_Parallel_choose_min_chunk - typo? #2345
Answered
by
CaseyCarter
mikekazakov
asked this question in
General
-
Hello, Line 1479 in 3c2fd04 _Static_partitioned_find_end_backward2 should look for the last results, therefore use _Parallel_choose_max_result /_Parallel_choose_max_chunk like _Static_partitioned_find_end_forward does, but for some reason _Parallel_choose_max_result /_Parallel_choose_min_chunk is used in _Static_partitioned_find_end_backward2 .Sorry if I'm missing something. |
Beta Was this translation helpful? Give feedback.
Answered by
CaseyCarter
Nov 19, 2021
Replies: 1 comment 1 reply
-
The difference is that |
Beta Was this translation helpful? Give feedback.
1 reply
Answer selected by
mikekazakov
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The difference is that
find_end_backward
uses_Static_partition_range_backward
, which partitions the range from the end to the beginning - backward - instead of from the beginning to the end. That way chunks with smaller ids are near the end of the range and will be processed first, which enables us to potentially cancel processing of chunks with bigger ids / toward the beginning of the range. (No need to apologize, this is fairly subtle.)