Remove hacking code

cpp/src/arrow/acero/swiss_join_avx2.cc

@@ -397,20 +397,6 @@ int RowArray::DecodeFixedLength_avx2(ResizableArrayData* output, int output_star
  int column_id, uint32_t fixed_length,
  int num_rows_to_append,
  const uint32_t* row_ids) const {
- // Benchmarking shows that when the data element width is <= 8, the scalar code almost
- // always outperforms the vectorized code - about 2X~3X faster when the whole data set
- // falls into L1~LLC, and the ratio goes down to about 1:1 as the data size increases
- // when most of the accesses hit the main memory. This is possibly because that decoding
- // is mostly copying scattered pieces of data and there is not enough computation to pay
- // off the cost of the heavy gather instructions.
- // For fixed length 0 (boolean column), the vectorized code wins by batching 8 bits into
- // a single byte instead of modifying the same byte 8 times in the scalar code.
- if (!(fixed_length == 0 || fixed_length > 8)) {
- DecodeFixedLength(output, output_start_row, column_id, fixed_length,
- num_rows_to_append, row_ids);
- return num_rows_to_append;
- }
-
  DCHECK_EQ(output_start_row % 8, 0);
 
  int num_rows_processed = 0;
@@ -483,10 +469,6 @@ int RowArray::DecodeFixedLength_avx2(ResizableArrayData* output, int output_star
 int RowArray::DecodeOffsets_avx2(ResizableArrayData* output, int output_start_row,
  int column_id, int num_rows_to_append,
  const uint32_t* row_ids) const {
- // Same reason as DecodeFixedLength_avx2.
- DecodeOffsets(output, output_start_row, column_id, num_rows_to_append, row_ids);
- return num_rows_to_append;
-
  uint32_t* offsets =
  reinterpret_cast<uint32_t*>(output->mutable_data(1)) + output_start_row;
  uint32_t current_length = (output_start_row == 0) ? 0 : offsets[0];