[PATCH i-g-t v3 0/7] benchmarks/kms_fb_stress: Test all possible combinations of formats

[Pekka Paalanen]

On Thu, 07 Mar 2024 16:41:02 -0300
Arthur Grillo <arthurgrillo at riseup.net> wrote:

> After some discussion[1], I changed how to tackle the problem.
> 
> With the upcoming addition of YUV format to VKMS[2] it would be great to
> test the their performance. So, test all possible combinations of
> formats.

Hi Arthur,

what do you gain by testing all possible combinations of formats on
planes?

I would have thought that if you can pick and choose a format, then you
don't need to test combinations. Each format alone (but as plane source
and writeback destination separately) would exercise all the VKMS code
there is for that format, given that VKMS uses the intermediate pixel
format for all blending etc. operations.


What will the standard benchmark results printout look like?

I imagine it should be a simple table that people could put in their
commit message, something like this:

kms_fb_stress test results, plane source:

            before    after    difference
  ARGB8888   3.6 s    3.7 s       +2.7 %
  ABGR8888   3.6 s    3.8 s       +5.6 %
  YUV420    15.5 s    9.9 s        -36 %
  YUV410         -    8.8 s          -

and another table for writeback destination, if that's interesting.

But that requires both before and after runs, so one would need to save
results in a file, and use a tool to generate the table. Well, maybe
something to consider in the future. One could also get even fancier
with statistical analysis, but I guess this would be plenty already.


Thanks,
pq

> Also, to focus in an combination subset, add command-line options to
> arbitrarily select the format of each plane and writeback.
> 
> - PATCH #1: Just a non-related assert that I found needed while developing
> - PATCH #2: Log the KMS structure
> - PATCH #3: Moving code, should be noop
> - PATCH #4 & #5: Fix a couple of bugs found while developing patch #6
> - PATCH #6: Test all possible combinations
> - PATCH #7: Add command-line options
> 
> [1]: https://lore.kernel.org/all/6979cd2e-2b00-4dc4-8e41-66b435cf7ea8@riseup.net/
> [2]: https://lore.kernel.org/all/20240304-yuv-v4-11-76beac8e9793@bootlin.com/
> 
> Signed-off-by: Arthur Grillo <arthurgrillo at riseup.net>
> ---
> Changes in v3:
> - Make the help message and commit message more clear on how the
>   command-line options work (Louis Chauvet)
> - Link to v2: https://lore.kernel.org/r/20240305-kms_fb_stress-dev-v2-0-78888a38e563@riseup.net
> 
> Changes in v2:
> - Change how the command-line options work, now they pin the specific
>   format and let the other change.
> - Add a patch to test all the possible formats combination
> - Add a couple of patches to fix some bugs
> - Move the KMS logging patch to the beginning of the series
> - Move the default_kms definition to the static struct data_t variable
> - Link to v1: https://lore.kernel.org/r/20240226-kms_fb_stress-dev-v1-0-1c14942b1244@riseup.net
> 
> ---
> Arthur Grillo (7):
>       benchmarks/kms_fb_stress: Assert that we have an supported pipe
>       benchmarks/kms_fb_stress: Log the KMS structure
>       benchmarks/kms_fb_stress: Move the stress procedure to its own function
>       benchmarks/kms_fb_stress: Free resources on the stress procedure
>       benchmarks/kms_fb_stress: Don't paint the FB's if the format is DRM_FORMAT_RGB565
>       benchmarks/kms_fb_stress: Test every possible format combinations
>       benchmarks/kms_fb_stress: Add command line options to pin the planes or writeback formats
> 
>  benchmarks/kms_fb_stress.c | 342 ++++++++++++++++++++++++++++++++++++---------
>  1 file changed, 273 insertions(+), 69 deletions(-)
> ---
> base-commit: 32369affc1fdb8194c4e7274d1d4dfd74f4aae88
> change-id: 20240226-kms_fb_stress-dev-c93ae176f901
> 
> Best regards,

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <https://lists.freedesktop.org/archives/igt-dev/attachments/20240320/79675075/attachment.sig>