new benches using pixelbomber

[fabi321]

breakwater::test::helpers::pixelflut_commands is never used. should it be removed?

For the record:

Processed image, pixel commands bytes: 38142000, 18.394096 bytes per pixel, using no optimizations
parse_draw_commands_ordered/1920 x 1080
                        time:   [9.8654 ms 9.9548 ms 10.085 ms]
                        change: [-0.2438% +0.4584% +1.4045%] (p = 0.34 > 0.05)
                        No change in performance detected.
Found 5 outliers among 100 measurements (5.00%)
  3 (3.00%) high mild
  2 (2.00%) high severe

Processed image, pixel commands bytes: 38142000, 18.394096 bytes per pixel, using no optimizations
parse_draw_commands_shuffled/1920 x 1080
                        time:   [18.889 ms 18.936 ms 18.986 ms]
                        change: [+0.3597% +0.6845% +1.0536%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild

Processed image, pixel commands bytes: 28271640, 13.634086 bytes per pixel, using both gray and offset optimizations
parse_mixed_draw_commands/1920 x 1080
                        time:   [8.5315 ms 8.6067 ms 8.7171 ms]
Found 11 outliers among 100 measurements (11.00%)
  3 (3.00%) high mild
  8 (8.00%) high severe

Processed image, pixel commands bytes: 28271640, 13.634086 bytes per pixel, using both gray and offset optimizations
parse_mixed_draw_commands_shuffled/1920 x 1080
                        time:   [8.6021 ms 8.6253 ms 8.6513 ms]
Found 5 outliers among 100 measurements (5.00%)
  1 (1.00%) high mild
  4 (4.00%) high severe

So using offset actually helps performance significantly ;)

Fixes #4

[fabi321]

the new bench results for me:

Processed image, pixel commands bytes: 38142000, 18.394096 bytes per pixel, using no optimizations
Benchmarking parse_draw_commands_ordered/1920 x 1080: Warming up for 10.000 s
Warning: Unable to complete 100 samples in 30.0s. You may wish to increase target time to 49.6s, enable flat sampling, or reduce sample count to 50.
parse_draw_commands_ordered/1920 x 1080
                        time:   [9.7931 ms 9.8223 ms 9.8558 ms]
                        change: [-2.0787% -1.0964% -0.3418%] (p = 0.01 < 0.05)
                        Change within noise threshold.
Found 5 outliers among 100 measurements (5.00%)
  3 (3.00%) high mild
  2 (2.00%) high severe

Processed image, pixel commands bytes: 38142000, 18.394096 bytes per pixel, using no optimizations
parse_draw_commands_shuffled/1920 x 1080
                        time:   [18.789 ms 18.838 ms 18.890 ms]
                        change: [-0.8874% -0.5180% -0.1368%] (p = 0.01 < 0.05)
                        Change within noise threshold.
Found 3 outliers among 100 measurements (3.00%)
  2 (2.00%) high mild
  1 (1.00%) high severe

Processed image, pixel commands bytes: 37105200, 17.894096 bytes per pixel, using only gray optimization
Benchmarking parse_mixed_draw_commands/1920 x 1080: Warming up for 10.000 s
Warning: Unable to complete 100 samples in 30.0s. You may wish to increase target time to 50.1s, enable flat sampling, or reduce sample count to 50.
parse_mixed_draw_commands/1920 x 1080
                        time:   [9.8759 ms 9.9003 ms 9.9274 ms]
                        change: [+14.179% +15.254% +16.186%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild

Processed image, pixel commands bytes: 29308440, 14.134086 bytes per pixel, using only offset optimization
Benchmarking parse_draw_commands_with_offset/1920 x 1080: Warming up for 10.000 s
Warning: Unable to complete 100 samples in 30.0s. You may wish to increase target time to 40.9s, enable flat sampling, or reduce sample count to 60.
parse_draw_commands_with_offset/1920 x 1080
                        time:   [8.2587 ms 8.2849 ms 8.3130 ms]
Found 5 outliers among 100 measurements (5.00%)
  3 (3.00%) high mild
  2 (2.00%) high severe

[fabi321]

A note on the mixed bench: it is 50% rrggbb, 25% rrggbbaa and 25% gg rn. Let me know if you want to go with this, or rather have dedicated benches for each type.

[sbernauer]

Many thanks for the nice benchmarks, sorry for taking so long!
Mixed benchmark and breakwater::test::helpers::pixelflut_commands would be fine for me as of now, but happy to let you decide on that and create a follow-up PR if desired :)