You are viewing a plain text version of this content. The canonical link for it is here.

Posted to github@arrow.apache.org by "zhoudan-intel (via GitHub)" <gi...@apache.org> on 2023/02/09 05:17:14 UTC

[GitHub] [arrow] zhoudan-intel commented on a diff in pull request #14674: ARROW-18362: [C++][Parquet] Accelerate Parquet bit-packing decoding with ICX AVX-512 instructions

zhoudan-intel commented on code in PR #14674:
URL: https://github.com/apache/arrow/pull/14674#discussion_r1100991838


##########
cpp/src/arrow/util/bpacking_avx512_icx.h:
##########
@@ -0,0 +1,1145 @@
+// Licensed to the Apache Software Foundation (ASF) under one

Review Comment:
   @pitrou Hey, thanks for your question. This code is written by hand now. I agree that AVX512 intrinsic is not obvious and not easy to understand. But it shows great performance improvement. For this case up to 55% performance improvement is observed on the Intel ICX platform. Regarding the maintainance issue, we can verify its function via unit tests, function tests, etc. If more functions/features are introduced in the future, we will add more test cases to ensure it works as designed. 



-- 
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: github-unsubscribe@arrow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org