schedule: zephyr_domain: make LL reporting build-time configurable and change the defaults #9661
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add Kconfig options to enable and disable the low-latency scheduler statistics reporting, and to set the window size. Also add an option to set reporting to be done only for every other statistics window. This can be useful to exclude impact of the logging itself. With most current configurations and logging backends, printing out the scheduling report at the end of the window, will show up as an execution time spike on the first iteration of the subsequent measurement window. This can mask spikes caused by other components in the system. Defaults match the earlier settings hardcoded in code, so this commit has no functional change. Signed-off-by: Kai Vehmanen <[email protected]>
Change the logic how overrun count is reported out in the low-latency scheduler statistics reporting. Instead of resetting the overrun count after every window, keep a cumulative count of overruns seen. Overruns are in practise a rare occurrence and this change makes the CONFIG_SCHEDULE_LL_STATS_LOG_EVERY_OTHER_WINDOW option more useful as one can still track overruns for all executed iterations, even if reporting is done at a slower pace. Signed-off-by: Kai Vehmanen <[email protected]>
Make SCHEDULE_LL_STATS_LOG_EVERY_OTHER_WINDOW=y the default. This in practise filters out the impact of the stat logging from the results. Another impact is that logging rate is halved to once per two seconds with the new default settings. Signed-off-by: Kai Vehmanen <[email protected]>
kv2019i
requested review from
pblaszko,
marcinszkudlinski,
dbaluta and
LaurentiuM1234
as code owners
|
The effect of this series looks like this on a light playback pipeline on Intel MTL target: Whether one wants to see the higher max values (they are real cycles consumed after all), depends on what one is trying to debug. |
lgirdwood
approved these changes
Nov 19, 2024
lyakh
approved these changes
Nov 19, 2024
marcinszkudlinski
approved these changes
Nov 19, 2024
|
sof-docs fail and Intel LNL fails all known and tracked in https://github.com/thesofproject/sof/issues?q=is%3Aissue+is%3Aopen+label%3ACI |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Here's a series to make the LL scheduler stat logging more configurable, and tune the defaults using the new tools.