Repository navigation

All pull requests

Add single-file sparse payload hybrid scan APIs
#24659mhaseeb123 opened Oct 16, 2026Contributor3 of 4 tasks
1
Split large JOIN and Parquet tests for finer-grained GPU scheduling
#24658vyasr opened Oct 16, 2026Contributor3 tasks done
2
Remove unneeded reinterpret_cast<cudf::bitmask_type*>()
#24656KyleFromNVIDIA opened Oct 16, 2026Member3 tasks
2
Remove fatbin linker script from libcudf builds
#24654robertmaynard opened Oct 16, 2026Contributor3 tasks done
1
Support streaming first and last groupby aggregations
#24653rjzamora opened Oct 16, 2026Contributor
4
Backport #24231, #24331, #24396, #24655: Reduce HashCSR hash join memory usage
#24652bdice opened Oct 16, 2026Contributor3 tasks done
5
Convert nvtext scratch rmm::device_uvector<T> to cuda::device_buffer<T>
#24651davidwendt opened Oct 16, 2026Contributor3 tasks done
1
Allow for reseting the Quent context in StreamingEngine._reset
#24649TomAugspurger opened Oct 16, 2026Contributor3 tasks done
8
wheels: use wildcards in auditwheel exclusions
#24648jameslamb opened Oct 16, 2026Member3 tasks done
2
Update docs to match actual project requirements
#24647robertmaynard opened Oct 16, 2026Contributor2 tasks done
1
Test cudf against cupy nightly in nightly test job
#24644TomAugspurger opened Oct 16, 2026Contributor3 tasks done
14
Add and migrate to new hybrid scan APIs
#24641mhaseeb123 opened Oct 16, 2026Contributor3 of 4 tasks
1
Add the DELTA_LENGTH_BYTE_ARRAY nz map consumer and a block-wide delta decode
#24639vyasr opened Oct 15, 2026Contributor3 tasks done
1
Convert quantiles' and reductions' scratch RMM vectors to CUDA buffers
#24638KyleFromNVIDIA opened Oct 15, 2026Member3 tasks
1
Improve matches reduce for hash joins
#24637scanhex12 opened Oct 15, 2026Contributor3 tasks
2
Make all scalar ctors explicit
#24636wence- opened Oct 15, 2026Contributor3 tasks done
8
Convert parquet's scratch rmm::device_uvector<T> to cuda::device_buffer<T>
#24635KyleFromNVIDIA opened Oct 15, 2026Member3 tasks
9
Convert json's scratch rmm::device_uvector<T> to cuda::device_buffer<T>
#24627KyleFromNVIDIA opened Oct 15, 2026Member3 tasks
3
Convert dictionary's scratch rmm::device_uvector<T> to cuda::device_buffer<T>
#24626KyleFromNVIDIA opened Oct 15, 2026Member3 tasks
3
Cache installed files for Conda C++ packaging
#24625vyasr opened Oct 15, 2026Contributor3 tasks done
5
Consolidate hybrid scan reader impl APIs
#24624mhaseeb123 opened Oct 15, 2026Contributor3 tasks done
3
Replace streaming groupby internals with a persistent slot table and warp aggregation
#24623PointKernel opened Oct 15, 2026Member3 tasks done
27
Fix nesting info for one-level Parquet list columns
#24622mhaseeb123 opened Oct 15, 2026Contributor3 tasks done
6
Use generalized maven docs deployment
#24621msarahan opened Oct 15, 2026Member
3
Fix Parquet reads of BYTE_STREAM_SPLIT 8- and 16-bit integer columns
#24618mattgara opened Oct 14, 20263 tasks done
1

Navigation Menu

Repository navigation

All pull requests

Add single-file sparse payload hybrid scan APIs

Split large JOIN and Parquet tests for finer-grained GPU scheduling

Remove unneeded `reinterpret_cast<cudf::bitmask_type*>()`

Remove fatbin linker script from libcudf builds

Support streaming `first` and `last` groupby aggregations

Backport #24231, #24331, #24396, #24655: Reduce HashCSR hash join memory usage

Convert nvtext scratch `rmm::device_uvector<T>` to `cuda::device_buffer<T>`

Allow for reseting the Quent context in `StreamingEngine._reset`

wheels: use wildcards in auditwheel exclusions

Update docs to match actual project requirements

Test cudf against cupy nightly in nightly test job

Add and migrate to new hybrid scan APIs

Add the DELTA_LENGTH_BYTE_ARRAY nz map consumer and a block-wide delta decode

Convert quantiles' and reductions' scratch RMM vectors to CUDA buffers

Improve matches reduce for hash joins

Make all scalar ctors explicit

Convert parquet's scratch `rmm::device_uvector<T>` to `cuda::device_buffer<T>`

Convert json's scratch `rmm::device_uvector<T>` to `cuda::device_buffer<T>`

Convert dictionary's scratch `rmm::device_uvector<T>` to `cuda::device_buffer<T>`

Cache installed files for Conda C++ packaging

Consolidate hybrid scan reader impl APIs

Replace streaming groupby internals with a persistent slot table and warp aggregation

Fix nesting info for one-level Parquet list columns

Use generalized maven docs deployment

Fix Parquet reads of BYTE_STREAM_SPLIT 8- and 16-bit integer columns

Pull requests