38 Pull requests merged by 12 people

• [STF] Implement CUDASTF_DOT_TIMING for the ctx.cuda_kernel construct

  #2950 merged Nov 25, 2024

• Drop some of the mdspan fold implementation

  #2949 merged Nov 25, 2024

• minor consistency improvements in concepts macros

  #2928 merged Nov 24, 2024

• Try to fix a clang warning:

  #2941 merged Nov 23, 2024

• Add missing qualifier for cuda namespace

  #2940 merged Nov 23, 2024

• Add tuple protocol to cuda::std::complex from C++26

  #2882 merged Nov 22, 2024

• implement C++26 std::span's constructor from std::initializer_list

  #2923 merged Nov 22, 2024

• Reorganize PTX headers to match generator

  #2925 merged Nov 22, 2024

• Improve build instructions for libcu++

  #2881 merged Nov 22, 2024

• Reorganize PTX docs to match generator

  #2929 merged Nov 22, 2024

• Reorganize PTX tests to match generator

  #2930 merged Nov 22, 2024

• move msvc compiler macros to new version

  #2885 merged Nov 22, 2024

• implement C++26 std::span::at

  #2924 merged Nov 22, 2024

• Replace inconsistent Doxygen macros with _CCCL_DOXYGEN_INVOKED

  #2921 merged Nov 21, 2024

• add "interface" to _CCCL_PUSH_MACROS

  #2919 merged Nov 21, 2024

• Refactor nvbench helper less_t

  #2905 merged Nov 21, 2024

• Try to work around issue with NVHPC in conjunction with older CTK versions

  #2889 merged Nov 21, 2024

• Move implementation of _LIBCUDACXX_TEMPLATE to CCCL

  #2832 merged Nov 21, 2024

• Fix old gcc version check

  #2904 merged Nov 20, 2024

• for_each_in_extent

  #2518 merged Nov 20, 2024

• Add thrust_create_target DISPATCH option.

  #2844 merged Nov 19, 2024

• Automate release branch creation

  #2685 merged Nov 19, 2024

• move _CCCL_COMPILER_CLANG to the new macro

  #2859 merged Nov 19, 2024

• backport to_underlying

  #2853 merged Nov 19, 2024

• Include use of NVHPC in CUB/Thrust magic namespace

  #2771 merged Nov 19, 2024

• Fix rst typos in benchmarking.html

  #2868 merged Nov 19, 2024

• [Docs/PTX] Add device tensor map init example

  #1983 merged Nov 19, 2024

• correct the names of shared_resource's async allocate/deallocate members

  #2880 merged Nov 19, 2024

• add missing DOXYGEN_* predefined macros when building the cudax docs

  #2858 merged Nov 19, 2024

• [Backport] Fix cluster launch error in branch/2.7.x

  #2866 merged Nov 18, 2024

• Make discovery mechanism for cuda/_include directory compatible with pip install --editable

  #2846 merged Nov 18, 2024

• Fix DeviceSegmentedSort NVTX range name

  #2857 merged Nov 18, 2024

• Add MatX build to CCCL CI

  #2682 merged Nov 18, 2024

• Fix race condition in block-RLD test harness.

  #2706 merged Nov 18, 2024

• Add benchmarking and tuning presets

  #2856 merged Nov 18, 2024

• Move _CCCL_COMPILER_GCC to the new macro

  #2850 merged Nov 18, 2024

• Add missing include

  #2855 merged Nov 18, 2024

• Fix wrong include in Thrust benchmark

  #2854 merged Nov 18, 2024

16 Pull requests opened by 9 people

• Drop memory resources in libcu++

  #2860 opened Nov 18, 2024

• Disable execution checks for tuple (#2780)

  #2867 opened Nov 18, 2024

• Tweak tuning database plot and comparison scripts

  #2883 opened Nov 19, 2024

• Add more CUB transform benchmarks

  #2906 opened Nov 20, 2024

• implement C++26 `std::ignore`

  #2922 opened Nov 21, 2024

• Add "Open with Codeanywhere" badge to .devcontainer/README.md

  #2926 opened Nov 21, 2024

• Reduce number of per-PR CI jobs.

  #2931 opened Nov 21, 2024

• [CUDAX] Add copy_bytes and fill_bytes overloads for mdspan

  #2932 opened Nov 22, 2024

• unify implementation of `[[deprecated]]` attribute

  #2934 opened Nov 22, 2024

• Regenerate `cuda::ptx` files and run format

  #2937 opened Nov 22, 2024

• cudax compiler bump

  #2943 opened Nov 22, 2024

• fix thread-reduce performance regression

  #2944 opened Nov 22, 2024

• add a `_CCCL_NO_CONCEPTS` config macro

  #2945 opened Nov 23, 2024

• add a `__type_switch` utility and use it the ptx generator

  #2946 opened Nov 24, 2024

• Avoid potential null dereference in `annotated_ptr`

  #2951 opened Nov 25, 2024

• make compiler version comparison utility generic

  #2952 opened Nov 25, 2024

16 Issues closed by 9 people

• Ambiguous cuda:: namespace issue

  #2939 closed Nov 23, 2024

• [BUG]: thrust::tuple<Eigen::Vector3f> copy Constructor fails in template dedcution

  #2936 closed Nov 22, 2024

• [BUG]: inconsistent use of macros to suppress doxygen document generation

  #2362 closed Nov 21, 2024

• Determine and finalize design for large input support in CUB

  #1454 closed Nov 21, 2024

• [EPIC] Self-hosted Windows Runners

  #523 closed Nov 20, 2024

• Get sccache working in Windows

  #312 closed Nov 20, 2024

• Initial version of "Finalize Release" workflow

  #2917 closed Nov 20, 2024

• Initial version of "New Release" workflow

  #2915 closed Nov 20, 2024

• MSVC: Compilation of <tuple> issues in VS 2017 and VS 2019

  #955 closed Nov 20, 2024

• Gather benchmark results for each CUB algorithm using different offset types

  #1787 closed Nov 20, 2024

• [FEA]: Provide `for_each_in_extent` to directly linearization/de-linearization multidimensional indices with `mdspan`

  #2475 closed Nov 20, 2024

• Extend CUB/Thrust magic namespace with NVHPC

  #2770 closed Nov 19, 2024

• [DOC]: PTX: Document how to initialize/modify a tensor map in device memory

  #1982 closed Nov 19, 2024

• [FEA]: Build MatX as part of our third-party CI testing

  #2264 closed Nov 18, 2024

• [BUG]: Race reported in cub::BlockStore

  #1903 closed Nov 18, 2024

• Create CMake presets for benchmarking and tuning

  #2839 closed Nov 18, 2024

50 Issues opened by 15 people

• [BUG]: Suboptimal swap performance on universal vectors

  #2948 opened Nov 25, 2024

• [BUG]: libcu++ should support the three-way comparison operator

  #2947 opened Nov 24, 2024

• [BUG]: UB in annotated_ptr

  #2942 opened Nov 22, 2024

• [FEA]: `cuda::span_collection`

  #2938 opened Nov 22, 2024

• Towards a libcudacxx developer guide

  #2935 opened Nov 22, 2024

• [BUG]: cuda::ptx takes long to compile

  #2933 opened Nov 22, 2024

• [FEA]: Add `__pipeline_arrive_on_noinc`

  #2927 opened Nov 21, 2024

• [FEA] Provide templates for device-side wrappers that can be type-checked on the host

  #2918 opened Nov 20, 2024

• Initial version of "Update RC" workflow

  #2916 opened Nov 20, 2024

• [FEA]: Come up with a plan for consolidating CCCL namespaces

  #2914 opened Nov 20, 2024

• [Theme]: Namespace Unification

  #2913 opened Nov 20, 2024

• add copy/fill_bytes overload for mdspan

  #2912 opened Nov 20, 2024

• [DOC]: Improve thrust API Reference list appearance

  #2911 opened Nov 20, 2024

• [FEA]: Implement `_CCCL_MAYBE_UNUSED`?

  #2910 opened Nov 20, 2024

• [DOC]: Clean-up reference example code snippets in CUB

  #2909 opened Nov 20, 2024

• [DOC]: Improve reference example code snippets in Thrust

  #2908 opened Nov 20, 2024

• [DOC]: Provide a Tutorials top level page in our CCCL docs

  #2903 opened Nov 20, 2024

• [DOC]: Extract from source and create a libcu++ examples standalone subsection under the libcu++ section

  #2902 opened Nov 20, 2024

• [DOC]: Extract from source and create a Thrust examples standalone subsection under the Thrust section

  #2901 opened Nov 20, 2024

• [DOC]: Extract CUB examples into a standalone subsection under the CUB section

  #2900 opened Nov 20, 2024

• [DOC]: Remove Releases/Contributing subsections from Thrust and libcu++

  #2899 opened Nov 20, 2024

• [DOC]: Consolidate and flatten the two Thrust API subsections

  #2898 opened Nov 20, 2024

• [DOC]: Unify Docs hierarchy tree

  #2897 opened Nov 20, 2024

• [DOC]: Flatten CUB API documentation section

  #2896 opened Nov 20, 2024

• Split Device-Wide Primitives into Reference and Explanation pages

  #2895 opened Nov 20, 2024

• Split Block-Wide Primitives into Reference and Explanation pages

  #2894 opened Nov 20, 2024

• Split Warp-Wide Primitives into Reference and Explanation pages

  #2893 opened Nov 20, 2024

• [DOC]: Split CUB primitives pages into Reference and Explanation pages [diataxis]

  #2892 opened Nov 20, 2024

• Explore a stricter separation of synchronous and asynchronous containers and APIs

  #2891 opened Nov 19, 2024

• [BUG]: sm_100 missing from <nv/target>

  #2890 opened Nov 19, 2024

• mdarray multi-dimensional owning abstraction

  #2888 opened Nov 19, 2024

• Investigate interaction between `cuda::vector` and `cuda::launch`

  #2887 opened Nov 19, 2024

• Provide a PoC implementation within cudax

  #2886 opened Nov 19, 2024

• [FEA]: Store peak memory bandwidth in tuning database

  #2884 opened Nov 19, 2024

• [FEA]: Migrate fail test utilities to top-level; reuse between projects

  #2879 opened Nov 19, 2024

• Design for runtime compilation of kernel templates with dynamically supplied types

  #2878 opened Nov 19, 2024

• Adjust Jitify2 interface to the rest of CUDAX / Design NVRTC interface for CUDAX

  #2877 opened Nov 19, 2024

• Meet with Jitify2 maintainers and discuss potential exposure in CUDAX

  #2876 opened Nov 19, 2024

• Expose runtime compilation of kernels

  #2875 opened Nov 19, 2024

• Explore integration of host side execution with container APIs

  #2874 opened Nov 19, 2024

• CUDAX stream host callback API

  #2873 opened Nov 19, 2024

• Stream ordered host execution

  #2872 opened Nov 19, 2024

• get stream priority range

  #2871 opened Nov 19, 2024

• get and set device flags

  #2870 opened Nov 19, 2024

• get/set device limit

  #2869 opened Nov 19, 2024

• [FEA]: Combine c2h and benchmark data generators.

  #2865 opened Nov 18, 2024

• [FEA]: Migrate nvbench utilities to top-level; reuse between projects

  #2864 opened Nov 18, 2024

• [DOC]: Convert to single sphinx project so full ToC is visible from subprojects

  #2863 opened Nov 18, 2024

• [DOC]: Migrate to nvidia's sphinx theme.

  #2862 opened Nov 18, 2024

• [EXTERNAL] Numba pointer arithmetic

  #2861 opened Nov 18, 2024

138 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Start reworking our math functions

  #2749 commented on Nov 25, 2024 • 20 new comments

• backport `unreachable`

  #2852 commented on Nov 22, 2024 • 13 new comments

• [STF] reduce access mode

  #2830 commented on Nov 25, 2024 • 3 new comments

• Add environment to encapsulate information needed for `cudax::vector`

  #2775 commented on Nov 25, 2024 • 1 new comment

• Higher-level, `cuda::vector<T, Properties...>` (like `thrust::host_vector/device_vector`)

  #2057 commented on Nov 20, 2024 • 0 new comments

• Provide Run-Length Decode API

  #599 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Port libcxx <ranges> views and backport to C++17

  #93 commented on Nov 20, 2024 • 0 new comments

• [DOC]: Write dev docs for how to use Windows images for local development

  #94 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Extend devcontainer/launch.sh to build custom containers on the fly

  #126 commented on Nov 20, 2024 • 0 new comments

• Thrust large input support

  #49 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Use P2322R6 to determine accumulator type in Thrust

  #153 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Intrusive Decoupled Look-Back

  #220 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Relax shuffle-based warp exchange requirements

  #271 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: Thrust/CUB shouldn't invoke user-defined operators on out-of-bounds data

  #459 commented on Nov 20, 2024 • 0 new comments

• Specialize relevant `cuda::(std::)` types for `__half/bfloat16/fp8`

  #525 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Transition to new Thrust benchmarks

  #557 commented on Nov 20, 2024 • 0 new comments

• Use parallel implementations for CPU execution policies

  #831 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Remove CDP (RDC) architecture filtering logic from Thrust/CUB tests

  #1137 commented on Nov 20, 2024 • 0 new comments

• Replace mdspan with upstream libc++ implementation

  #1185 commented on Nov 20, 2024 • 0 new comments

• Proof-of-concept parallel ranges algorithms

  #1213 commented on Nov 20, 2024 • 0 new comments

• [DOC]: Add CUB examples utilizing multi-dimensional thread blocks

  #1659 commented on Nov 20, 2024 • 0 new comments

• expose device properties, meminfo etc

  #2081 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Improve performance of `thrust::all_of`.

  #2113 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Synchronous data structures that use a cuda::mr::resource to allocate their memory

  #2129 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Aynchronous data structures that use a cuda::mr::asyncresource to allocate their memory

  #2131 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Non-owning view types with properties to ensure type safe access to allocated memory

  #2132 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Prototype a design to ensure asynchronous operations on different streams work nicely with `cuda::launch`

  #2143 commented on Nov 20, 2024 • 0 new comments

• Publish `cuda-cooperative` Python package

  #2148 commented on Nov 20, 2024 • 0 new comments

• Roll out address stability optimization

  #2403 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Multi-dimesional TMA exposure

  #39 commented on Nov 20, 2024 • 0 new comments

• [EPIC] std::simd support in libcu++

  #30 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Refactor user-facing Thrust types to use libcu++

  #34 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Evaluate and retune radix sort and merge sort

  #55 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Implement 1D TMA support in memcpy_async

  #58 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Explore potential improvements for DeviceMemcpy::Batched

  #59 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Add `thread_scope_cluster` and implement `atomic(_ref)<thread_scope_cluster>`

  #73 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Add support for remote shared memory to barrier<thread_scope_block>

  #75 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Provide `std::dims`

  #2810 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Provide `submdspan` for padded layouts

  #2809 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: Add Hopper features to `cuda::ptx`

  #1340 commented on Nov 21, 2024 • 0 new comments

• Benchmark and plot before/after performance of BabelStream based on `thrust::transform`

  #2717 commented on Nov 21, 2024 • 0 new comments

• [Do Not Merge] Implement `<ranges>`

  #198 commented on Nov 19, 2024 • 0 new comments

• Add compute-sanitizer testing to CI.

  #1879 commented on Nov 20, 2024 • 0 new comments

• Add simple kernel for deterministic reduction

  #2234 commented on Nov 21, 2024 • 0 new comments

• `basic_any`: a utility for defining type-erasing wrappers in terms of an interface description

  #2633 commented on Nov 22, 2024 • 0 new comments

• Add CI with STF MathLib builds

  #2651 commented on Nov 19, 2024 • 0 new comments

• Configure `workflow-run-job-linux` to use sccache-dist build cluster

  #2672 commented on Nov 25, 2024 • 0 new comments

• Add missing template parameter to BlockRadixRank example.

  #2736 commented on Nov 19, 2024 • 0 new comments

• [WIP] Support fancy iterators in cuda.parallel

  #2788 commented on Nov 23, 2024 • 0 new comments

• backport std integer comparison functions to C++11

  #2805 commented on Nov 22, 2024 • 0 new comments

• Try to avoid UB in `thrust::reference`

  #2813 commented on Nov 25, 2024 • 0 new comments

• new type-erased memory resources

  #2824 commented on Nov 22, 2024 • 0 new comments

• Test proclaim_copyable_arguments for lambdas

  #2833 commented on Nov 23, 2024 • 0 new comments

• [BUG]: MSVC < 2022 doesn't properly handle thrust's member function detector.

  #1731 commented on Nov 20, 2024 • 0 new comments

• [EPIC] RAPIDS Should not need to patch CCCL

  #1939 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Extend cub::DeviceSegmentedSort to support custom comparators

  #1577 commented on Nov 20, 2024 • 0 new comments

• Benchmark regression testing

  #2011 commented on Nov 20, 2024 • 0 new comments

• Standard abstraction for specifying thread grid hierarchy and dimensions

  #2037 commented on Nov 20, 2024 • 0 new comments

• Provide cmath for floating_point<M,E>

  #2161 commented on Nov 20, 2024 • 0 new comments

• Make bfloat alias for `floating_point<8,7>` that dispatches to accelerated instructions where possible

  #2182 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Consider compiling CCCL tests with enabled warnings for local memory and register spilling

  #2253 commented on Nov 20, 2024 • 0 new comments

• [DOC]: Clarity check of CUB device-scope docs

  #2323 commented on Nov 20, 2024 • 0 new comments

• Return algorithm diagrams to CUB docs

  #2319 commented on Nov 20, 2024 • 0 new comments

• Make half_t alias for `floating_point<5,10>` that dispatches to accelerated instructions where possible

  #2181 commented on Nov 20, 2024 • 0 new comments

• [EPIC] CUDA Next asynchronous programming model

  #2041 commented on Nov 20, 2024 • 0 new comments

• [DOC]: Clarity check of CUB block-scope docs

  #2324 commented on Nov 20, 2024 • 0 new comments

• [DOC]: Clarity check of CUB warp-scope docs

  #2325 commented on Nov 20, 2024 • 0 new comments

• [THEME] CUDA Runtime Modernization

  #1646 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Abstract build step in cccl.c.parallel

  #2525 commented on Nov 20, 2024 • 0 new comments

• [EXTERNAL] Contribute back Numba documentation updates

  #2722 commented on Nov 18, 2024 • 0 new comments

• [DOC]: `copy_n` docs formatting bug

  #2748 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: Improve documentation content by adopting diataxis

  #2841 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: Fix sphinx

  #2807 commented on Nov 20, 2024 • 0 new comments

• [EPIC] make the Thrust parallel algorithms available in `cuda::std`

  #2818 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Implement cuda.parallel.device_for

  #2537 commented on Nov 20, 2024 • 0 new comments

• [DOC]: Design approach to CCCL documentation authoring

  #2327 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Add more benchmarks for `thrust::transform`

  #2814 commented on Nov 20, 2024 • 0 new comments

• [THEME] Add support for 128b `atomic_ref` in device code.

  #2048 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: `mdspan`-based algorithms

  #2471 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Improve separation and elminate redundancy between Thrust and CUB

  #24 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Consolidate kernels between Thrust and CUB

  #26 commented on Nov 20, 2024 • 0 new comments

• [EPIC] CUB Performance Tuning

  #27 commented on Nov 20, 2024 • 0 new comments

• [EPIC] CUB Test Catch2 Migration

  #28 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Port CUB device-scope algorithm tests to use Catch2

  #29 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Extended Floating-Point Support

  #31 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Replace & refactor Thrust/CUB types w/ libcu++

  #33 commented on Nov 20, 2024 • 0 new comments

• [THEME] Hopper CUDA C++ Feature Exposure

  #35 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: Unify testing infrastructure

  #2806 commented on Nov 19, 2024 • 0 new comments

• [FEA]: Make cuda.parallel available on PyPI

  #2555 commented on Nov 19, 2024 • 0 new comments

• [FEA]: Match cuda.parallel support matrix with CuPy

  #2554 commented on Nov 19, 2024 • 0 new comments

• [FEA]: Implement cuda.parallel version of radix sort

  #2552 commented on Nov 19, 2024 • 0 new comments

• [FEA]: Implement cuda.parallel version of partition

  #2551 commented on Nov 19, 2024 • 0 new comments

• [FEA]: Implement cuda.parallel version of select

  #2550 commented on Nov 19, 2024 • 0 new comments

• [FEA]: Implement cuda.parallel.segmented_reduce algorithm

  #2549 commented on Nov 19, 2024 • 0 new comments

• [FEA]: Implement cuda.parallel.merge_sort algorithm

  #2546 commented on Nov 19, 2024 • 0 new comments

• [FEA]: Implement cuda.parallel version of segmented radix sort

  #2553 commented on Nov 19, 2024 • 0 new comments

• [FEA]: Provide `mdarray`

  #2474 commented on Nov 19, 2024 • 0 new comments

• [DOC]: Format function signatures

  #2339 commented on Nov 19, 2024 • 0 new comments

• [FEA]: Research stateful operators for cuda.parallel

  #2538 commented on Nov 19, 2024 • 0 new comments

• [FEA]: Implement cuda.parallel version of `cub::DoubleBuffer`

  #2548 commented on Nov 19, 2024 • 0 new comments

• [FEA]: Support fancy iterators in cuda.parallel

  #2479 commented on Nov 20, 2024 • 0 new comments

• [BUG]: Proclaiming copyable arguments for lambdas fails to compile

  #2834 commented on Nov 20, 2024 • 0 new comments

• [DOC]: `thrust::partition_copy` docs are wrong

  #2792 commented on Nov 20, 2024 • 0 new comments

• [DOC]: `thrust::stable_sort` docs are wrong

  #2747 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: Improve usability of architecture specific features in libcudacxx

  #1083 commented on Nov 20, 2024 • 0 new comments

• Redesign libcudacxx architecture specific testing

  #1084 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Add SM90 to architecture list for CI

  #1092 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Expand CUB APIs for user-specified algorithmic properties

  #1186 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Clarify support for CCCL headers with host-only translation units

  #1374 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Move rapids repositories away from thrust types in API interfaces

  #1382 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Roadmap for cuda/memory_resource

  #1502 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Third-party testing in CI

  #1507 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: Reproducible floating-point reductions

  #1558 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Run tests through `compute-sanitizer` in CI

  #1618 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: Setup nightly/weekly CI

  #1619 commented on Nov 20, 2024 • 0 new comments

• Provide generic implementation of `floating_point<M, E>` limited to arithmetic operations

  #1666 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Optimize `thrust::transform` for newer architectures

  #1947 commented on Nov 20, 2024 • 0 new comments

• Create generic floating-point wrapping type for <NumExponentBits, NumMantissaBits>

  #1665 commented on Nov 20, 2024 • 0 new comments

• Modern versions of cudaMemset and cudaMemcpy

  #2000 commented on Nov 20, 2024 • 0 new comments

• `cuda::launch` kernel-launch API

  #2038 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: Atomics Improvements

  #2047 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: TMA Exposure

  #36 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: Hopper Cluster support

  #37 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: 1D TMA/BULK exposure via `memcpy_async`

  #38 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Fork/Join Parallel Ranges

  #44 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Heterogeneous, sequential ranges support

  #45 commented on Nov 20, 2024 • 0 new comments

• [THEME] Asynchronous Parallel Algorithms

  #46 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Universal 64-bit index type support in Thrust/CUB algorithms

  #47 commented on Nov 20, 2024 • 0 new comments

• [FEA]: CUB large input support

  #50 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Docs Overhaul

  #51 commented on Nov 20, 2024 • 0 new comments

• [EPIC]: Implement `<cuda/std/ranges>` and associated headers

  #61 commented on Nov 20, 2024 • 0 new comments

• [EPIC] GitHub Actions CI

  #68 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Track future breaking changes

  #101 commented on Nov 20, 2024 • 0 new comments

• [EPIC] Setup Windows container and script infrastructure

  #248 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Vectorize memory operations in CUB algorithms

  #307 commented on Nov 20, 2024 • 0 new comments

• [FEA]: Automate release process

  #653 commented on Nov 20, 2024 • 0 new comments

• `thrust::all_of` is slower than a naive reduction

  #720 commented on Nov 20, 2024 • 0 new comments

• Add additional protections against undefined uses of CUDA extended lambdas

  #1004 commented on Nov 20, 2024 • 0 new comments