Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
To: Riana Tauro <riana.tauro@intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>,
	<intel-xe@lists.freedesktop.org>, <anshuman.gupta@intel.com>,
	<rodrigo.vivi@intel.com>, <vinay.belgaumkar@intel.com>,
	<aravind.iddamsetty@intel.com>, <john.c.harrison@intel.com>,
	<ashutosh.dixit@intel.com>, <soham.purkait@intel.com>
Subject: Re: [PATCH v2 8/8] drm/xe/pmu: Add PMU support for engine busyness
Date: Thu, 19 Dec 2024 10:46:06 -0800	[thread overview]
Message-ID: <Z2Rp7q09O3knfGqS@orsosgc001> (raw)
In-Reply-To: <d3d24cd9-c54b-43bd-a9d9-d7682b4fe10e@intel.com>

On Thu, Dec 19, 2024 at 11:17:16AM +0530, Riana Tauro wrote:
>
>
>On 12/18/2024 8:21 PM, Riana Tauro wrote:
>>
>>
>>On 12/18/2024 7:43 PM, Lucas De Marchi wrote:
>>>On Wed, Dec 18, 2024 at 10:42:53AM +0530, Riana Tauro wrote:
>>>>Hi Lucas
>>>>
>>>>On 12/13/2024 11:28 AM, Lucas De Marchi wrote:
>>>>>On Thu, Nov 21, 2024 at 12:09:04PM +0530, Riana Tauro wrote:
>>>>>>PMU provides two counters (<engine>-busy-ticks-gt<n>,
>>>>>><engine>-total-ticks-gt<n>) to calculate engine busyness. 
>>>>>>When querying
>>>>>>engine busyness, user must group these 2 counters using the perf_event
>>>>>>group mechanism to ensure both counters are sampled together.
>>>>>>
>>>>>>To list engine busyness counters use the following
>>>>>>
>>>>>>./perf list
>>>>>> xe_0000_03_00.0/bcs0-busy-ticks-gt0/               [Kernel PMU event]
>>>>>
>>>>>this will need a rebase on latest versions of the pmu patches as we
>>>>>moved to have gt as a param rather than mangling the event name.
>>>>
>>>>In case of multiple gts, there might be different engines for each gt
>>>>So should we display a common name and unsupported in case the 
>>>>engine does not belong to the gt?
>>>>
>>>>sudo ./perf stat -e xe_0000_00_02.0/vcs0-busy-ticks,gt_id=0/ -I 1000
>>>
>>>why is engine not a param, like gt?
>Wouldn't it be better to have engine name in the event name? Otherwise 
>anyone using perf tool needs to find the engine class to name mapping
>
>@Umesh any inputs here?

sorry, I am not clear on how even gt_id is being passed here. How does 
the gt_id fit into the struct perf_event_attr {} that we pass as the 
event configuration? Are we using the config1/config2 fields to pass 
these additional params?

fwiu, whether we pass engine as a parameter or use the name in the 
event, we would still need to validate if the engine belongs to that gt.

Thanks,
Umesh

>>are you suggesting to have both engine and instance too as parameters?
>>Something like this?
>>
>>xe_0000_00_02.0/busy-ticks,engine_class=1,engine_instance=0,gt_id=0/
>>
>>
>>>
>>>Lucas De Marchi
>>
>

  reply	other threads:[~2024-12-19 18:46 UTC|newest]

Thread overview: 38+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-11-21  6:38 [PATCH v2 0/8] Add PMU support for single engine busyness Riana Tauro
2024-11-21  6:33 ` ✓ CI.Patch_applied: success for Add PMU support for single engine busyness (rev2) Patchwork
2024-11-21  6:34 ` ✗ CI.checkpatch: warning " Patchwork
2024-11-21  6:35 ` ✓ CI.KUnit: success " Patchwork
2024-11-21  6:38 ` [PATCH v2 1/8] [DO NOT REVIEW] drm/xe/pmu: Enable PMU interface Riana Tauro
2024-11-21  6:38 ` [PATCH v2 2/8] [DO NOT REVIEW] drm/xe/pmu: Add GT C6 events Riana Tauro
2024-11-21  6:38 ` [PATCH v2 3/8] [DO NOT REVIEW] drm/xe/pmu: Add GT frequency events Riana Tauro
2024-11-21  6:39 ` [PATCH v2 4/8] drm/xe: add function to convert xe hw engine class to user class Riana Tauro
2024-12-10 23:48   ` Umesh Nerlige Ramappa
2024-12-18  7:33     ` Riana Tauro
2024-12-19  0:13       ` Umesh Nerlige Ramappa
2024-11-21  6:39 ` [PATCH v2 5/8] drm/xe: Add single engine busyness support Riana Tauro
2024-12-12 23:56   ` Umesh Nerlige Ramappa
2024-12-13  0:52     ` Lucas De Marchi
2024-12-13  7:06       ` Riana Tauro
2024-12-13 14:15         ` Lucas De Marchi
2024-12-13  6:22     ` Riana Tauro
2024-11-21  6:39 ` [PATCH v2 6/8] drm/xe/trace: Add trace for xe_engine_activity Riana Tauro
2024-11-21  6:39 ` [PATCH v2 7/8] drm/xe/guc: Expose engine busyness only for supported GuC version Riana Tauro
2024-12-13  5:44   ` Lucas De Marchi
2024-12-13  6:00     ` Riana Tauro
2024-12-13  6:04       ` Lucas De Marchi
2024-12-13  6:10         ` Riana Tauro
2024-12-13 17:26           ` Lucas De Marchi
2024-12-19  0:19   ` John Harrison
2024-11-21  6:39 ` [PATCH v2 8/8] drm/xe/pmu: Add PMU support for engine busyness Riana Tauro
2024-12-13  5:58   ` Lucas De Marchi
2024-12-18  5:12     ` Riana Tauro
2024-12-18 14:13       ` Lucas De Marchi
2024-12-18 14:51         ` Riana Tauro
2024-12-19  5:47           ` Riana Tauro
2024-12-19 18:46             ` Umesh Nerlige Ramappa [this message]
2025-01-06 14:06               ` Riana Tauro
2024-11-21  6:53 ` ✓ CI.Build: success for Add PMU support for single engine busyness (rev2) Patchwork
2024-11-21  6:55 ` ✗ CI.Hooks: failure " Patchwork
2024-11-21  6:57 ` ✓ CI.checksparse: success " Patchwork
2024-11-21  7:17 ` ✗ Xe.CI.BAT: failure " Patchwork
2024-11-21 13:52 ` ✗ Xe.CI.Full: " Patchwork

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=Z2Rp7q09O3knfGqS@orsosgc001 \
    --to=umesh.nerlige.ramappa@intel.com \
    --cc=anshuman.gupta@intel.com \
    --cc=aravind.iddamsetty@intel.com \
    --cc=ashutosh.dixit@intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=john.c.harrison@intel.com \
    --cc=lucas.demarchi@intel.com \
    --cc=riana.tauro@intel.com \
    --cc=rodrigo.vivi@intel.com \
    --cc=soham.purkait@intel.com \
    --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