All issues

[BUG]: `DeviceReduce::ReduceByKey` is not run-to-run deterministic for fp64 sums over multi-tile runs

#9995

· guillaume-michel opened

on Jul 17, 2026

[FEA]: Use mbarriers to improve barrier/communication latency in cluster segmented TopK

#9994

· pauleonix opened

on Jul 17, 2026

[BUG]: Invalid instruction generated by `cuda::ptx::shr` for non-stdint type

#9992

· davebayer opened

on Jul 17, 2026

[cuco][cudax] Replace cudaMemcpyAsync with cudaMemcpyBatchAsync

#9989

· sleeepyjack opened

on Jul 16, 2026

[cuco][cudax] Dedup empty value sentinel

#9988

· sleeepyjack opened

on Jul 16, 2026

[cuco][cudax] Extend type range for key and value types in cuco

#9987

· sleeepyjack opened

on Jul 16, 2026

[CUCO][FEA]: Use plain CUDA atomics instead of cuda::atomic_ref

#9983

· PointKernel opened

on Jul 16, 2026

[cuco][cudax] Keep cudax cuco up to date with recent cuco changes

#9981

· sleeepyjack opened

on Jul 16, 2026

[FEA]: Enable `__float128` in `cuda::complex`

#9976

· davebayer opened

on Jul 16, 2026

[FEA]: Enable `__float128` in `cuda::std::complex`

#9975

· davebayer opened

on Jul 16, 2026

[FEA]: Enable `__float128` in `cuda::sincos`

#9974

· davebayer opened

on Jul 16, 2026

[FEA]: Enable `__float128` in `cuda::std::isunordered`

#9973

· davebayer opened

on Jul 16, 2026