iceberg-kitten opened a new pull request, #1639:
URL: https://github.com/apache/arrow-go/pull/1639

   ### Why
   The RLE decoder expands literal bits into an intermediate uint64 buffer, 
then converts each word to a boolean.
   
   ### What
   Decode bit-packed literals directly into the output bool slice.
   
   ### Implementation
   Use `BitReader.GetBatchBools` for the literal path. Preserve partial-byte 
state between calls and use the bulk byte-to-bool decoder once aligned.
   
   ### Benchmark
   
   Decode one Parquet RLE bit-packed literal run of 65,536 booleans per 
operation, using the batch sizes below.
   
   | Batch size | Time/op before → after | Change |
   | --- | ---: | ---: |
   | 32 | 872 → 108 µs | -87.7% |
   | 256 | 259 → 18 µs | -93.0% |
   | 4,096 | 186 → 6.02 µs | -96.8% |
   | 65,536 | 177 → 5.22 µs | -97.0% |
   
   Medians of 8 runs per revision on Linux amd64, Intel Core i9-12900HK, Go 
1.26.11, pinned to one efficiency core. Baseline is `main` at `aba35416`. Use 
the same benchmark on both revisions. Copy the PR's benchmark file into the 
baseline checkout before running. All cases reported 0 allocs/op.
   
   ```sh
   go test ./parquet/internal/utils -run '^$' -bench 
'^BenchmarkRLEGetBatchBools$' -benchmem -benchtime=120ms -count=8 -cpu=1
   ```
   


-- 
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