[PARQUET-1841] [C++] Experiment to see if using SIMD shuffle operations for DecodeSpaced improves performance

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: parquet-cpp
• Assignee: Micah Kornfield
• Reporter: Micah Kornfield
• Created: 14/Apr/20 04:41
• Updated: 27/Apr/20 03:23

Description

Followup from PARQUET-1840 for current benchmarks it seems that doing removing the memset somehow either has no impact or is slightly worse.  We should investigate using SIMD operations to speed up spacing. 

 

As part of this we can see if moving the memset to only cover uninitialized values after moving all required values provides any speedup.