9prady9 opened a new pull request #11674:
URL: https://github.com/apache/arrow/pull/11674


   Currently there are two versions of Bitmap operations in the API.
   
   1. Returns a Buffer object
   2. Updates the output pointer passed in as function argument
   
   With this change, the second (2) version of the function can now return 
bitmap operation result's validity count. All bitmap ops are updated to reflect 
the same.
   
   This new output argument passed to the function should be pointer to single 
value (only first value is used if pointer to an array is passed). If this 
pointer is set to `nullptr`, the output validity count isn't calculated.
   
   When output validity count is calculated, there is a performance drop of 
10-20% in number of bytes processed per second as per bit-util-benchmark.
   
   
![chart](https://user-images.githubusercontent.com/3270458/141252701-a406dde4-d0b0-4b6a-b392-d9e915ed34a0.png)
   
   In the aligned bitmapOp case, I wonder if running popcount of word instead 
of looking up each byte in hash is faster - unlikely I would think as a first 
thought.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to