Sets the cap on how many threads parquet_table%prefetch/%materialize_all may use to read
several columns at once, each on its own reader.
Read per prefetch, so it takes effect immediately. 0 restores automatic behaviour. The value
is a cap: the table never uses more threads than OpenMP offers, so setting this higher
than OMP_NUM_THREADS changes nothing. 1 makes the prefetch serial, which is also what
happens automatically whenever the parallel path is not applicable.
| Type | Intent | Optional | Attributes | Name | ||
|---|---|---|---|---|---|---|
| integer, | intent(in) | :: | n |
thread cap, or 0 for automatic; must be >= 0. |
subroutine parquet_set_prefetch_threads(n) integer, intent(in) :: n !! thread cap, or 0 for automatic; must be >= 0. if (n < 0) error stop "parquet_set_prefetch_threads: n must be >= 0 (0 means automatic)" cfg_prefetch_threads = n end subroutine parquet_set_prefetch_threads