From: Rodrigo Vivi <rodrigo.vivi@intel.com>
To: "Tvrtko Ursulin" <tursulin@ursulin.net>,
"Lucas De Marchi" <lucas.demarchi@intel.com>,
"Thomas Hellström" <thomas.hellstrom@linux.intel.com>
Cc: Vinay Belgaumkar <vinay.belgaumkar@intel.com>,
<intel-gfx@lists.freedesktop.org>,
<igt-dev@lists.freedesktop.org>,
"Lucas De Marchi" <lucas.demarchi@intel.com>,
Kamil Konieczny <kamil.konieczny@linux.intel.com>
Subject: Re: [PATCH i-g-t v2 0/4] tools/gputop: Add PMU stats
Date: Fri, 20 Dec 2024 14:13:26 -0500 [thread overview]
Message-ID: <Z2XB1jhvfreb3kME@intel.com> (raw)
In-Reply-To: <7d77e830-54c1-42c0-9402-16937697d2bb@ursulin.net>
On Fri, Dec 20, 2024 at 10:15:04AM +0000, Tvrtko Ursulin wrote:
>
> On 20/12/2024 00:37, Vinay Belgaumkar wrote:
> > Use the PMU support being added in
> > https://patchwork.freedesktop.org/series/139121/ to add freq/c6 stats.
>
> Brace yourself for the customary "why". So yes, why?
>
> Gputop so far was a reference showcase for DRM fdinfo, nothing more. Why add
> a _subset_ of PMU stats to it? Any other drivers could be wired up? How far
> do people intend to grow it, considering alternatives with nicer UI and more
> featureful exist?
Well, currently intel_gpu_top doesn't support Xe and we really don't
have any intention to add xe support there.
We don't believe it is a better UI and more features.
Hopefully someday we can get qmassa [1] part of the distros and make that to
be our default to use the uapis that we add in Xe.
But for now we were in the hope that we could use gputop for that which is
the current tool that supports Xe metrics. But well, I just noticed that
at least in Fedora it is not packaged :/
$ rpm -ql igt-gpu-tools-1.27.1-0.4.20230215git45da871.fc40.x86_64 | grep top
/usr/bin/intel_gpu_top
/usr/share/man/man1/intel_gpu_top.1.gz
[1] - https://github.com/ulissesf/qmassa
So, our options are:
1. Add all the Xe support in intel_gpu_top
2. Convince distros to build and package gputop along with intel_gpu_top in igt
3. Convince distros to adopt qmaasa
Meanwhile our PMU are blocked...
Lucas, Thomas, thoughts?
>
> Or in case the conclusion ends up being "yes", then lets at least share some
> more code between intel_gpu_top and this work. Ie. make it in a way gputop
> completely subsumes and replaces intel_gpu_top might be an idea.
with this I agree as well.
>
> Regards,
>
> Tvrtko
>
> > Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
> > Cc: Lucas De Marchi <lucas.demarchi@intel.com>
> > Cc: Kamil Konieczny <kamil.konieczny@linux.intel.com>
> > Signed-off-by: Vinay Belgaumkar <vinay.belgaumkar@intel.com>
> >
> > Vinay Belgaumkar (4):
> > tools/gputop: Define data structs for PMU stats
> > lib/igt_drm_clients: Add pdev and driver info
> > lib/igt_perf: Add utils to extract PMU event info
> > tools/gputop: Add GT freq and c6 stats
> >
> > lib/igt_drm_clients.c | 6 ++
> > lib/igt_drm_clients.h | 4 +
> > lib/igt_perf.c | 68 +++++++++++++
> > lib/igt_perf.h | 2 +
> > tools/gputop.c | 225 ++++++++++++++++++++++++++++++++++++++++++
> > tools/meson.build | 2 +-
> > 6 files changed, 306 insertions(+), 1 deletion(-)
> >
next prev parent reply other threads:[~2024-12-20 19:13 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-12-20 0:37 [PATCH i-g-t v2 0/4] tools/gputop: Add PMU stats Vinay Belgaumkar
2024-12-20 0:37 ` [PATCH i-g-t v2 1/4] tools/gputop: Define data structs for " Vinay Belgaumkar
2024-12-20 0:37 ` [PATCH i-g-t v2 2/4] lib/igt_drm_clients: Add pdev and driver info Vinay Belgaumkar
2024-12-20 0:37 ` [PATCH i-g-t v2 3/4] lib/igt_perf: Add utils to extract PMU event info Vinay Belgaumkar
2024-12-20 0:37 ` [PATCH i-g-t v2 4/4] tools/gputop: Add GT freq and c6 stats Vinay Belgaumkar
2024-12-20 10:15 ` [PATCH i-g-t v2 0/4] tools/gputop: Add PMU stats Tvrtko Ursulin
2024-12-20 19:13 ` Rodrigo Vivi [this message]
2024-12-20 19:32 ` Lucas De Marchi
2024-12-22 17:47 ` Tvrtko Ursulin
2025-01-13 23:19 ` Belgaumkar, Vinay
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=Z2XB1jhvfreb3kME@intel.com \
--to=rodrigo.vivi@intel.com \
--cc=igt-dev@lists.freedesktop.org \
--cc=intel-gfx@lists.freedesktop.org \
--cc=kamil.konieczny@linux.intel.com \
--cc=lucas.demarchi@intel.com \
--cc=thomas.hellstrom@linux.intel.com \
--cc=tursulin@ursulin.net \
--cc=vinay.belgaumkar@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