Modest improvement to FixedLenByteArray BYTE_STREAM_SPLIT arrow decoder

[etseidl]

Which issue does this PR close?

Related to #6219.

Rationale for this change

When this decoder was added, little thought was given to performance. The changes here have an impact that varies by architecture.

Old 2.2GHz Core i7

group                                                                                                   master                                 opt_read
-----                                                                                                   ------                                 --------
arrow_array_reader/BYTE_ARRAY/Decimal128Array/plain encoded, mandatory, no NULLs                        1.00      2.2±0.04ms        ? ?/sec    1.03      2.3±0.48ms        ? ?/sec
arrow_array_reader/BYTE_ARRAY/Decimal128Array/plain encoded, optional, half NULLs                       1.00      2.3±0.04ms        ? ?/sec    1.01      2.3±0.04ms        ? ?/sec
arrow_array_reader/BYTE_ARRAY/Decimal128Array/plain encoded, optional, no NULLs                         1.00      2.2±0.09ms        ? ?/sec    1.00      2.2±0.04ms        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Decimal128Array/byte_stream_split encoded, mandatory, no NULLs     1.74      3.3±0.07ms        ? ?/sec    1.00  1904.9±45.03µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Decimal128Array/byte_stream_split encoded, optional, half NULLs    1.29      3.1±0.05ms        ? ?/sec    1.00      2.4±0.05ms        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Decimal128Array/byte_stream_split encoded, optional, no NULLs      1.72      3.3±0.06ms        ? ?/sec    1.00  1923.3±37.33µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Float16Array/byte_stream_split encoded, mandatory, no NULLs        1.28   831.4±14.84µs        ? ?/sec    1.00   651.8±13.35µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Float16Array/byte_stream_split encoded, optional, half NULLs       1.12  1473.3±33.07µs        ? ?/sec    1.00  1314.2±26.78µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Float16Array/byte_stream_split encoded, optional, no NULLs         1.27   841.0±17.96µs        ? ?/sec    1.00   659.7±13.25µs        ? ?/sec
arrow_array_reader/FIXED_LENGTH_BYTE_ARRAY/Decimal128Array/plain encoded, mandatory, no NULLs           1.00  1027.5±24.48µs        ? ?/sec    1.00  1028.8±24.90µs        ? ?/sec
arrow_array_reader/FIXED_LENGTH_BYTE_ARRAY/Decimal128Array/plain encoded, optional, half NULLs          1.00  1927.8±45.32µs        ? ?/sec    1.00  1927.2±38.97µs        ? ?/sec
arrow_array_reader/FIXED_LENGTH_BYTE_ARRAY/Decimal128Array/plain encoded, optional, no NULLs            1.00  1034.4±27.58µs        ? ?/sec    1.00  1034.2±32.82µs        ? ?/sec

newer 3.6GHz Core i7-12700K

group                                                                                                   master                                 opt_read
-----                                                                                                   ------                                 --------
arrow_array_reader/BYTE_ARRAY/Decimal128Array/plain encoded, mandatory, no NULLs                        1.00    873.0±8.77µs        ? ?/sec    1.00    875.9±5.62µs        ? ?/sec
arrow_array_reader/BYTE_ARRAY/Decimal128Array/plain encoded, optional, half NULLs                       1.00   1039.7±7.19µs        ? ?/sec    1.01   1048.3±6.85µs        ? ?/sec
arrow_array_reader/BYTE_ARRAY/Decimal128Array/plain encoded, optional, no NULLs                         1.00    874.6±4.54µs        ? ?/sec    1.00    878.5±9.08µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Decimal128Array/byte_stream_split encoded, mandatory, no NULLs     1.30   963.5±10.85µs        ? ?/sec    1.00    743.8±3.88µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Decimal128Array/byte_stream_split encoded, optional, half NULLs    1.10  1216.2±16.67µs        ? ?/sec    1.00   1103.1±6.30µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Decimal128Array/byte_stream_split encoded, optional, no NULLs      1.29    966.9±7.71µs        ? ?/sec    1.00    747.5±3.44µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Float16Array/byte_stream_split encoded, mandatory, no NULLs        1.36    347.3±2.89µs        ? ?/sec    1.00    255.5±2.11µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Float16Array/byte_stream_split encoded, optional, half NULLs       1.04    684.2±2.43µs        ? ?/sec    1.00    655.9±4.87µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Float16Array/byte_stream_split encoded, optional, no NULLs         1.37    347.6±5.69µs        ? ?/sec    1.00    254.5±1.07µs        ? ?/sec
arrow_array_reader/FIXED_LENGTH_BYTE_ARRAY/Decimal128Array/plain encoded, mandatory, no NULLs           1.01    367.8±5.65µs        ? ?/sec    1.00    365.7±2.58µs        ? ?/sec
arrow_array_reader/FIXED_LENGTH_BYTE_ARRAY/Decimal128Array/plain encoded, optional, half NULLs          1.01    927.8±8.27µs        ? ?/sec    1.00    919.6±4.94µs        ? ?/sec
arrow_array_reader/FIXED_LENGTH_BYTE_ARRAY/Decimal128Array/plain encoded, optional, no NULLs            1.00    373.1±3.23µs        ? ?/sec    1.00    373.0±5.44µs        ? ?/sec

What changes are included in this PR?

Replaces the use of Vec::push with direct access via slices.

Are there any user-facing changes?

No

[etseidl]

Updated numbers.

2.2GHz Core i7

arrow_array_reader/BYTE_STREAM_SPLIT/Decimal128Array/byte_stream_split encoded, mandatory, no NULLs     1.99      3.3±0.07ms        ? ?/sec    1.00  1669.6±32.91µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Decimal128Array/byte_stream_split encoded, optional, half NULLs    1.37      3.1±0.05ms        ? ?/sec    1.00      2.2±0.02ms        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Decimal128Array/byte_stream_split encoded, optional, no NULLs      1.97      3.3±0.06ms        ? ?/sec    1.00  1679.6±34.58µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Float16Array/byte_stream_split encoded, mandatory, no NULLs        1.34   831.4±14.84µs        ? ?/sec    1.00   619.9±11.99µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Float16Array/byte_stream_split encoded, optional, half NULLs       1.09  1473.3±33.07µs        ? ?/sec    1.00  1356.3±33.21µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Float16Array/byte_stream_split encoded, optional, no NULLs         1.33   841.0±17.96µs        ? ?/sec    1.00   631.5±16.27µs        ? ?/sec

3.6GHz

arrow_array_reader/BYTE_STREAM_SPLIT/Decimal128Array/byte_stream_split encoded, mandatory, no NULLs     1.54   963.5±10.85µs        ? ?/sec    1.00    626.7±4.82µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Decimal128Array/byte_stream_split encoded, optional, half NULLs    1.16  1216.2±16.67µs        ? ?/sec    1.00   1049.0±6.41µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Decimal128Array/byte_stream_split encoded, optional, no NULLs      1.53    966.9±7.71µs        ? ?/sec    1.00    630.4±3.76µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Float16Array/byte_stream_split encoded, mandatory, no NULLs        1.45    347.3±2.89µs        ? ?/sec    1.00    239.8±2.33µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Float16Array/byte_stream_split encoded, optional, half NULLs       1.03    684.2±2.43µs        ? ?/sec    1.00    662.6±3.21µs        ? ?/sec
arrow_array_reader/BYTE_STREAM_SPLIT/Float16Array/byte_stream_split encoded, optional, no NULLs         1.47    347.6±5.69µs        ? ?/sec    1.00    236.0±1.40µs        ? ?/sec

[wiedld]

Change to avoid bounds checking.
👍🏼 benchmarked results.

[alamb]

Thanks @etseidl and @wiedld for the review