Repository navigation

All pull requests

[FEA] Add fixed-width list support to JIT transforms
#24767lamarrr opened Oct 22, 2026Contributor3 tasks
1
[FEA] Unify JIT storage access and mutable string representations
#24766lamarrr opened Oct 22, 2026Contributor3 tasks
1
Use std::span in Parquet readers internals
#24765mhaseeb123 opened Oct 22, 2026Contributor3 tasks done
1
Reimplement Categorical support in cudf-polars using dictionary32 types
#24764mroeschke opened Oct 22, 2026Contributor3 tasks
2
Fix instances of __host__ code calling __device__ code
#24755KyleFromNVIDIA opened Oct 21, 2026Member3 tasks
6
Use Parquet NaN counts in stats filtering
#24754mhaseeb123 opened Oct 21, 2026Contributor3 of 4 tasks
1
Support Parquet IEEE 754 nan_count statistics fields
#24753mhaseeb123 opened Oct 21, 2026Contributor3 tasks done
3
Migrate allocate_like to cudf::memory_resources
#24751PointKernel opened Oct 21, 2026Member3 tasks done
2
Migrate tile to cudf::memory_resources
#24750PointKernel opened Oct 21, 2026Member3 tasks done
2
Migrate repeat APIs to cudf::memory_resources
#24749PointKernel opened Oct 21, 2026Member3 tasks done
2
Migrate apply_mask APIs to cudf::memory_resources
#24748PointKernel opened Oct 21, 2026Member3 tasks done
2
Bound HashCSR groupby build kernel registers on Ampere and Hopper
#24747PointKernel opened Oct 21, 2026Member3 tasks done
4
Allow PDS-H benchmarks to accept SQL queries
#24746mroeschke opened Oct 21, 2026Contributor3 tasks
4
Ignore NaN Parquet stats values in stats filtering
#24745mhaseeb123 opened Oct 21, 2026Contributor3 tasks done
8
Allow specialized host buffers to handle device copies
#24743zpuller opened Oct 21, 2026Contributor3 tasks done
1
Update polars xfail/skip list since polars 2.0 upgrade
#24742Matt711 opened Oct 21, 2026Member3 tasks
2
Avoid duplicate sort keys in cudf-polars
#24739rjzamora opened Oct 21, 2026Contributor
2
Fix disengaged optional dereference in null-aware transform_kernel
#24738davidwendt opened Oct 21, 2026Contributor3 tasks done
7
Persist ORC calendar metadata and support legacy DATE rebasing
#24733wjxiz1992 opened Oct 21, 2026Contributor3 tasks done
3
Preserve empty dictionary32 key types in gather, scatter, copy, column constructors
#24732mroeschke opened Oct 21, 2026Contributor3 tasks
1
Track peak memory in join benchmarks
#24731PointKernel opened Oct 20, 2026Member3 tasks done
3
Fix access_columns cleanup on enter failure
#24729fallintoplace opened Oct 20, 2026Contributor
4
Use logarithmic radix binning to schedule strings::contains
#24728davidwendt opened Oct 20, 2026Contributor3 tasks done
2
[FEA] Introduce libcudacxxlite for faster JIT compilation
#24726lamarrr opened Oct 20, 2026Contributor3 tasks done
2
Convert IO text's scratch rmm::device_uvector<T> to cuda::device_buffer<T>
#24724KyleFromNVIDIA opened Oct 20, 2026Member3 tasks
5

Navigation Menu

Repository navigation

All pull requests

[FEA] Add fixed-width list support to JIT transforms

[FEA] Unify JIT storage access and mutable string representations

Use std::span in Parquet readers internals

Reimplement Categorical support in cudf-polars using dictionary32 types

Fix instances of `host` code calling `device` code

Use Parquet NaN counts in stats filtering

Support Parquet IEEE 754 `nan_count` statistics fields

Migrate allocate_like to cudf::memory_resources

Migrate tile to cudf::memory_resources

Migrate repeat APIs to cudf::memory_resources

Migrate apply_mask APIs to cudf::memory_resources

Bound HashCSR groupby build kernel registers on Ampere and Hopper

Allow PDS-H benchmarks to accept SQL queries

Ignore `NaN` Parquet stats values in stats filtering

Allow specialized host buffers to handle device copies

Update polars xfail/skip list since polars 2.0 upgrade

Avoid duplicate sort keys in cudf-polars

Fix disengaged optional dereference in null-aware transform_kernel

Persist ORC calendar metadata and support legacy DATE rebasing

Preserve empty dictionary32 key types in gather, scatter, copy, column constructors

Track peak memory in join benchmarks

Fix access_columns cleanup on enter failure

Use logarithmic radix binning to schedule strings::contains

[FEA] Introduce libcudacxxlite for faster JIT compilation

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

Pull requests