Issues · NVIDIA/cccl · GitHub

Labels Milestones New issue

[FEA]: Expose compiler detection macros

#8247

· Jacobfaib opened

on Mar 31, 2026

[FEA]: Reenable cluster launches in groups tests

#8246

· davebayer opened

on Mar 31, 2026

Implement C++26 `P4037R1`: Supporting `signed char` and `unsigned char` in random number generation

#8243

· fbusato opened

on Mar 31, 2026

[BUG]: `cub::DeviceScan::ExclusiveScan` and `InclusiveScan` produce incorrect results with custom binary operator

#8242

· shrshi opened

on Mar 31, 2026

Update RAPIDS to no longer use deprecated allocate APIs w/ default alignment

#8241

· jrhemstad opened

on Mar 31, 2026

Add convenience functions for make_buffer to provide a shorthand for usage of the default mempools

#8240

· pciolkosz opened

on Mar 31, 2026

Expose type erased memory resource casts

#8239

· pciolkosz opened

on Mar 31, 2026

[cuda.compute]: Add Benchmark CI for cuda.compute

#8237

· NaderAlAwar opened

on Mar 31, 2026

[FEA]: Optimize `cub::DeviceUniqueByKey` by conditionally loading values

#8234

· NaderAlAwar opened

on Mar 31, 2026

[BUG]: Unqualified `warpReduce` in `warpspeed` scan kernel causes ADL collision

#8233

· viclafargue opened

on Mar 31, 2026

Check for SASS changes to `cub::DeviceTransform` in CI

#8229

· bernhardmgruber opened

on Mar 31, 2026

[FEA]: Implement CUDA backend for parallel `cuda::std::stable_partition`

#8228

· miscco opened

on Mar 31, 2026