ARROW-10120: [C++] Add two-level nested Parquet read to Arrow benchmarks

[pitrou]

No description provided.

[pitrou]

(wrong benchmark numbers deleted)

[pitrou]

Benchmark numbers on AMD Ryzen:

BM_ReadStructColumn/0                    1770216 ns      1768552 ns          391 bytes_per_second=6.62618G/s items_per_second=592.901M/s
BM_ReadStructColumn/1                    6769274 ns      6762710 ns          104 bytes_per_second=1.73285G/s items_per_second=155.053M/s
BM_ReadStructColumn/50                  10427316 ns     10419932 ns           65 bytes_per_second=1.12465G/s items_per_second=100.632M/s
BM_ReadStructColumn/99                   2910162 ns      2907881 ns          239 bytes_per_second=4.03G/s items_per_second=360.598M/s

BM_ReadStructOfStructColumn/0            3916658 ns      3912652 ns          179 bytes_per_second=5.99018G/s items_per_second=267.996M/s
BM_ReadStructOfStructColumn/1           13437999 ns     13425673 ns           52 bytes_per_second=1.74572G/s items_per_second=78.1023M/s
BM_ReadStructOfStructColumn/50          15649910 ns     15638181 ns           44 bytes_per_second=1.49874G/s items_per_second=67.0523M/s
BM_ReadStructOfStructColumn/99           7022917 ns      7015954 ns           95 bytes_per_second=3.3406G/s items_per_second=149.456M/s

BM_ReadStructOfListColumn/0             17565225 ns     17553576 ns           40 bytes_per_second=683.621M/s items_per_second=59.7357M/s
BM_ReadStructOfListColumn/1             20743098 ns     20729469 ns           34 bytes_per_second=578.886M/s items_per_second=50.5838M/s
BM_ReadStructOfListColumn/50            35540775 ns     35528805 ns           19 bytes_per_second=337.754M/s items_per_second=29.5134M/s
BM_ReadStructOfListColumn/99            14650196 ns     14640694 ns           47 bytes_per_second=819.633M/s items_per_second=71.6207M/s

BM_ReadListColumn/0                      7604549 ns      7601412 ns           92 bytes_per_second=1052.44M/s items_per_second=137.945M/s
BM_ReadListColumn/1                      9118144 ns      9114307 ns           77 bytes_per_second=877.741M/s items_per_second=115.047M/s
BM_ReadListColumn/50                    16743022 ns     16738957 ns           41 bytes_per_second=477.927M/s items_per_second=62.6428M/s
BM_ReadListColumn/99                     6798257 ns      6794603 ns          104 bytes_per_second=1.14981G/s items_per_second=154.325M/s

BM_ReadListOfStructColumn/0             15266355 ns     15256162 ns           45 bytes_per_second=786.567M/s items_per_second=68.7313M/s
BM_ReadListOfStructColumn/1             19382854 ns     19372448 ns           36 bytes_per_second=619.436M/s items_per_second=54.1272M/s
BM_ReadListOfStructColumn/50            30720332 ns     30711989 ns           23 bytes_per_second=390.727M/s items_per_second=34.1422M/s
BM_ReadListOfStructColumn/99            12476630 ns     12467956 ns           56 bytes_per_second=962.467M/s items_per_second=84.1017M/s

BM_ReadListOfListColumn/0                8938648 ns      8934638 ns           77 bytes_per_second=895.392M/s items_per_second=117.361M/s
BM_ReadListOfListColumn/1               10470577 ns     10466690 ns           66 bytes_per_second=764.329M/s items_per_second=100.182M/s
BM_ReadListOfListColumn/50              18194899 ns     18190700 ns           39 bytes_per_second=439.785M/s items_per_second=57.6435M/s
BM_ReadListOfListColumn/99               7500037 ns      7496263 ns           93 bytes_per_second=1067.2M/s items_per_second=139.88M/s

[github-actions]

https://issues.apache.org/jira/browse/ARROW-10120

[pitrou]

@emkornfield If you can approve this quickly I can then exercise #8320 on it.

[pitrou]

+1, will merge.