Make conversion to/from arrays generic, take 2

[Shnatsel]

A different take on #286: reduce duplicate code and allow all the same generic programming perks, without creating extra traits and disrupting the existing trait hierarchy.

This is a draft. I am going to port Vello to this to see how it feels in practice before committing to the design.

[Shnatsel]

Vello does not need any code changes, unlike for #286.

Vello benchmarks improvements:

• 18% on fine/blend/normal_u8_avx2
• 70% on fine/blend/clear_u8_sse4_2
• 23% on fine/blend/dest_over_u8_sse4_2
• 10% on fine/image/extend/reflect_u8_sse4_2
• 11% on fine/image/extend/reflect_u8_avx2

Vello benchmark regressions:

• 18% on fine/blend/clear_u8_avx2
• 16% on fine/blend/src_out_u8_sse4_2
• 271% on fine/blend/plus_u8_sse4_2
• 261% on fine/blend/plus_lighter_u8_sse4_2

The rest is noise - small movements both ways.

Full benchmark run

allocator/alloc_1000_varied_8_128
                        time:   [297.68 µs 297.71 µs 297.76 µs]
                        change: [+0.7412% +0.7922% +0.8431%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 3 outliers among 100 measurements (3.00%)
  1 (1.00%) low mild
  1 (1.00%) high mild
  1 (1.00%) high severe

allocator/alloc_until_full_32x32
                        time:   [330.19 µs 330.21 µs 330.22 µs]
                        change: [-0.5530% -0.5196% -0.4898%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 6 outliers among 100 measurements (6.00%)
  1 (1.00%) low severe
  2 (2.00%) high mild
  3 (3.00%) high severe

allocator/churn_500_steady_state
                        time:   [242.03 µs 242.06 µs 242.09 µs]
                        change: [+1.2235% +1.2658% +1.3059%] (p = 0.00 < 0.05)
                        Performance has regressed.

tile_aaa/Ghostscript_Tiger
                        time:   [72.006 µs 72.012 µs 72.019 µs]
                        change: [+4.5791% +4.6784% +4.7726%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 2 outliers among 50 measurements (4.00%)
  1 (2.00%) high mild
  1 (2.00%) high severe
tile_aaa/coat_of_arms   time:   [942.53 µs 942.60 µs 942.68 µs]
                        change: [+3.2055% +3.2779% +3.3493%] (p = 0.00 < 0.05)
                        Performance has regressed.
tile_aaa/Saimaa_Canal_(map)
                        time:   [2.6468 ms 2.6470 ms 2.6472 ms]
                        change: [-1.5211% -1.3925% -1.2637%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 2 outliers among 50 measurements (4.00%)
  1 (2.00%) low mild
  1 (2.00%) high severe
tile_aaa/heraldry       time:   [545.83 µs 546.11 µs 546.46 µs]
                        change: [+3.1864% +3.2484% +3.3097%] (p = 0.00 < 0.05)
                        Performance has regressed.

tile_msaa/Ghostscript_Tiger
                        time:   [206.01 µs 206.01 µs 206.03 µs]
                        change: [-0.2539% -0.1679% -0.0873%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 50 measurements (8.00%)
  2 (4.00%) high mild
  2 (4.00%) high severe
tile_msaa/coat_of_arms  time:   [1.8940 ms 1.8941 ms 1.8942 ms]
                        change: [+0.6204% +0.6445% +0.6677%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 6 outliers among 50 measurements (12.00%)
  1 (2.00%) low severe
  1 (2.00%) low mild
  4 (8.00%) high mild
tile_msaa/Saimaa_Canal_(map)
                        time:   [5.5862 ms 5.5870 ms 5.5877 ms]
                        change: [-0.9080% -0.8621% -0.8159%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 2 outliers among 50 measurements (4.00%)
  1 (2.00%) high mild
  1 (2.00%) high severe
tile_msaa/heraldry      time:   [1.2647 ms 1.2648 ms 1.2649 ms]
                        change: [-0.3442% -0.3166% -0.2882%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 50 measurements (8.00%)
  1 (2.00%) low mild
  1 (2.00%) high mild
  2 (4.00%) high severe

tile_aaa_shift50/Ghostscript_Tiger
                        time:   [56.659 µs 56.661 µs 56.663 µs]
                        change: [+5.6428% +5.6885% +5.7315%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 4 outliers among 50 measurements (8.00%)
  2 (4.00%) high mild
  2 (4.00%) high severe
tile_aaa_shift50/coat_of_arms
                        time:   [647.46 µs 647.57 µs 647.68 µs]
                        change: [+6.5828% +6.6188% +6.6523%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 1 outliers among 50 measurements (2.00%)
  1 (2.00%) high mild
tile_aaa_shift50/Saimaa_Canal_(map)
                        time:   [2.3434 ms 2.3437 ms 2.3439 ms]
                        change: [+1.6456% +1.6751% +1.7025%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 3 outliers among 50 measurements (6.00%)
  1 (2.00%) low mild
  2 (4.00%) high mild
tile_aaa_shift50/heraldry
                        time:   [273.84 µs 273.86 µs 273.87 µs]
                        change: [+4.7110% +5.3995% +5.7679%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 3 outliers among 50 measurements (6.00%)
  1 (2.00%) high mild
  2 (4.00%) high severe

render_strips/Ghostscript_Tiger_simd
                        time:   [124.61 µs 124.61 µs 124.62 µs]
                        change: [-0.0266% -0.0122% +0.0083%] (p = 0.16 > 0.05)
                        No change in performance detected.
Found 4 outliers among 50 measurements (8.00%)
  1 (2.00%) high mild
  3 (6.00%) high severe
render_strips/coat_of_arms_simd
                        time:   [1.3299 ms 1.3300 ms 1.3301 ms]
                        change: [-0.1053% -0.0339% +0.0271%] (p = 0.35 > 0.05)
                        No change in performance detected.
Found 3 outliers among 50 measurements (6.00%)
  3 (6.00%) high severe
render_strips/Saimaa_Canal_(map)_simd
                        time:   [3.5440 ms 3.5442 ms 3.5444 ms]
                        change: [+0.1073% +0.1552% +0.2037%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 50 measurements (8.00%)
  1 (2.00%) high mild
  3 (6.00%) high severe
render_strips/heraldry_simd
                        time:   [878.70 µs 878.79 µs 878.89 µs]
                        change: [-0.2476% -0.2244% -0.1937%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 3 outliers among 50 measurements (6.00%)
  1 (2.00%) high mild
  2 (4.00%) high severe

render_strips_culled50/Ghostscript_Tiger
                        time:   [49.341 µs 49.343 µs 49.345 µs]
                        change: [-0.0610% -0.0343% -0.0136%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 50 measurements (8.00%)
  2 (4.00%) high mild
  2 (4.00%) high severe
render_strips_culled50/coat_of_arms
                        time:   [558.66 µs 558.72 µs 558.77 µs]
                        change: [+0.1571% +0.1985% +0.2518%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 50 measurements (8.00%)
  4 (8.00%) high severe
render_strips_culled50/Saimaa_Canal_(map)
                        time:   [2.5417 ms 2.5420 ms 2.5423 ms]
                        change: [-0.2105% -0.1556% -0.1043%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 5 outliers among 50 measurements (10.00%)
  2 (4.00%) high mild
  3 (6.00%) high severe
render_strips_culled50/heraldry
                        time:   [393.97 µs 394.02 µs 394.08 µs]
                        change: [+0.0446% +0.1354% +0.2144%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 3 outliers among 50 measurements (6.00%)
  2 (4.00%) high mild
  1 (2.00%) high severe

render_rect/14x14_via_path
                        time:   [200.12 ns 200.25 ns 200.38 ns]
                        change: [-2.5070% -2.3662% -2.2245%] (p = 0.00 < 0.05)
                        Performance has improved.
render_rect/14x14_via_rect
                        time:   [112.93 ns 112.95 ns 112.97 ns]
                        change: [+1.4954% +1.5447% +1.5901%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 4 outliers among 50 measurements (8.00%)
  1 (2.00%) low severe
  2 (4.00%) high mild
  1 (2.00%) high severe

flatten/Ghostscript_Tiger
                        time:   [154.64 µs 154.66 µs 154.69 µs]
                        change: [-1.0189% -0.8792% -0.7439%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 5 outliers among 100 measurements (5.00%)
  5 (5.00%) high severe
flatten/coat_of_arms    time:   [593.41 µs 593.48 µs 593.55 µs]
                        change: [-0.5604% -0.3820% -0.2059%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 5 outliers among 100 measurements (5.00%)
  2 (2.00%) high mild
  3 (3.00%) high severe
flatten/Saimaa_Canal_(map)
                        time:   [1.1010 ms 1.1015 ms 1.1021 ms]
                        change: [-1.1109% -0.1692% +0.7302%] (p = 0.74 > 0.05)
                        No change in performance detected.
Found 12 outliers among 100 measurements (12.00%)
  4 (4.00%) high mild
  8 (8.00%) high severe
flatten/heraldry        time:   [315.02 µs 315.55 µs 316.08 µs]
                        change: [+0.7514% +0.9131% +1.0871%] (p = 0.00 < 0.05)
                        Change within noise threshold.

strokes/Ghostscript_Tiger
                        time:   [124.35 µs 124.35 µs 124.36 µs]
                        change: [-0.2591% -0.2347% -0.2142%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 100 measurements (4.00%)
  2 (2.00%) high mild
  2 (2.00%) high severe
strokes/coat_of_arms    time:   [1.1359 ms 1.1360 ms 1.1361 ms]
                        change: [-1.1469% -1.0162% -0.8845%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 6 outliers among 100 measurements (6.00%)
  1 (1.00%) low mild
  5 (5.00%) high mild
strokes/Saimaa_Canal_(map)
                        time:   [7.4082 ms 7.4113 ms 7.4147 ms]
                        change: [-0.7548% -0.6869% -0.6180%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 2 outliers among 100 measurements (2.00%)
  1 (1.00%) high mild
  1 (1.00%) high severe
strokes/heraldry        time:   [847.68 µs 848.10 µs 848.69 µs]
                        change: [-0.2445% -0.2165% -0.1833%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 13 outliers among 100 measurements (13.00%)
  4 (4.00%) high mild
  9 (9.00%) high severe

glyph/cached_hinted     time:   [37.745 µs 37.757 µs 37.769 µs]
                        change: [-2.7768% -2.6917% -2.6058%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 6 outliers among 100 measurements (6.00%)
  1 (1.00%) low mild
  4 (4.00%) high mild
  1 (1.00%) high severe
glyph/uncached_hinted   time:   [52.507 µs 52.515 µs 52.523 µs]
                        change: [-1.0035% -0.9088% -0.8162%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 6 outliers among 100 measurements (6.00%)
  5 (5.00%) high mild
  1 (1.00%) high severe
glyph/cached_unhinted   time:   [37.369 µs 37.415 µs 37.458 µs]
                        change: [-3.5066% -3.3864% -3.2692%] (p = 0.00 < 0.05)
                        Performance has improved.
glyph/uncached_unhinted time:   [51.519 µs 51.527 µs 51.536 µs]
                        change: [-0.9800% -0.8883% -0.7965%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 12 outliers among 100 measurements (12.00%)
  1 (1.00%) low severe
  1 (1.00%) low mild
  7 (7.00%) high mild
  3 (3.00%) high severe

fine/fill/opaque_short_u8_sse4_2
                        time:   [6.9125 ns 6.9197 ns 6.9269 ns]
                        change: [-7.1323% -7.0262% -6.9287%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
  4 (4.00%) high mild

fine/fill/opaque_short_u8_avx2
                        time:   [5.6816 ns 5.6961 ns 5.7139 ns]
                        change: [+3.3571% +3.6992% +4.0928%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 18 outliers among 100 measurements (18.00%)
  2 (2.00%) high mild
  16 (16.00%) high severe

fine/fill/opaque_long_u8_sse4_2
                        time:   [47.271 ns 47.273 ns 47.275 ns]
                        change: [+0.2224% +0.2565% +0.2880%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 8 outliers among 100 measurements (8.00%)
  2 (2.00%) low mild
  3 (3.00%) high mild
  3 (3.00%) high severe

fine/fill/opaque_long_u8_avx2
                        time:   [25.925 ns 25.926 ns 25.927 ns]
                        change: [-0.0445% -0.0175% +0.0065%] (p = 0.18 > 0.05)
                        No change in performance detected.
Found 10 outliers among 100 measurements (10.00%)
  3 (3.00%) high mild
  7 (7.00%) high severe

fine/fill/transparent_short_u8_sse4_2
                        time:   [19.247 ns 19.259 ns 19.269 ns]
                        change: [+1.5220% +1.6900% +1.8519%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 2 outliers among 100 measurements (2.00%)
  2 (2.00%) low mild

fine/fill/transparent_short_u8_avx2
                        time:   [12.474 ns 12.480 ns 12.488 ns]
                        change: [+1.8306% +1.9179% +2.0019%] (p = 0.00 < 0.05)
                        Performance has regressed.

fine/fill/transparent_long_u8_sse4_2
                        time:   [139.19 ns 139.29 ns 139.40 ns]
                        change: [+1.6259% +1.7250% +1.8195%] (p = 0.00 < 0.05)
                        Performance has regressed.

fine/fill/transparent_long_u8_avx2
                        time:   [86.006 ns 86.047 ns 86.085 ns]
                        change: [+0.2716% +0.3282% +0.3819%] (p = 0.00 < 0.05)
                        Change within noise threshold.

fine/strip/solid_single_u8_sse4_2
                        time:   [8.8285 ns 8.8351 ns 8.8416 ns]
                        change: [+0.3323% +0.3947% +0.4533%] (p = 0.00 < 0.05)
                        Change within noise threshold.

fine/strip/solid_single_u8_avx2
                        time:   [7.1687 ns 7.1741 ns 7.1803 ns]
                        change: [+0.3781% +0.4361% +0.5040%] (p = 0.00 < 0.05)
                        Change within noise threshold.

fine/strip/solid_short_u8_sse4_2
                        time:   [15.547 ns 15.552 ns 15.558 ns]
                        change: [-2.0124% -1.8137% -1.6073%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 15 outliers among 100 measurements (15.00%)
  7 (7.00%) low severe
  5 (5.00%) low mild
  2 (2.00%) high mild
  1 (1.00%) high severe

fine/strip/solid_short_u8_avx2
                        time:   [10.552 ns 10.560 ns 10.568 ns]
                        change: [+0.2110% +0.2988% +0.3791%] (p = 0.00 < 0.05)
                        Change within noise threshold.

fine/strip/solid_medium_u8_sse4_2
                        time:   [28.864 ns 28.905 ns 28.952 ns]
                        change: [+0.1659% +0.2999% +0.4315%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 12 outliers among 100 measurements (12.00%)
  6 (6.00%) high mild
  6 (6.00%) high severe

fine/strip/solid_medium_u8_avx2
                        time:   [20.159 ns 20.166 ns 20.173 ns]
                        change: [-0.0286% +0.0137% +0.0565%] (p = 0.53 > 0.05)
                        No change in performance detected.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) low mild

fine/strip/solid_long_u8_sse4_2
                        time:   [109.61 ns 109.66 ns 109.73 ns]
                        change: [+0.1200% +0.2040% +0.2894%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 8 outliers among 100 measurements (8.00%)
  1 (1.00%) low severe
  7 (7.00%) high mild

fine/strip/solid_long_u8_avx2
                        time:   [73.543 ns 73.562 ns 73.584 ns]
                        change: [-0.6422% -0.6064% -0.5697%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 6 outliers among 100 measurements (6.00%)
  4 (4.00%) high mild
  2 (2.00%) high severe

fine/pack/pack_block_u8_sse4_2
                        time:   [53.030 ns 53.036 ns 53.042 ns]
                        change: [-0.9191% -0.8391% -0.7582%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 100 measurements (4.00%)
  2 (2.00%) high mild
  2 (2.00%) high severe

fine/pack/pack_block_u8_avx2
                        time:   [51.374 ns 51.396 ns 51.419 ns]
                        change: [+0.0332% +0.1234% +0.2061%] (p = 0.01 < 0.05)
                        Change within noise threshold.
Found 6 outliers among 100 measurements (6.00%)
  1 (1.00%) low mild
  4 (4.00%) high mild
  1 (1.00%) high severe

fine/pack/unpack_block_u8_sse4_2
                        time:   [58.952 ns 59.012 ns 59.070 ns]
                        change: [+0.6420% +0.7483% +0.8528%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 3 outliers among 100 measurements (3.00%)
  3 (3.00%) high mild

fine/pack/unpack_block_u8_avx2
                        time:   [53.578 ns 53.642 ns 53.698 ns]
                        change: [-0.0433% +0.0352% +0.1108%] (p = 0.38 > 0.05)
                        No change in performance detected.
Found 11 outliers among 100 measurements (11.00%)
  11 (11.00%) high mild

fine/gradient/linear/opaque_u8_sse4_2
                        time:   [557.61 ns 558.23 ns 558.87 ns]
                        change: [+0.7263% +0.8539% +0.9784%] (p = 0.00 < 0.05)
                        Change within noise threshold.

fine/gradient/linear/opaque_u8_avx2
                        time:   [505.38 ns 505.90 ns 506.50 ns]
                        change: [+0.0086% +0.1247% +0.2477%] (p = 0.04 < 0.05)
                        Change within noise threshold.

fine/gradient/radial/opaque_u8_sse4_2
                        time:   [743.73 ns 744.43 ns 745.19 ns]
                        change: [+0.3918% +0.4679% +0.5444%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 9 outliers among 100 measurements (9.00%)
  8 (8.00%) high mild
  1 (1.00%) high severe

fine/gradient/radial/opaque_u8_avx2
                        time:   [601.95 ns 602.46 ns 602.95 ns]
                        change: [+0.0238% +0.1064% +0.1968%] (p = 0.02 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high severe

fine/gradient/radial/opaque_conical_u8_sse4_2
                        time:   [832.52 ns 833.21 ns 834.04 ns]
                        change: [+0.5035% +0.6115% +0.7269%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild

fine/gradient/radial/opaque_conical_u8_avx2
                        time:   [686.30 ns 686.90 ns 687.47 ns]
                        change: [-0.5709% -0.4683% -0.3724%] (p = 0.00 < 0.05)
                        Change within noise threshold.

fine/gradient/sweep/opaque_u8_sse4_2
                        time:   [1.3760 µs 1.3766 µs 1.3772 µs]
                        change: [+0.3403% +0.3900% +0.4404%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high severe

fine/gradient/sweep/opaque_u8_avx2
                        time:   [936.19 ns 936.37 ns 936.54 ns]
                        change: [-0.8035% -0.7796% -0.7545%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 100 measurements (4.00%)
  2 (2.00%) low mild
  2 (2.00%) high mild

fine/gradient/extend/pad_u8_sse4_2
                        time:   [555.41 ns 555.69 ns 555.97 ns]
                        change: [+0.1575% +0.2609% +0.3637%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 100 measurements (4.00%)
  3 (3.00%) high mild
  1 (1.00%) high severe

fine/gradient/extend/pad_u8_avx2
                        time:   [503.30 ns 503.50 ns 503.73 ns]
                        change: [-0.4940% -0.4254% -0.3577%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 12 outliers among 100 measurements (12.00%)
  1 (1.00%) low mild
  7 (7.00%) high mild
  4 (4.00%) high severe

fine/gradient/extend/repeat_u8_sse4_2
                        time:   [619.77 ns 620.29 ns 620.83 ns]
                        change: [-0.4322% -0.3339% -0.2341%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild

fine/gradient/extend/repeat_u8_avx2
                        time:   [535.85 ns 535.99 ns 536.14 ns]
                        change: [-0.3477% -0.3064% -0.2643%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 100 measurements (4.00%)
  1 (1.00%) high mild
  3 (3.00%) high severe

fine/gradient/extend/reflect_u8_sse4_2
                        time:   [687.61 ns 688.05 ns 688.49 ns]
                        change: [+0.0514% +0.1413% +0.2313%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) low mild

fine/gradient/extend/reflect_u8_avx2
                        time:   [589.12 ns 589.29 ns 589.48 ns]
                        change: [+0.0205% +0.0632% +0.1067%] (p = 0.01 < 0.05)
                        Change within noise threshold.
Found 6 outliers among 100 measurements (6.00%)
  1 (1.00%) low mild
  5 (5.00%) high mild

fine/gradient/many_stops_u8_sse4_2
                        time:   [806.94 ns 807.21 ns 807.47 ns]
                        change: [-0.1546% -0.0975% -0.0431%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 2 outliers among 100 measurements (2.00%)
  1 (1.00%) low mild
  1 (1.00%) high mild

fine/gradient/many_stops_u8_avx2
                        time:   [670.90 ns 671.05 ns 671.21 ns]
                        change: [-0.2032% -0.1639% -0.1243%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 8 outliers among 100 measurements (8.00%)
  1 (1.00%) low mild
  5 (5.00%) high mild
  2 (2.00%) high severe

fine/gradient/transparent_u8_sse4_2
                        time:   [740.42 ns 740.63 ns 740.86 ns]
                        change: [+0.1559% +0.2519% +0.3574%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 9 outliers among 100 measurements (9.00%)
  7 (7.00%) high mild
  2 (2.00%) high severe

fine/gradient/transparent_u8_avx2
                        time:   [639.12 ns 639.26 ns 639.41 ns]
                        change: [-0.1199% -0.0798% -0.0382%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 3 outliers among 100 measurements (3.00%)
  1 (1.00%) high mild
  2 (2.00%) high severe

fine/rounded_blurred_rect/with_transform_u8_sse4_2
                        time:   [10.516 µs 10.518 µs 10.520 µs]
                        change: [+0.0222% +0.0555% +0.0877%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 7 outliers among 100 measurements (7.00%)
  5 (5.00%) low mild
  2 (2.00%) high mild

fine/rounded_blurred_rect/with_transform_u8_avx2
                        time:   [9.7963 µs 9.7986 µs 9.8012 µs]
                        change: [-0.1660% -0.1379% -0.1114%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 5 outliers among 100 measurements (5.00%)
  5 (5.00%) high mild

fine/rounded_blurred_rect/no_transform_u8_sse4_2
                        time:   [10.645 µs 10.648 µs 10.652 µs]
                        change: [+0.0527% +0.0989% +0.1463%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 100 measurements (4.00%)
  3 (3.00%) high mild
  1 (1.00%) high severe

fine/rounded_blurred_rect/no_transform_u8_avx2
                        time:   [9.9414 µs 9.9447 µs 9.9481 µs]
                        change: [+0.0641% +0.1083% +0.1553%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 3 outliers among 100 measurements (3.00%)
  2 (2.00%) high mild
  1 (1.00%) high severe

fine/blend/normal_u8_sse4_2
                        time:   [47.427 ns 47.462 ns 47.499 ns]
                        change: [+0.4121% +0.4897% +0.5677%] (p = 0.00 < 0.05)
                        Change within noise threshold.

fine/blend/normal_u8_avx2
                        time:   [26.026 ns 26.040 ns 26.057 ns]
                        change: [-18.332% -18.287% -18.234%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 14 outliers among 100 measurements (14.00%)
  10 (10.00%) high mild
  4 (4.00%) high severe

fine/blend/multiply_u8_sse4_2
                        time:   [2.3564 µs 2.4088 µs 2.4589 µs]
                        change: [-5.4069% -3.3899% -1.1018%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) low mild

fine/blend/multiply_u8_avx2
                        time:   [1.2784 µs 1.2791 µs 1.2798 µs]
                        change: [-0.1201% +0.0756% +0.2390%] (p = 0.43 > 0.05)
                        No change in performance detected.
Found 6 outliers among 100 measurements (6.00%)
  5 (5.00%) high mild
  1 (1.00%) high severe

fine/blend/screen_u8_sse4_2
                        time:   [1.8329 µs 1.8761 µs 1.9252 µs]
                        change: [-3.2994% -1.8209% -0.1020%] (p = 0.03 < 0.05)
                        Change within noise threshold.
Found 21 outliers among 100 measurements (21.00%)
  5 (5.00%) high mild
  16 (16.00%) high severe

fine/blend/screen_u8_avx2
                        time:   [1.2665 µs 1.2682 µs 1.2705 µs]
                        change: [-4.3774% -3.9390% -3.5269%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 11 outliers among 100 measurements (11.00%)
  10 (10.00%) high mild
  1 (1.00%) high severe

fine/blend/overlay_u8_sse4_2
                        time:   [2.0567 µs 2.0648 µs 2.0736 µs]
                        change: [-3.1647% -2.4825% -1.8082%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
  2 (2.00%) high mild

fine/blend/overlay_u8_avx2
                        time:   [1.2911 µs 1.2938 µs 1.2963 µs]
                        change: [-0.5660% -0.3972% -0.2219%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 5 outliers among 100 measurements (5.00%)
  4 (4.00%) high mild
  1 (1.00%) high severe

fine/blend/darken_u8_sse4_2
                        time:   [1.6757 µs 1.6919 µs 1.7099 µs]
                        change: [+4.2429% +5.4786% +6.7590%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 18 outliers among 100 measurements (18.00%)
  11 (11.00%) high mild
  7 (7.00%) high severe

fine/blend/darken_u8_avx2
                        time:   [1.3373 µs 1.3398 µs 1.3427 µs]
                        change: [+0.8885% +1.0691% +1.2835%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 6 outliers among 100 measurements (6.00%)
  6 (6.00%) high mild

fine/blend/lighten_u8_sse4_2
                        time:   [1.6637 µs 1.6707 µs 1.6792 µs]
                        change: [-5.9808% -4.5247% -3.1404%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high severe

fine/blend/lighten_u8_avx2
                        time:   [1.3511 µs 1.3515 µs 1.3519 µs]
                        change: [-0.1222% -0.0684% -0.0163%] (p = 0.01 < 0.05)
                        Change within noise threshold.
Found 15 outliers among 100 measurements (15.00%)
  5 (5.00%) low mild
  10 (10.00%) high mild

fine/blend/color_dodge_u8_sse4_2
                        time:   [6.1797 µs 6.2307 µs 6.2868 µs]
                        change: [+1.8494% +2.6377% +3.4231%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 3 outliers among 100 measurements (3.00%)
  3 (3.00%) high mild

fine/blend/color_dodge_u8_avx2
                        time:   [4.2804 µs 4.2826 µs 4.2847 µs]
                        change: [-3.1557% -3.1048% -3.0553%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 7 outliers among 100 measurements (7.00%)
  4 (4.00%) low mild
  3 (3.00%) high mild

fine/blend/color_burn_u8_sse4_2
                        time:   [6.1894 µs 6.2227 µs 6.2594 µs]
                        change: [+0.5146% +1.1612% +1.8307%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 2 outliers among 100 measurements (2.00%)
  2 (2.00%) high mild

fine/blend/color_burn_u8_avx2
                        time:   [4.4286 µs 4.4302 µs 4.4319 µs]
                        change: [-1.0331% -0.9792% -0.9216%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) low mild

fine/blend/hard_light_u8_sse4_2
                        time:   [2.2752 µs 2.3004 µs 2.3327 µs]
                        change: [-2.6572% -0.9095% +0.8145%] (p = 0.31 > 0.05)
                        No change in performance detected.
Found 2 outliers among 100 measurements (2.00%)
  1 (1.00%) high mild
  1 (1.00%) high severe

fine/blend/hard_light_u8_avx2
                        time:   [1.3132 µs 1.3157 µs 1.3179 µs]
                        change: [+0.1370% +0.3661% +0.6023%] (p = 0.00 < 0.05)
                        Change within noise threshold.

fine/blend/soft_light_u8_sse4_2
                        time:   [5.9459 µs 5.9679 µs 5.9887 µs]
                        change: [+0.4284% +1.4208% +2.3797%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 2 outliers among 100 measurements (2.00%)
  2 (2.00%) high mild

fine/blend/soft_light_u8_avx2
                        time:   [5.1935 µs 5.1976 µs 5.2018 µs]
                        change: [+0.1220% +0.1683% +0.2222%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild

fine/blend/difference_u8_sse4_2
                        time:   [2.2687 µs 2.4010 µs 2.5483 µs]
                        change: [-4.6277% -1.2819% +2.4445%] (p = 0.50 > 0.05)
                        No change in performance detected.
Found 15 outliers among 100 measurements (15.00%)
  1 (1.00%) high mild
  14 (14.00%) high severe

fine/blend/difference_u8_avx2
                        time:   [1.2534 µs 1.2555 µs 1.2576 µs]
                        change: [+0.7522% +0.9206% +1.0865%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) low mild

fine/blend/exclusion_u8_sse4_2
                        time:   [1.7478 µs 1.8423 µs 1.9440 µs]
                        change: [+4.5130% +7.0576% +10.322%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 13 outliers among 100 measurements (13.00%)
  13 (13.00%) high severe

fine/blend/exclusion_u8_avx2
                        time:   [1.3728 µs 1.3749 µs 1.3770 µs]
                        change: [-0.1715% -0.0236% +0.1273%] (p = 0.76 > 0.05)
                        No change in performance detected.
Found 9 outliers among 100 measurements (9.00%)
  8 (8.00%) high mild
  1 (1.00%) high severe

fine/blend/hue_u8_sse4_2
                        time:   [8.8913 µs 8.9303 µs 8.9724 µs]
                        change: [+4.6586% +4.9657% +5.3021%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 6 outliers among 100 measurements (6.00%)
  2 (2.00%) high mild
  4 (4.00%) high severe

fine/blend/hue_u8_avx2  time:   [7.5768 µs 7.5825 µs 7.5882 µs]
                        change: [+1.4558% +1.5616% +1.6679%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 8 outliers among 100 measurements (8.00%)
  4 (4.00%) low mild
  2 (2.00%) high mild
  2 (2.00%) high severe

fine/blend/saturation_u8_sse4_2
                        time:   [8.6111 µs 8.6221 µs 8.6332 µs]
                        change: [-2.2313% -1.9316% -1.6141%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild

fine/blend/saturation_u8_avx2
                        time:   [7.5972 µs 7.6032 µs 7.6087 µs]
                        change: [+1.4421% +1.5354% +1.6274%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 13 outliers among 100 measurements (13.00%)
  1 (1.00%) low severe
  5 (5.00%) low mild
  5 (5.00%) high mild
  2 (2.00%) high severe

fine/blend/color_u8_sse4_2
                        time:   [7.2227 µs 7.2323 µs 7.2421 µs]
                        change: [-6.0269% -5.6816% -5.3208%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
  3 (3.00%) high mild

fine/blend/color_u8_avx2
                        time:   [6.2791 µs 6.2829 µs 6.2872 µs]
                        change: [-6.1108% -6.0246% -5.9408%] (p = 0.00 < 0.05)
                        Performance has improved.

fine/blend/luminosity_u8_sse4_2
                        time:   [7.0038 µs 7.0183 µs 7.0337 µs]
                        change: [+0.8341% +1.0873% +1.3563%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 3 outliers among 100 measurements (3.00%)
  3 (3.00%) high mild

fine/blend/luminosity_u8_avx2
                        time:   [6.3247 µs 6.3258 µs 6.3270 µs]
                        change: [-5.6774% -5.6139% -5.5502%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 6 outliers among 100 measurements (6.00%)
  4 (4.00%) high mild
  2 (2.00%) high severe

fine/blend/clear_u8_sse4_2
                        time:   [769.66 ns 800.20 ns 838.71 ns]
                        change: [-70.956% -70.001% -69.057%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 14 outliers among 100 measurements (14.00%)
  14 (14.00%) high severe

fine/blend/clear_u8_avx2
                        time:   [414.78 ns 427.33 ns 441.23 ns]
                        change: [+13.160% +18.047% +23.021%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 3 outliers among 100 measurements (3.00%)
  3 (3.00%) high mild

fine/blend/copy_u8_sse4_2
                        time:   [701.26 ns 709.18 ns 717.76 ns]
                        change: [-2.5891% -1.7479% -0.8735%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 6 outliers among 100 measurements (6.00%)
  6 (6.00%) high mild

fine/blend/copy_u8_avx2 time:   [332.29 ns 333.05 ns 334.00 ns]
                        change: [-2.0480% -1.8584% -1.6574%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 9 outliers among 100 measurements (9.00%)
  4 (4.00%) high mild
  5 (5.00%) high severe

fine/blend/dest_u8_sse4_2
                        time:   [621.74 ns 623.40 ns 625.22 ns]
                        change: [-3.5330% -3.2362% -2.9137%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
  1 (1.00%) low mild
  3 (3.00%) high mild

fine/blend/dest_u8_avx2 time:   [333.83 ns 334.01 ns 334.19 ns]
                        change: [+1.9951% +2.0718% +2.1501%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 9 outliers among 100 measurements (9.00%)
  3 (3.00%) low mild
  6 (6.00%) high mild

fine/blend/dest_over_u8_sse4_2
                        time:   [1.5267 µs 1.5732 µs 1.6150 µs]
                        change: [-25.806% -23.333% -20.881%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) low mild

fine/blend/dest_over_u8_avx2
                        time:   [405.62 ns 405.95 ns 406.31 ns]
                        change: [+0.1987% +0.2542% +0.3133%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 5 outliers among 100 measurements (5.00%)
  5 (5.00%) high mild

fine/blend/src_out_u8_sse4_2
                        time:   [732.45 ns 755.53 ns 782.43 ns]
                        change: [+14.329% +16.534% +19.247%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 9 outliers among 100 measurements (9.00%)
  6 (6.00%) high mild
  3 (3.00%) high severe

fine/blend/src_out_u8_avx2
                        time:   [351.70 ns 352.03 ns 352.39 ns]
                        change: [-0.3847% -0.2523% -0.1181%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 100 measurements (4.00%)
  4 (4.00%) high mild

fine/blend/src_in_u8_sse4_2
                        time:   [999.38 ns 1.0428 µs 1.0893 µs]
                        change: [+3.5879% +8.5566% +13.742%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 3 outliers among 100 measurements (3.00%)
  3 (3.00%) high mild

fine/blend/src_in_u8_avx2
                        time:   [345.04 ns 345.78 ns 346.67 ns]
                        change: [+0.1911% +0.3431% +0.5116%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 9 outliers among 100 measurements (9.00%)
  8 (8.00%) high mild
  1 (1.00%) high severe

fine/blend/dest_in_u8_sse4_2
                        time:   [674.01 ns 676.37 ns 679.00 ns]
                        change: [+5.6029% +5.9960% +6.4338%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 7 outliers among 100 measurements (7.00%)
  7 (7.00%) high mild

fine/blend/dest_in_u8_avx2
                        time:   [349.30 ns 349.83 ns 350.43 ns]
                        change: [-0.4608% -0.2749% -0.0827%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 17 outliers among 100 measurements (17.00%)
  15 (15.00%) high mild
  2 (2.00%) high severe

fine/blend/dest_out_u8_sse4_2
                        time:   [632.70 ns 636.54 ns 640.93 ns]
                        change: [-2.5117% -1.9782% -1.4275%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
  4 (4.00%) high mild

fine/blend/dest_out_u8_avx2
                        time:   [352.07 ns 352.34 ns 352.63 ns]
                        change: [-0.7760% -0.6166% -0.4429%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 12 outliers among 100 measurements (12.00%)
  3 (3.00%) high mild
  9 (9.00%) high severe

fine/blend/src_atop_u8_sse4_2
                        time:   [2.3023 µs 2.3363 µs 2.3695 µs]
                        change: [+4.7026% +6.5639% +8.5803%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 10 outliers among 100 measurements (10.00%)
  1 (1.00%) low severe
  9 (9.00%) low mild

fine/blend/src_atop_u8_avx2
                        time:   [432.57 ns 432.93 ns 433.31 ns]
                        change: [-0.4606% -0.3826% -0.2974%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 3 outliers among 100 measurements (3.00%)
  1 (1.00%) high mild
  2 (2.00%) high severe

fine/blend/dest_atop_u8_sse4_2
                        time:   [1.7374 µs 1.7665 µs 1.7971 µs]
                        change: [-2.4494% -0.8432% +0.8725%] (p = 0.32 > 0.05)
                        No change in performance detected.

fine/blend/dest_atop_u8_avx2
                        time:   [434.69 ns 434.82 ns 434.95 ns]
                        change: [-0.1680% -0.1180% -0.0665%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 5 outliers among 100 measurements (5.00%)
  1 (1.00%) low mild
  4 (4.00%) high mild

fine/blend/xor_u8_sse4_2
                        time:   [2.3208 µs 2.3292 µs 2.3370 µs]
                        change: [-2.2415% -1.8288% -1.4306%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) low mild

fine/blend/xor_u8_avx2  time:   [443.22 ns 443.38 ns 443.56 ns]
                        change: [-0.0584% +0.0465% +0.1509%] (p = 0.39 > 0.05)
                        No change in performance detected.
Found 6 outliers among 100 measurements (6.00%)
  6 (6.00%) high mild

fine/blend/plus_u8_sse4_2
                        time:   [2.3860 µs 2.3911 µs 2.3958 µs]
                        change: [+270.86% +271.63% +272.37%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 9 outliers among 100 measurements (9.00%)
  2 (2.00%) low severe
  7 (7.00%) low mild

fine/blend/plus_u8_avx2 time:   [367.90 ns 368.11 ns 368.34 ns]
                        change: [-0.4824% -0.3732% -0.2575%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 13 outliers among 100 measurements (13.00%)
  1 (1.00%) low mild
  8 (8.00%) high mild
  4 (4.00%) high severe

fine/blend/plus_lighter_u8_sse4_2
                        time:   [2.3564 µs 2.3648 µs 2.3728 µs]
                        change: [+261.08% +262.86% +264.57%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 6 outliers among 100 measurements (6.00%)
  1 (1.00%) low severe
  4 (4.00%) low mild
  1 (1.00%) high mild

fine/blend/plus_lighter_u8_avx2
                        time:   [369.10 ns 369.33 ns 369.56 ns]
                        change: [-0.0191% +0.0484% +0.1121%] (p = 0.16 > 0.05)
                        No change in performance detected.

fine/image/transform/none_u8_sse4_2
                        time:   [613.44 ns 613.55 ns 613.66 ns]
                        change: [+0.0219% +0.0728% +0.1200%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild

fine/image/transform/none_u8_avx2
                        time:   [615.22 ns 615.36 ns 615.57 ns]
                        change: [-0.2091% -0.1639% -0.1204%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 14 outliers among 100 measurements (14.00%)
  1 (1.00%) low severe
  1 (1.00%) low mild
  1 (1.00%) high mild
  11 (11.00%) high severe

fine/image/transform/scale_u8_sse4_2
                        time:   [612.55 ns 612.66 ns 612.79 ns]
                        change: [+0.0045% +0.0545% +0.1001%] (p = 0.03 < 0.05)
                        Change within noise threshold.
Found 4 outliers among 100 measurements (4.00%)
  1 (1.00%) low mild
  3 (3.00%) high mild

fine/image/transform/scale_u8_avx2
                        time:   [615.62 ns 615.82 ns 616.04 ns]
                        change: [-0.5152% -0.4771% -0.4382%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 19 outliers among 100 measurements (19.00%)
  2 (2.00%) low mild
  5 (5.00%) high mild
  12 (12.00%) high severe

fine/image/transform/rotate_u8_sse4_2
                        time:   [776.76 ns 776.84 ns 776.95 ns]
                        change: [-5.8679% -5.8284% -5.7838%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 6 outliers among 100 measurements (6.00%)
  5 (5.00%) high mild
  1 (1.00%) high severe

fine/image/transform/rotate_u8_avx2
                        time:   [795.37 ns 795.48 ns 795.62 ns]
                        change: [-5.4546% -5.4328% -5.4112%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
  3 (3.00%) high mild
  1 (1.00%) high severe

fine/image/quality/low_u8_sse4_2
                        time:   [612.59 ns 612.67 ns 612.77 ns]
                        change: [-0.1397% -0.0861% -0.0366%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 2 outliers among 100 measurements (2.00%)
  1 (1.00%) low severe
  1 (1.00%) high mild

fine/image/quality/low_u8_avx2
                        time:   [615.54 ns 615.76 ns 616.00 ns]
                        change: [-0.8438% -0.7706% -0.6984%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 13 outliers among 100 measurements (13.00%)
  1 (1.00%) low mild
  12 (12.00%) high mild

fine/image/quality/medium_u8_sse4_2
                        time:   [2.4362 µs 2.4374 µs 2.4387 µs]
                        change: [-0.0651% +0.0029% +0.0677%] (p = 0.93 > 0.05)
                        No change in performance detected.
Found 2 outliers among 100 measurements (2.00%)
  1 (1.00%) high mild
  1 (1.00%) high severe

fine/image/quality/medium_u8_avx2
                        time:   [5.2194 µs 5.3467 µs 5.4814 µs]
                        change: [-1.4855% +2.1459% +6.2357%] (p = 0.27 > 0.05)
                        No change in performance detected.
Found 6 outliers among 100 measurements (6.00%)
  6 (6.00%) high mild

fine/image/quality/high_u8_sse4_2
                        time:   [13.729 µs 13.730 µs 13.731 µs]
                        change: [-0.1407% -0.1233% -0.1068%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 3 outliers among 100 measurements (3.00%)
  3 (3.00%) high mild

fine/image/quality/high_u8_avx2
                        time:   [11.074 µs 11.082 µs 11.092 µs]
                        change: [+2.5554% +2.6397% +2.7251%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 12 outliers among 100 measurements (12.00%)
  2 (2.00%) high mild
  10 (10.00%) high severe

fine/image/extend/pad_u8_sse4_2
                        time:   [612.47 ns 612.56 ns 612.66 ns]
                        change: [-0.8765% -0.8087% -0.7401%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 7 outliers among 100 measurements (7.00%)
  6 (6.00%) high mild
  1 (1.00%) high severe

fine/image/extend/pad_u8_avx2
                        time:   [615.30 ns 615.53 ns 615.79 ns]
                        change: [-0.4973% -0.4500% -0.4024%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 14 outliers among 100 measurements (14.00%)
  7 (7.00%) high mild
  7 (7.00%) high severe

fine/image/extend/repeat_u8_sse4_2
                        time:   [603.49 ns 603.61 ns 603.74 ns]
                        change: [-3.3875% -3.3541% -3.3216%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 5 outliers among 100 measurements (5.00%)
  2 (2.00%) high mild
  3 (3.00%) high severe

fine/image/extend/repeat_u8_avx2
                        time:   [613.32 ns 613.77 ns 614.21 ns]
                        change: [-0.9102% -0.8460% -0.7816%] (p = 0.00 < 0.05)
                        Change within noise threshold.

fine/image/extend/reflect_u8_sse4_2
                        time:   [970.34 ns 971.06 ns 971.86 ns]
                        change: [-10.314% -10.247% -10.177%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild

fine/image/extend/reflect_u8_avx2
                        time:   [1.0065 µs 1.0070 µs 1.0075 µs]
                        change: [-11.485% -11.426% -11.360%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
  2 (2.00%) high mild

sort/Ghostscript_Tiger  time:   [92.455 µs 92.648 µs 92.802 µs]
                        change: [+0.2407% +0.3925% +0.5602%] (p = 0.00 < 0.05)
                        Change within noise threshold.
sort/coat_of_arms       time:   [1.0200 ms 1.0218 ms 1.0235 ms]
                        change: [-0.0576% +0.0624% +0.1902%] (p = 0.34 > 0.05)
                        No change in performance detected.
sort/Saimaa_Canal_(map) time:   [3.1034 ms 3.1079 ms 3.1119 ms]
                        change: [+0.5295% +0.6388% +0.7504%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 50 measurements (2.00%)
  1 (2.00%) high mild
sort/heraldry           time:   [659.22 µs 659.48 µs 659.81 µs]
                        change: [-0.6923% -0.6446% -0.5950%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 5 outliers among 50 measurements (10.00%)
  2 (4.00%) high mild
  3 (6.00%) high severe

images/overlapping      time:   [2.8189 ms 2.8214 ms 2.8239 ms]
                        change: [-0.0488% +0.0516% +0.1439%] (p = 0.26 > 0.05)
                        No change in performance detected.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild

Codex analysis of assembly changes

Verdict

The refactor introduces a small, localized performance regression in vello_cpu’s generated x86 code. Most load/store-heavy kernels are unchanged, but several hot loops now have less favorable optimization.

Compared revisions:

• Before: 0a7408a2217767e2232140309bdb1d8eb97050d0
• After: d7ce2b1d13728b1a19171b3a08a0b000f0e4f606
• Rust 1.97.1, LLVM 22.1.6, optimized release builds
• Checked default, f32, and all-feature configurations

What stayed unchanged

The straightforward load/store users—copy_solid, apply_mask, apply_tint, and alpha-composite loops—produce identical assembly across SSE2, SSE4.2, AVX2, and AVX-512. The SIMD code in vello_common is also unchanged.

This confirms the new generic array conversion implementations inline successfully in the normal cases.

What regressed

The most notable change is the u8 blend loop in [lowp/mod.rs (line 478)](/home/shnatsel/Code/vello/sparse_strips/vello_cpu/src/fine/lowp/mod.rs:478):

• Previously LLVM hoisted default_mix outside the loop and emitted two specialized loop bodies.
• Afterwards it emits one smaller loop with the default_mix branch inside every 32-byte SIMD iteration.
• This affects SSE2, SSE4.2, AVX2, and AVX-512.
• Code size improves, but steady-state throughput is likely slightly worse because of the extra per-chunk branch and additional stack traffic.

There are also register-allocation regressions in the high-quality filtered-image path in [image.rs (line 181)](/home/shnatsel/Code/vello/sparse_strips/vello_cpu/src/fine/common/image.rs:181):

• AVX2 paint_f32: 780 → 793 instructions; stack frame 552 → 600 bytes.
• AVX2 paint_u8: 816 → 829 instructions.
• Image bounds are repeatedly compared against a stack-resident length instead of a register.
• AVX-512 gradient painting grows from 225 → 249 instructions.

Static SIMD instruction totals actually decrease slightly because LLVM merges duplicated loops:

Level	Static instruction delta
SSE2	−63
SSE4.2	−53
AVX2	−37
AVX-512	−32
Fallback	unchanged

That reduction is primarily a code-size improvement, not a throughput improvement: some once-per-call decisions moved into hot loops.

So I would classify the result as mostly unchanged, but slightly worse in important u8 blend and filtered-image paths. It does not look like a broad load/store regression, but it is enough that I would not call the refactor performance-neutral without benchmark confirmation. No workspace files were changed during the analysis.