Jihoon Son

Add throughput metrics for REDUCTION_BENCH/REDUCTION_NVBENCH benchmarks

7

## Description This PR addresses https://github.com/rapidsai/cudf/issues/13735 for reduction benchmarks. There are 3 new utils added. - `int64_t estimate_size(cudf::table_view)` returns a size estimate for the given table. https://github.com/rapidsai/cudf/pull/13984 was a previous...

[FEA] Dictionary type support for `row_bit_count`

3

**Is your feature request related to a problem? Please describe.** `row_bit_count` currently does not support the dictionary type. **Describe the solution you'd like** It seems that we could add a...

Fix empty cluster handling in tdigest merge

1

## Description This PR fixes an edge case bug in the tdigest merge. When there are multiple distinct keys but all values are empty clusters, the value column is currently...

Add a new NVTX range for task GPU ownership

1

If you have ever been curious about the semaphore-based GPU concurrency control mechanism, you may have wondered at some point exactly what tasks are holding the semaphore at a given...

https://github.com/NVIDIA/spark-rapids/pull/11596 adds a new NVTX range to track which task owns the GPU during query processing. Currently, the colors of the task range and the GPU ownership are fixed to...

[BUG] `cudaErrorInvalidDevice` in ParquetChunkedReader with a small number of executor cores

11

**Describe the bug** A simple join query fails with the error below. ``` 24/10/07 17:06:59 WARN TaskSetManager: Lost task 0.0 in stage 31.0 (TID 12031) (10.110.47.50 executor 4): java.io.IOException: Error...

[FEA] Support metadata row index for Delta tables with DVs

3

**Is your feature request related to a problem? Please describe.** In Delta tables with DVs, `spark.databricks.delta.deletionVectors.useMetadataRowIndex` is the configuration to enable optimizations such as file split and predicate pushdown. Since...

**Is your feature request related to a problem? Please describe.** On Databricks, AQE can optimize the shuffle at query time using the auto-optimized shuffle feature. This feature can be enabled...

[BUG] CudfColumnSizeOverflowException in populating a column of input file name

6

**Describe the bug** `GpuInputFileName.columnarEval` populates a string column with a scalar value which is the input file path. Since the file path length can vary, this can throw the `CudfColumnSizeOverflowException`...

**Is your feature request related to a problem? Please describe.** Similar to #8415, we should support the merge with the "when not matched by source" clause, so that it won't...