Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Rodrigo Vivi <rodrigo.vivi@intel.com>
To: <intel-gfx@lists.freedesktop.org>
Subject: Re: [PATCH 0/3] Enable GuC SLPC default balancing strategies v2
Date: Thu, 9 Jan 2025 17:28:59 -0500	[thread overview]
Message-ID: <Z4BNq5ZtV711sBzs@intel.com> (raw)
In-Reply-To: <20250108141318.63823-1-rodrigo.vivi@intel.com>

On Wed, Jan 08, 2025 at 09:13:14AM -0500, Rodrigo Vivi wrote:
> The goal here is to improve the GPU performance in some cases
> where we are TDP limited, with both GPU and CPU with high
> utilization and PCODE without power to fulfill both sides.
> Of course, only matters for integrated parts with GuC
> SLPC enabled: ADL and MTL.
> 
> Apparently the ADL on our CI was not happy about that
> with the igt_spinners timming out [1]. So, this new attempt
> also lift the BAT restriction on the igt_spinner with the
> first patch.
> 
> Another possibility would be to make this change for MTL
> only, but I'm afraid that this is not the ideal, because:
> 1. ADL users couldn't get the performance benefit of this change
> 2. This would sound like a hack to bypass CI, where the ADL there
>    might have some TDP limitation and MTL there having more power.

For the records, I have changed my mind here. IBC on ADL was doing
balancing even when not TDP limited and perhaps even depending on
display metrics. Let's enable it only on MTL+

> 
> [1] - https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_142676v2/bat-twl-2/igt@i915_selftest@live@gt_engines.html
> 
> Thanks,
> Rodrigo.
> 
> Rodrigo Vivi (3):
>   drm/i915: Increase the timeout of the spinners
>   drm/i915/guc/slpc: Enable GuC SLPC default strategies
>   drm/i915/guc/slpc: Print more SLPC debug status information
> 
>  drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.c  | 34 ++++++++++----------
>  drivers/gpu/drm/i915/selftests/igt_spinner.c |  2 +-
>  2 files changed, 18 insertions(+), 18 deletions(-)
> 
> -- 
> 2.47.1
> 

      parent reply	other threads:[~2025-01-09 22:29 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-01-08 14:13 [PATCH 0/3] Enable GuC SLPC default balancing strategies v2 Rodrigo Vivi
2025-01-08 14:13 ` [PATCH 1/3] drm/i915: Increase the timeout of the spinners Rodrigo Vivi
2025-01-08 14:13 ` [PATCH 2/3] drm/i915/guc/slpc: Enable GuC SLPC default strategies Rodrigo Vivi
2025-01-08 19:11   ` Belgaumkar, Vinay
2025-01-08 19:30     ` Dixit, Ashutosh
2025-01-08 21:19       ` Rodrigo Vivi
2025-01-09  5:18         ` Dixit, Ashutosh
2025-01-09 19:00           ` Rodrigo Vivi
2025-01-09 21:32             ` Dixit, Ashutosh
2025-01-08 14:13 ` [PATCH 3/3] drm/i915/guc/slpc: Print more SLPC debug status information Rodrigo Vivi
2025-01-08 19:08   ` Belgaumkar, Vinay
2025-01-08 16:10 ` ✓ i915.CI.BAT: success for Enable GuC SLPC default balancing strategies (rev3) Patchwork
2025-01-09 18:06 ` ✓ i915.CI.Full: " Patchwork
2025-01-09 22:28 ` Rodrigo Vivi [this message]

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=Z4BNq5ZtV711sBzs@intel.com \
    --to=rodrigo.vivi@intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    /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