seems pretty clear.
The problem is temporal partitioning on GPUs is much harder to enforce
unless you have a special case like SR-IOV. Spatial partitioning, on
AMD GPUs at least, is widely available and easily enforced. What is
the point of implementing temporal style cgroups if no one can enforce
it effectively?
Alex
>
> Thanks.
>
> --
> tejun
next prev parent reply other threads:[~2021-05-07 19:55 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20201007152355.2446741-1-Kenny.Ho@amd.com>
[not found] ` <CAOWid-d=a1Q3R92s7GrzxWhXx7_dc8NQvQg7i7RYTVv3+jHxkQ@mail.gmail.com>
[not found] ` <20201103053244.khibmr66p7lhv7ge@ast-mbp.dhcp.thefacebook.com>
[not found] ` <CAOWid-eQSPru0nm8+Xo3r6C0pJGq+5r8mzM8BL2dgNn2c9mt2Q@mail.gmail.com>
[not found] ` <CAADnVQKuoZDB-Xga5STHdGSxvSP=B6jQ40kLdpL1u+J98bv65A@mail.gmail.com>
[not found] ` <CAOWid-czZphRz6Y-H3OcObKCH=bLLC3=bOZaSB-6YBE56+Qzrg@mail.gmail.com>
[not found] ` <20201103210418.q7hddyl7rvdplike@ast-mbp.dhcp.thefacebook.com>
[not found] ` <CAOWid-djQ_NRfCbOTnZQ-A8Pr7jMP7KuZEJDSsvzWkdw7qc=yA@mail.gmail.com>
[not found] ` <20201103232805.6uq4zg3gdvw2iiki@ast-mbp.dhcp.thefacebook.com>
2021-02-01 14:49 ` [RFC] Add BPF_PROG_TYPE_CGROUP_IOCTL Daniel Vetter
2021-02-01 16:47 ` Kenny Ho
2021-02-01 16:51 ` Kenny Ho
2021-02-03 11:09 ` Daniel Vetter
2021-02-03 19:01 ` Kenny Ho
2021-02-05 13:49 ` Daniel Vetter
2021-05-07 2:06 ` Kenny Ho
2021-05-07 8:59 ` Daniel Vetter
2021-05-07 15:33 ` Kenny Ho
2021-05-07 16:13 ` Daniel Vetter
2021-05-07 16:19 ` Alex Deucher
2021-05-07 16:26 ` Daniel Vetter
2021-05-07 16:31 ` Alex Deucher
2021-05-07 16:50 ` Alex Deucher
2021-05-07 16:54 ` Daniel Vetter
2021-05-07 17:04 ` Kenny Ho
2021-05-07 19:33 ` Tejun Heo
2021-05-07 19:55 ` Alex Deucher [this message]
2021-05-07 20:59 ` Tejun Heo
2021-05-07 22:30 ` Alex Deucher
2021-05-07 23:45 ` Tejun Heo
2021-05-11 15:48 ` Alex Deucher
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='CADnq5_Mwd-xHZQ4pt34=FPk2Gq3ij1FNHWsEz1LdS7_Dyo00iQ@mail.gmail.com' \
--to=alexdeucher@gmail.com \
--cc=Kenny.Ho@amd.com \
--cc=alexander.deucher@amd.com \
--cc=alexei.starovoitov@gmail.com \
--cc=amd-gfx@lists.freedesktop.org \
--cc=andriin@fb.com \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=brian.welty@intel.com \
--cc=cgroups@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=dri-devel@lists.freedesktop.org \
--cc=john.fastabend@gmail.com \
--cc=kafai@fb.com \
--cc=kpsingh@chromium.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=songliubraving@fb.com \
--cc=tj@kernel.org \
--cc=viro@zeniv.linux.org.uk \
--cc=y2kenny@gmail.com \
--cc=yhs@fb.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;
as well as URLs for NNTP newsgroup(s).