From: Marcin Bernatowicz <marcin.bernatowicz@linux.intel.com>
To: igt-dev@lists.freedesktop.org
Cc: Marcin Bernatowicz <marcin.bernatowicz@linux.intel.com>,
Adam Miszczak <adam.miszczak@linux.intel.com>,
Jakub Kolakowski <jakub1.kolakowski@intel.com>,
Kamil Konieczny <kamil.konieczny@linux.intel.com>,
Lukasz Laguna <lukasz.laguna@intel.com>,
Satyanarayana K V P <satyanarayana.k.v.p@intel.com>
Subject: [PATCH v3 i-g-t 0/3] tests/intel/xe_sriov_scheduling: Refactor submission/measurement
Date: Mon, 25 Aug 2025 10:22:51 +0200 [thread overview]
Message-ID: <20250825082254.444880-1-marcin.bernatowicz@linux.intel.com> (raw)
Refactor submission/measurement to better saturate HW and make
throughput comparisons more robust, especially with short jobs.
Add the --inflight option.
- Drive a K-in-flight pipeline per VF using per-slot BO/addr/spin and
binary out-fences; add subm_exec_slot()/subm_wait_slot() (prefill +
refill).
- Record complete_ts[] and per-slot submit_ts[]; build the common
window from completions [max(first), min(last)] and compute
throughput as count/window.
- Push durations as submit-to-completion (complete_ts - submit_ts) and
print "mean submit->signal latency".
- Add --inflight (0=auto; non-preempt defaults to 1; short jobs pick
higher K); print chosen K in the banner.
v2: - drop redundant num_syncs init; simplify subm_exec_slot (Lukasz)
- free complete_ts (Lukasz)
Signed-off-by: Marcin Bernatowicz <marcin.bernatowicz@linux.intel.com>
Cc: Adam Miszczak <adam.miszczak@linux.intel.com>
Cc: Jakub Kolakowski <jakub1.kolakowski@intel.com>
Cc: Kamil Konieczny <kamil.konieczny@linux.intel.com>
Cc: Lukasz Laguna <lukasz.laguna@intel.com>
Cc: Satyanarayana K V P <satyanarayana.k.v.p@intel.com>
Marcin Bernatowicz (3):
tests/intel/xe_sriov_scheduling: Keep K submissions in flight
tests/intel/xe_sriov_scheduling: Compute throughput from completion
timestamps
tests/intel/xe_sriov_scheduling: Make in-flight submissions
configurable
tests/intel/xe_sriov_scheduling.c | 237 +++++++++++++++++++++---------
1 file changed, 167 insertions(+), 70 deletions(-)
--
2.31.1
next reply other threads:[~2025-08-25 8:23 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-08-25 8:22 Marcin Bernatowicz [this message]
2025-08-25 8:22 ` [PATCH v3 i-g-t 1/3] tests/intel/xe_sriov_scheduling: Keep K submissions in flight Marcin Bernatowicz
2025-08-25 9:15 ` Laguna, Lukasz
2025-08-25 8:22 ` [PATCH v3 i-g-t 2/3] tests/intel/xe_sriov_scheduling: Compute throughput from completion timestamps Marcin Bernatowicz
2025-08-25 9:15 ` Laguna, Lukasz
2025-08-25 8:22 ` [PATCH v3 i-g-t 3/3] tests/intel/xe_sriov_scheduling: Make in-flight submissions configurable Marcin Bernatowicz
2025-08-25 10:37 ` ✓ Xe.CI.BAT: success for tests/intel/xe_sriov_scheduling: Refactor submission/measurement (rev2) Patchwork
2025-08-25 10:49 ` ✓ i915.CI.BAT: " Patchwork
2025-08-25 12:41 ` ✗ Xe.CI.Full: failure " Patchwork
2025-08-25 13:27 ` ✗ i915.CI.Full: " Patchwork
2025-08-26 14:51 ` Bernatowicz, Marcin
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=20250825082254.444880-1-marcin.bernatowicz@linux.intel.com \
--to=marcin.bernatowicz@linux.intel.com \
--cc=adam.miszczak@linux.intel.com \
--cc=igt-dev@lists.freedesktop.org \
--cc=jakub1.kolakowski@intel.com \
--cc=kamil.konieczny@linux.intel.com \
--cc=lukasz.laguna@intel.com \
--cc=satyanarayana.k.v.p@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 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.