Move the TreeDesc fields and finish adding cache line markers to State

[brian-pane]

No performance regression measured at compression levels 1-9 on my test system

[brian-pane]

This PR is stacked on top of #278. It groups the smaller fields into cache lines to make it easier to do future changes such as reducing usize to u16 and removing fields that are faster to calculate.

[brian-pane]

Rebasing fixed some of the problems, but I can’t reproduce the CI test failure on aarch64

[folkertdev]

I can't reproduce them either, maybe github CI is having issues? I did just retry them without any success, but maybe it takes a little while. (the only reason I can think of is some sort of file/disk corruption)

[bjorn3]

The github actions problem should be resolved: https://www.githubstatus.com/incidents/dk61qxd21mtl However I just retried CI an it is still broken.

[folkertdev]

well so in https://github.com/trifectatechfoundation/zlib-rs/pull/280/files I just made whitespace changes and CI still fails with the same error as here. I just don't see how that could be us.

[folkertdev]

I'm getting some good instruction count wins locally now, and actual wall_time in some cases

Benchmark 2 (62 runs): target/release/examples/blogpost-compress 1 rs silesia-small.tar
  measurement          mean ± σ            min … max           outliers         delta
  wall_time          81.7ms ±  935us    80.4ms … 87.4ms          5 ( 8%)          -  1.3% ±  0.4%
  peak_rss           26.7MB ± 50.5KB    26.6MB … 26.7MB         11 (18%)          +  0.4% ±  0.1%
  cpu_cycles          299M  ± 2.72M      296M  …  315M           2 ( 3%)        ⚡-  1.6% ±  0.3%
  instructions        601M  ±  274       601M  …  601M           0 ( 0%)        ⚡-  9.1% ±  0.0%
  cache_references   19.9M  ±  166K     19.7M  … 20.6M           2 ( 3%)          +  0.5% ±  0.3%
  cache_misses        393K  ± 84.7K      293K  …  714K           1 ( 2%)        ⚡- 10.7% ±  7.1%
  branch_misses      2.98M  ± 4.51K     2.97M  … 2.99M           0 ( 0%)        💩+  1.5% ±  0.1%
Benchmark 2 (37 runs): target/release/examples/blogpost-compress 2 rs silesia-small.tar
  measurement          mean ± σ            min … max           outliers         delta
  wall_time           135ms ± 3.56ms     133ms …  153ms          2 ( 5%)          -  1.3% ±  0.9%
  peak_rss           25.0MB ± 59.0KB    24.9MB … 25.0MB          0 ( 0%)          -  0.0% ±  0.1%
  cpu_cycles          546M  ± 15.4M      538M  …  629M           2 ( 5%)          -  1.4% ±  1.0%
  instructions       1.12G  ±  280      1.12G  … 1.12G           0 ( 0%)        ⚡-  4.9% ±  0.0%
  cache_references   34.3M  ±  394K     33.7M  … 35.5M           1 ( 3%)          +  0.0% ±  0.5%
  cache_misses       1.05M  ±  233K      789K  … 1.65M           2 ( 5%)          +  3.4% ± 10.0%
  branch_misses      7.01M  ± 3.71K     7.00M  … 7.02M           1 ( 3%)          +  0.7% ±  0.0%
Benchmark 2 (34 runs): target/release/examples/blogpost-compress 3 rs silesia-small.tar
  measurement          mean ± σ            min … max           outliers         delta
  wall_time           151ms ± 1.55ms     148ms …  155ms          0 ( 0%)        ⚡-  2.9% ±  0.5%
  peak_rss           24.7MB ± 66.4KB    24.6MB … 24.8MB          0 ( 0%)          -  0.1% ±  0.1%
  cpu_cycles          615M  ± 6.70M      607M  …  637M           1 ( 3%)        ⚡-  3.4% ±  0.5%
  instructions       1.46G  ±  250      1.46G  … 1.46G           0 ( 0%)        ⚡-  3.7% ±  0.0%
  cache_references   43.9M  ±  630K     43.1M  … 45.3M           5 (15%)          -  0.7% ±  0.6%
  cache_misses       1.16M  ±  338K      850K  … 2.21M           7 (21%)          +  7.8% ± 14.1%
  branch_misses      7.86M  ± 4.05K     7.85M  … 7.87M           0 ( 0%)        💩+  1.0% ±  0.0%
Benchmark 2 (29 runs): target/release/examples/blogpost-compress 4 rs silesia-small.tar
  measurement          mean ± σ            min … max           outliers         delta
  wall_time           173ms ± 2.75ms     172ms …  186ms          2 ( 7%)        ⚡-  3.4% ±  1.6%
  peak_rss           24.6MB ± 61.7KB    24.5MB … 24.6MB          0 ( 0%)          -  0.1% ±  0.1%
  cpu_cycles          722M  ± 8.77M      715M  …  762M           3 (10%)        ⚡-  3.4% ±  1.4%
  instructions       1.53G  ±  293      1.53G  … 1.53G           0 ( 0%)        ⚡-  3.5% ±  0.0%
  cache_references   62.7M  ±  701K     61.9M  … 65.1M           1 ( 3%)          -  0.4% ±  0.6%
  cache_misses       1.82M  ±  356K     1.42M  … 2.79M           2 ( 7%)        💩+ 28.4% ± 12.5%
  branch_misses      8.54M  ± 17.8K     8.52M  … 8.60M           5 (17%)          +  0.9% ±  0.1%
Benchmark 1 (26 runs): target/release/examples/compress-baseline 5 rs silesia-small.tar
  measurement          mean ± σ            min … max           outliers         delta
  wall_time           189ms ± 1.46ms     187ms …  193ms          1 ( 4%)        ⚡-  2.2% ±  0.7%
  peak_rss           24.6MB ± 47.4KB    24.5MB … 24.6MB          4 (15%)          +  0.1% ±  0.1%
  cpu_cycles          792M  ± 5.36M      785M  …  808M           0 ( 0%)        ⚡-  2.5% ±  0.8%
  instructions       1.74G  ±  273      1.74G  … 1.74G           0 ( 0%)        ⚡-  2.9% ±  0.0%
  cache_references   68.6M  ±  680K     67.8M  … 69.9M           0 ( 0%)          +  0.1% ±  0.6%
  cache_misses       1.87M  ±  344K     1.42M  … 2.56M           0 ( 0%)        💩+ 17.6% ± 13.8%
  branch_misses      9.15M  ± 7.72K     9.14M  … 9.17M           0 ( 0%)          +  0.9% ±  0.0%
Benchmark 2 (22 runs): target/release/examples/blogpost-compress 6 rs silesia-small.tar
  measurement          mean ± σ            min … max           outliers         delta
  wall_time           228ms ± 3.56ms     224ms …  238ms          3 (14%)          -  0.6% ±  0.7%
  peak_rss           24.6MB ± 51.7KB    24.5MB … 24.6MB          4 (18%)          +  0.0% ±  0.1%
  cpu_cycles          975M  ± 14.5M      962M  … 1.02G           1 ( 5%)          -  0.9% ±  0.7%
  instructions       1.88G  ±  313      1.88G  … 1.88G           0 ( 0%)        ⚡-  2.6% ±  0.0%
  cache_references    107M  ±  923K      105M  …  109M           0 ( 0%)          +  0.9% ±  0.4%
  cache_misses       2.53M  ±  484K     1.98M  … 3.65M           0 ( 0%)        💩+ 18.3% ± 11.9%
  branch_misses      9.37M  ± 21.1K     9.35M  … 9.42M           3 (14%)          +  1.0% ±  0.1%
Benchmark 2 (12 runs): target/release/examples/blogpost-compress 9 rs silesia-small.tar
  measurement          mean ± σ            min … max           outliers         delta
  wall_time           431ms ± 5.06ms     428ms …  447ms          1 ( 8%)          -  0.4% ±  0.7%
  peak_rss           24.4MB ± 87.6KB    24.2MB … 24.5MB          0 ( 0%)          -  0.1% ±  0.3%
  cpu_cycles         1.91G  ± 20.5M     1.90G  … 1.98G           1 ( 8%)          -  0.3% ±  0.7%
  instructions       3.32G  ±  713      3.32G  … 3.32G           1 ( 8%)        ⚡-  1.8% ±  0.0%
  cache_references    197M  ± 1.23M      195M  …  199M           1 ( 8%)          +  0.5% ±  0.4%
  cache_misses       3.56M  ±  728K     2.51M  … 4.65M           0 ( 0%)        💩+ 77.8% ± 25.5%
  branch_misses      19.0M  ± 37.5K     18.9M  … 19.1M           1 ( 8%)          -  0.5% ±  0.2%