viirya opened a new pull request, #56159: URL: https://github.com/apache/spark/pull/56159
### What changes were proposed in this pull request? Adds a `putNotNulls` case to `WritableColumnVectorBulkFillBenchmark` (introduced in SPARK-57042 / #56084). The case mirrors the existing `putNulls` case and exercises both `OnHeapColumnVector.putNotNulls` and `OffHeapColumnVector.putNotNulls` across the standard count sweep (1, 8, 64, 512, 4096, 65536). `putNotNulls` early-outs unless `numNulls > 0`, so the OnHeap / OffHeap factories seed one null into the column vector at setup before the measured loop begins. ### Why are the changes needed? `putNotNulls` is the inverse of `putNulls` and runs once per batch from `WritableColumnVector.reset()` when `numNulls > 0`. SPARK-57111 (#56156) is switching its implementation from a per-byte loop to `Arrays.fill` / `Platform.setMemory` (with the `SET_MEMORY_THRESHOLD` fallback). Landing the benchmark case in master first means SPARK-57111 will produce a like-for-like `Run benchmarks` GHA baseline → patched comparison, the same way SPARK-57042 / #56084 enabled the comparison for SPARK-57036 / #56082. ### Does this PR introduce _any_ user-facing change? No. Benchmark-only. ### How was this patch tested? The benchmark compiles and runs: ``` build/sbt "sql/Test/runMain org.apache.spark.sql.execution.vectorized.WritableColumnVectorBulkFillBenchmark" ``` ### Was this patch authored or co-authored using generative AI tooling? Generated-by: Claude Code (Claude Opus 4.7) -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
