Skip to content

Latest commit

 

History

History
1655 lines (1253 loc) · 64.1 KB

CHANGELOG.md

File metadata and controls

1655 lines (1253 loc) · 64.1 KB
Raw

CUB 2.1.0

Breaking Changes

  • NVIDIA#553: Deprecate the CUB_USE_COOPERATIVE_GROUPS macro, as all supported CTK distributions provide CG. This macro will be removed in a future version of CUB.

New Features

  • NVIDIA#359: Add new DeviceBatchMemcpy algorithm.
  • NVIDIA#565: Add DeviceMergeSort::StableSortKeysCopy API. Thanks to David Wendt (@davidwendt) for this contribution.
  • NVIDIA#585: Add SM90 tuning policy for DeviceRadixSort. Thanks to Andy Adinets (@canonizer) for this contribution.
  • NVIDIA#586: Introduce a new mechanism to opt-out of compiling CDP support in CUB algorithms by defining CUB_DISABLE_CDP.
  • NVIDIA#589: Support 64-bit indexing in DeviceReduce.
  • NVIDIA#607: Support 128-bit integers in radix sort.

Bug Fixes

  • NVIDIA#547: Resolve several long-running issues resulting from using multiple versions of CUB within the same process. Adds an inline namespace that encodes CUB version and targeted PTX architectures.
  • NVIDIA#562: Fix bug in BlockShuffle resulting from an invalid thread offset. Thanks to @sjfeng1999 for this contribution.
  • NVIDIA#564: Fix bug in BlockRadixRank when used with blocks that are not a multiple of 32 threads.
  • NVIDIA#579: Ensure that all threads in the logical warp participate in the index-shuffle for BlockRadixRank. Thanks to Andy Adinets (@canonizer) for this contribution.
  • NVIDIA#582: Fix reordering in CUB member initializer lists.
  • NVIDIA#589: Fix DeviceSegmentedSort when used with bool keys.
  • NVIDIA#590: Fix CUB's CMake install rules. Thanks to Robert Maynard (@robertmaynard) for this contribution.
  • NVIDIA#592: Fix overflow in DeviceReduce.
  • NVIDIA#598: Fix DeviceRunLengthEncode when the first item is a NaN.
  • NVIDIA#611: Fix WarpScanExclusive for vector types.

Other Enhancements

  • NVIDIA#537: Add detailed and expanded version of a CUB developer overview.
  • NVIDIA#549: Fix BlockReduceRaking docs for non-commutative operations. Thanks to Tobias Ribizel (@upsj) for this contribution.
  • NVIDIA#606: Optimize CUB's decoupled-lookback implementation.

CUB 2.0.1

Other Enhancements

  • Skip device-side synchronization on SM90+. These syncs are a debugging-only feature and not required for correctness, and a warning will be emitted if this happens.

CUB 2.0.0

Summary

The CUB 2.0.0 major release adds a dependency on libcu++ and contains several breaking changes. These include new diagnostics when inspecting device-only lambdas from the host, an updated method of determining accumulator types for algorithms like Reduce and Scan, and a compile-time replacement for the runtime debug_synchronous debugging flags.

This release also includes several new features. DeviceHistogram now supports __half and better handles various edge cases. WarpReduce now performs correctly when restricted to a single-thread “warp”, and will use the __reduce_add_sync accelerated intrinsic (introduced with Ampere) when appropriate. DeviceRadixSort learned to handle the case where begin_bit == end_bit.

Several algorithms also have updated documentation, with a particular focus on clarifying which operations can and cannot be performed in-place.

Breaking Changes

  • NVIDIA#448 Add libcu++ dependency (v1.8.0+).
  • NVIDIA#448: The following macros are no longer defined by default. They can be re-enabled by defining CUB_PROVIDE_LEGACY_ARCH_MACROS. These will be completely removed in a future release.
    • CUB_IS_HOST_CODE: Replace with NV_IF_TARGET.
    • CUB_IS_DEVICE_CODE: Replace with NV_IF_TARGET.
    • CUB_INCLUDE_HOST_CODE: Replace with NV_IF_TARGET.
    • CUB_INCLUDE_DEVICE_CODE: Replace with NV_IF_TARGET.
  • NVIDIA#486: CUB's CUDA Runtime support macros have been updated to support NV_IF_TARGET. They are now defined consistently across all host/device compilation passes. This should not affect most usages of these macros, but may require changes for some edge cases.
    • CUB_RUNTIME_FUNCTION: Execution space annotations for functions that invoke CUDA Runtime APIs.
      • Old behavior:
        • RDC enabled: Defined to __host__ __device__
        • RDC not enabled:
          • NVCC host pass: Defined to __host__ __device__
          • NVCC device pass: Defined to __host__
      • New behavior:
        • RDC enabled: Defined to __host__ __device__
        • RDC not enabled: Defined to __host__
    • CUB_RUNTIME_ENABLED: No change in behavior, but no longer used in CUB. Provided for legacy support only. Legacy behavior:
      • RDC enabled: Macro is defined.
      • RDC not enabled:
        • NVCC host pass: Macro is defined.
        • NVCC device pass: Macro is not defined.
    • CUB_RDC_ENABLED: New macro, may be combined with NV_IF_TARGET to replace most usages of CUB_RUNTIME_ENABLED. Behavior:
      • RDC enabled: Macro is defined.
      • RDC not enabled: Macro is not defined.
  • NVIDIA#509: A compile-time error is now emitted when a __device__-only lambda's return type is queried from host code (requires libcu++ ≥ 1.9.0).
    • Due to limitations in the CUDA programming model, the result of this query is unreliable, and will silently return an incorrect result. This leads to difficult to debug errors.
    • When using libcu++ 1.9.0, an error will be emitted with information about work-arounds:
      • Use a named function object with a __device__-only implementation of operator().
      • Use a __host__ __device__ lambda.
      • Use cuda::proclaim_return_type (Added in libcu++ 1.9.0)
  • NVIDIA#509: Use the result type of the binary reduction operator for accumulating intermediate results in the DeviceReduce algorithm, following guidance from http://wg21.link/P2322R6.
    • This change requires host-side introspection of the binary operator's signature, and device-only extended lambda functions can no longer be used.
    • In addition to the behavioral changes, the interfaces for the Dispatch*Reduce layer have changed:
      • DispatchReduce:
        • Now accepts accumulator type as last parameter.
        • Now accepts initializer type instead of output iterator value type.
        • Constructor now accepts init as initial type instead of output iterator value type.
      • DispatchSegmentedReduce:
        • Accepts accumulator type as last parameter.
        • Accepts initializer type instead of output iterator value type.
    • Thread operators now accept parameters using different types: Equality , Inequality, InequalityWrapper, Sum, Difference, Division, Max , ArgMax, Min, ArgMin.
    • ThreadReduce now accepts accumulator type and uses a different type for prefix.
  • NVIDIA#511: Use the result type of the binary operator for accumulating intermediate results in the DeviceScan, DeviceScanByKey, and DeviceReduceByKey algorithms, following guidance from http://wg21.link/P2322R6.
    • This change requires host-side introspection of the binary operator's signature, and device-only extended lambda functions can no longer be used.
    • In addition to the behavioral changes, the interfaces for the Dispatch layer have changed:
      • DispatchScannow accepts accumulator type as a template parameter.
      • DispatchScanByKeynow accepts accumulator type as a template parameter.
      • DispatchReduceByKeynow accepts accumulator type as the last template parameter.
  • NVIDIA#527: Deprecate the debug_synchronous flags on device algorithms.
    • This flag no longer has any effect. Define CUB_DEBUG_SYNC during compilation to enable these checks.
    • Moving this option from run-time to compile-time avoids the compilation overhead of unused debugging paths in production code.

New Features

  • NVIDIA#514: Support __half in DeviceHistogram.
  • NVIDIA#516: Add support for single-threaded invocations of WarpReduce.
  • NVIDIA#516: Use __reduce_add_sync hardware acceleration for WarpReduce on supported architectures.

Bug Fixes

  • NVIDIA#481: Fix the device-wide radix sort implementations to simply copy the input to the output when begin_bit == end_bit.
  • NVIDIA#487: Fix DeviceHistogram::Even for a variety of edge cases:
    • Bin ids are now correctly computed when mixing different types for SampleT and LevelT.
    • Bin ids are now correctly computed when LevelT is an integral type and the number of levels does not evenly divide the level range.
  • NVIDIA#508: Ensure that temp_storage_bytes is properly set in the AdjacentDifferenceCopy device algorithms.
  • NVIDIA#508: Remove excessive calls to the binary operator given to the AdjacentDifferenceCopy device algorithms.
  • NVIDIA#533: Fix debugging utilities when RDC is disabled.

Other Enhancements

  • NVIDIA#448: Removed special case code for unsupported CUDA architectures.
  • NVIDIA#448: Replace several usages of __CUDA_ARCH__ with <nv/target> to handle host/device code divergence.
  • NVIDIA#448: Mark unused PTX arch parameters as legacy.
  • NVIDIA#476: Enabled additional debug logging for the onesweep radix sort implementation. Thanks to @canonizer for this contribution.
  • NVIDIA#480: Add CUB_DISABLE_BF16_SUPPORT to avoid including the cuda_bf16.h header or using the __nv_bfloat16 type.
  • NVIDIA#486: Add debug log messages for post-kernel debug synchronizations.
  • NVIDIA#490: Clarify documentation for in-place usage of DeviceScan algorithms.
  • NVIDIA#494: Clarify documentation for in-place usage of DeviceHistogram algorithms.
  • NVIDIA#495: Clarify documentation for in-place usage of DevicePartition algorithms.
  • NVIDIA#499: Clarify documentation for in-place usage of Device*Sort algorithms.
  • NVIDIA#500: Clarify documentation for in-place usage of DeviceReduce algorithms.
  • NVIDIA#501: Clarify documentation for in-place usage of DeviceRunLengthEncode algorithms.
  • NVIDIA#503: Clarify documentation for in-place usage of DeviceSelect algorithms.
  • NVIDIA#518: Fix typo in WarpMergeSort documentation.
  • NVIDIA#519: Clarify segmented sort documentation regarding the handling of elements that are not included in any segment.

CUB 1.17.2

Summary

CUB 1.17.2 is a minor bugfix release.

  • NVIDIA#547: Introduce an annotated inline namespace to prevent issues with collisions and mismatched kernel configurations across libraries. The new namespace encodes the CUB version and target SM architectures.

CUB 1.17.1

Summary

CUB 1.17.1 is a minor bugfix release.

  • NVIDIA#508: Ensure that temp_storage_bytes is properly set in the AdjacentDifferenceCopy device algorithms.
  • NVIDIA#508: Remove excessive calls to the binary operator given to the AdjacentDifferenceCopy device algorithms.
  • Fix device-side debug synchronous behavior in DeviceSegmentedSort.

CUB 1.17.0

Summary

CUB 1.17.0 is the final minor release of the 1.X series. It provides a variety of bug fixes and miscellaneous enhancements, detailed below.

Known Issues

"Run-to-run" Determinism Broken

Several CUB device algorithms are documented to provide deterministic results (per device) for non-associative reduction operators (e.g. floating-point addition). Unfortunately, the implementations of these algorithms contain performance optimizations that violate this guarantee. The DeviceReduce::ReduceByKey and DeviceScan algorithms are known to be affected. We're currently evaluating the scope and impact of correcting this in a future CUB release. See NVIDIA/cub#471 for details.

Bug Fixes

  • NVIDIA#444: Fixed DeviceSelect to work with discard iterators and mixed input/output types.
  • NVIDIA#452: Fixed install issue when CMAKE_INSTALL_LIBDIR contained nested directories. Thanks to @robertmaynard for this contribution.
  • NVIDIA#462: Fixed bug that produced incorrect results from DeviceSegmentedSort on sm_61 and sm_70.
  • NVIDIA#464: Fixed DeviceSelect::Flagged so that flags are normalized to 0 or 1.
  • NVIDIA#468: Fixed overflow issues in DeviceRadixSort given num_items close to 2^32. Thanks to @canonizer for this contribution.
  • NVIDIA#498: Fixed compiler regression in BlockAdjacentDifference. Thanks to @MKKnorr for this contribution.

Other Enhancements

  • NVIDIA#445: Remove device-sync in DeviceSegmentedSort when launched via CDP.
  • NVIDIA#449: Fixed invalid link in documentation. Thanks to @kshitij12345 for this contribution.
  • NVIDIA#450: BlockDiscontinuity: Replaced recursive-template loop unrolling with #pragma unroll. Thanks to @kshitij12345 for this contribution.
  • NVIDIA#451: Replaced the deprecated TexRefInputIterator implementation with an alias to TexObjInputIterator. This fully removes all usages of the deprecated CUDA texture reference APIs from CUB.
  • NVIDIA#456: BlockAdjacentDifference: Replaced recursive-template loop unrolling with #pragma unroll. Thanks to @kshitij12345 for this contribution.
  • NVIDIA#466: cub::DeviceAdjacentDifference API has been updated to use the new OffsetT deduction approach described in NVIDIA#212.
  • NVIDIA#470: Fix several doxygen-related warnings. Thanks to @karthikeyann for this contribution.

CUB 1.16.0

Summary

CUB 1.16.0 is a major release providing several improvements to the device scope algorithms. DeviceRadixSort now supports large (64-bit indexed) input data. A new UniqueByKey algorithm has been added to DeviceSelect. DeviceAdjacentDifference provides new SubtractLeft and SubtractRight functionality.

This release also deprecates several obsolete APIs, including type traits and BlockAdjacentDifference algorithms. Many bugfixes and documentation updates are also included.

64-bit Offsets in DeviceRadixSort Public APIs

Users frequently want to process large datasets using CUB's device-scope algorithms, but the current public APIs limit input data sizes to those that can be indexed by a 32-bit integer. Beginning with this release, CUB is updating these APIs to support 64-bit offsets, as discussed in NVIDIA#212.

The device-scope algorithms will be updated with 64-bit offset support incrementally, starting with the cub::DeviceRadixSort family of algorithms. Thanks to @canonizer for contributing this functionality.

New DeviceSelect::UniqueByKey Algorithm

cub::DeviceSelect now provides a UniqueByKey algorithm, which has been ported from Thrust. Thanks to @zasdfgbnm for this contribution.

New DeviceAdjacentDifference Algorithms

The new cub::DeviceAdjacentDifference interface, also ported from Thrust, provides SubtractLeft and SubtractRight algorithms as CUB kernels.

Deprecation Notices

Synchronous CUDA Dynamic Parallelism Support

A future version of CUB will change the debug_synchronous behavior of device-scope algorithms when invoked via CUDA Dynamic Parallelism (CDP).

This will only affect calls to CUB device-scope algorithms launched from device-side code with debug_synchronous = true. Such invocations will continue to print extra debugging information, but they will no longer synchronize after kernel launches.

Deprecated Traits

CUB provided a variety of metaprogramming type traits in order to support C++03. Since C++14 is now required, these traits have been deprecated in favor of their STL equivalents, as shown below:

Deprecated CUB Trait Replacement STL Trait
cub::If std::conditional
cub::Equals std::is_same
cub::IsPointer std::is_pointer
cub::IsVolatile std::is_volatile
cub::RemoveQualifiers std::remove_cv
cub::EnableIf std::enable_if

CUB now uses the STL traits internally, resulting in a ~6% improvement in compile time.

Misnamed cub::BlockAdjacentDifference APIs

The algorithms in cub::BlockAdjacentDifference have been deprecated, as their names did not clearly describe their intent. The FlagHeads method is now SubtractLeft, and FlagTails has been replaced by SubtractRight.

Breaking Changes

  • NVIDIA#331: Deprecate the misnamed BlockAdjacentDifference::FlagHeads and FlagTails methods. Use the new SubtractLeft and SubtractRight methods instead.
  • NVIDIA#364: Deprecate some obsolete type traits. These should be replaced by the equivalent traits in <type_traits> as described above.

New Features

  • NVIDIA#331: Port the thrust::adjacent_difference kernel and expose it as cub::DeviceAdjacentDifference.
  • NVIDIA#405: Port the thrust::unique_by_key kernel and expose it as cub::DeviceSelect::UniqueByKey. Thanks to @zasdfgbnm for this contribution.

Enhancements

  • NVIDIA#340: Allow 64-bit offsets in DeviceRadixSort public APIs. Thanks to @canonizer for this contribution.
  • NVIDIA#400: Implement a significant reduction in DeviceMergeSort compilation time.
  • NVIDIA#415: Support user-defined CMAKE_INSTALL_INCLUDEDIR values in Thrust's CMake install rules. Thanks for @robertmaynard for this contribution.

Bug Fixes

  • NVIDIA#381: Fix shared memory alignment in dyn_smem example.
  • NVIDIA#393: Fix some collisions with the min/max macros defined in windows.h.
  • NVIDIA#404: Fix bad cast in util_device.
  • NVIDIA#410: Fix CDP issues in DeviceSegmentedSort.
  • NVIDIA#411: Ensure that the nv_exec_check_disable pragma is only used on nvcc.
  • NVIDIA#418: Fix -Wsizeof-array-div warning on gcc 11. Thanks to @robertmaynard for this contribution.
  • NVIDIA#420: Fix new uninitialized variable warning in DiscardIterator on gcc 10.
  • NVIDIA#423: Fix some collisions with the small macro defined in windows.h.
  • NVIDIA#426: Fix some issues with version handling in CUB's CMake packages.
  • NVIDIA#430: Remove documentation for DeviceSpmv parameters that are absent from public APIs.
  • NVIDIA#432: Remove incorrect documentation for DeviceScan algorithms that guaranteed run-to-run deterministic results for floating-point addition.

CUB 1.15.0 (NVIDIA HPC SDK 22.1, CUDA Toolkit 11.6)

Summary

CUB 1.15.0 includes a new cub::DeviceSegmentedSort algorithm, which demonstrates up to 5000x speedup compared to cub::DeviceSegmentedRadixSort when sorting a large number of small segments. A new cub::FutureValue<T> helper allows the cub::DeviceScan algorithms to lazily load the initial_value from a pointer. cub::DeviceScan also added ScanByKey functionality.

The new DeviceSegmentedSort algorithm partitions segments into size groups. Each group is processed with specialized kernels using a variety of sorting algorithms. This approach varies the number of threads allocated for sorting each segment and utilizes the GPU more efficiently.

cub::FutureValue<T> provides the ability to use the result of a previous kernel as a scalar input to a CUB device-scope algorithm without unnecessary synchronization:

int *d_intermediate_result = ...;
intermediate_kernel<<<blocks, threads>>>(d_intermediate_result,  // output
                                         arg1,                   // input
                                         arg2);                  // input

// Wrap the intermediate pointer in a FutureValue -- no need to explicitly
// sync when both kernels are stream-ordered. The pointer is read after
// the ExclusiveScan kernel starts executing.
cub::FutureValue<int> init_value(d_intermediate_result);

cub::DeviceScan::ExclusiveScan(d_temp_storage,
                               temp_storage_bytes,
                               d_in,
                               d_out,
                               cub::Sum(),
                               init_value,
                               num_items);

Previously, an explicit synchronization would have been necessary to obtain the intermediate result, which was passed by value into ExclusiveScan. This new feature enables better performance in workflows that use cub::DeviceScan.

Deprecation Notices

A future version of CUB will change the debug_synchronous behavior of device-scope algorithms when invoked via CUDA Dynamic Parallelism (CDP).

This will only affect calls to CUB device-scope algorithms launched from device-side code with debug_synchronous = true. These algorithms will continue to print extra debugging information, but they will no longer synchronize after kernel launches.

Breaking Changes

  • NVIDIA#305: The template parameters of cub::DispatchScan have changed to support the new cub::FutureValue helper. More details under "New Features".
  • NVIDIA#377: Remove broken operator->() from cub::TransformInputIterator, since this cannot be implemented without returning a temporary object's address. Thanks to Xiang Gao (@zasdfgbnm) for this contribution.

New Features

  • NVIDIA#305: Add overloads to cub::DeviceScan algorithms that allow the output of a previous kernel to be used as initial_value without explicit synchronization. See the new cub::FutureValue helper for details. Thanks to Xiang Gao (@zasdfgbnm) for this contribution.
  • NVIDIA#354: Add cub::BlockRunLengthDecode algorithm. Thanks to Elias Stehle (@elstehle) for this contribution.
  • NVIDIA#357: Add cub::DeviceSegmentedSort, an optimized version of cub::DeviceSegmentedSort with improved load balancing and small array performance.
  • NVIDIA#376: Add "by key" overloads to cub::DeviceScan. Thanks to Xiang Gao (@zasdfgbnm) for this contribution.

Bug Fixes

  • NVIDIA#349: Doxygen and unused variable fixes.
  • NVIDIA#363: Maintenance updates for the new cub::DeviceMergeSort algorithms.
  • NVIDIA#382: Fix several -Wconversion warnings. Thanks to Matt Stack (@matt-stack) for this contribution.
  • NVIDIA#388: Fix debug assertion on MSVC when using cub::CachingDeviceAllocator.
  • NVIDIA#395: Support building with __CUDA_NO_HALF_CONVERSIONS__. Thanks to Xiang Gao (@zasdfgbnm) for this contribution.

CUB 1.14.0 (NVIDIA HPC SDK 21.9)

Summary

CUB 1.14.0 is a major release accompanying the NVIDIA HPC SDK 21.9.

This release provides the often-requested merge sort algorithm, ported from the thrust::sort implementation. Merge sort provides more flexibility than the existing radix sort by supporting arbitrary data types and comparators, though radix sorting is still faster for supported inputs. This functionality is provided through the new cub::DeviceMergeSort and cub::BlockMergeSort algorithms.

The namespace wrapping mechanism has been overhauled for 1.14. The existing macros (CUB_NS_PREFIX/CUB_NS_POSTFIX) can now be replaced by a single macro, CUB_WRAPPED_NAMESPACE, which is set to the name of the desired wrapped namespace. Defining a similar THRUST_CUB_WRAPPED_NAMESPACE macro will embed both thrust:: and cub:: symbols in the same external namespace. The prefix/postfix macros are still supported, but now require a new CUB_NS_QUALIFIER macro to be defined, which provides the fully qualified CUB namespace (e.g. ::foo::cub). See cub/util_namespace.cuh for details.

Breaking Changes

  • NVIDIA#350: When the CUB_NS_[PRE|POST]FIX macros are set, CUB_NS_QUALIFIER must also be defined to the fully qualified CUB namespace (e.g. #define CUB_NS_QUALIFIER ::foo::cub). Note that this is handled automatically when using the new [THRUST_]CUB_WRAPPED_NAMESPACE mechanism.

New Features

  • NVIDIA#322: Ported the merge sort algorithm from Thrust: cub::BlockMergeSort and cub::DeviceMergeSort are now available.
  • NVIDIA#326: Simplify the namespace wrapper macros, and detect when Thrust's symbols are in a wrapped namespace.

Bug Fixes

  • NVIDIA#160, NVIDIA#163, NVIDIA#352: Fixed several bugs in cub::DeviceSpmv and added basic tests for this algorithm. Thanks to James Wyles and Seunghwa Kang for their contributions.
  • NVIDIA#328: Fixed error handling bug and incorrect debugging output in cub::CachingDeviceAllocator. Thanks to Felix Kallenborn for this contribution.
  • NVIDIA#335: Fixed a compile error affecting clang and NVRTC. Thanks to Jiading Guo for this contribution.
  • NVIDIA#351: Fixed some errors in the cub::DeviceHistogram documentation.

Enhancements

  • NVIDIA#348: Add an example that demonstrates how to use dynamic shared memory with a CUB block algorithm. Thanks to Matthias Jouanneaux for this contribution.

CUB 1.13.1 (CUDA Toolkit 11.5)

CUB 1.13.1 is a minor release accompanying the CUDA Toolkit 11.5.

This release provides a new hook for embedding the cub:: namespace inside a custom namespace. This is intended to work around various issues related to linking multiple shared libraries that use CUB. The existing CUB_NS_PREFIX and CUB_NS_POSTFIX macros already provided this capability; this update provides a simpler mechanism that is extended to and integrated with Thrust. Simply define THRUST_CUB_WRAPPED_NAMESPACE to a namespace name, and both thrust:: and cub:: will be placed inside the new namespace. Using different wrapped namespaces for each shared library will prevent issues like those reported in NVIDIA/thrust#1401.

New Features

  • NVIDIA#326: Add THRUST_CUB_WRAPPED_NAMESPACE hooks.

CUB 1.13.0 (NVIDIA HPC SDK 21.7)

CUB 1.13.0 is the major release accompanying the NVIDIA HPC SDK 21.7 release.

Notable new features include support for striped data arrangements in block load/store utilities, bfloat16 radix sort support, and fewer restrictions on offset iterators in segmented device algorithms. Several bugs in cub::BlockShuffle, cub::BlockDiscontinuity, and cub::DeviceHistogram have been addressed. The amount of code generated in cub::DeviceScan has been greatly reduced, leading to significant compile-time improvements when targeting multiple PTX architectures.

This release also includes several user-contributed documentation fixes that will be reflected in CUB's online documentation in the coming weeks.

Breaking Changes

  • NVIDIA#320: Deprecated cub::TexRefInputIterator<T, UNIQUE_ID>. Use cub::TexObjInputIterator<T> as a replacement.

New Features

  • NVIDIA#274: Add BLOCK_LOAD_STRIPED and BLOCK_STORE_STRIPED functionality to cub::BlockLoadAlgorithm and cub::BlockStoreAlgorithm. Thanks to Matthew Nicely (@mnicely) for this contribution.
  • NVIDIA#291: cub::DeviceSegmentedRadixSort and cub::DeviceSegmentedReduce now support different types for begin/end offset iterators. Thanks to Sergey Pavlov (@psvvsp) for this contribution.
  • NVIDIA#306: Add bfloat16 support to cub::DeviceRadixSort. Thanks to Xiang Gao (@zasdfgbnm) for this contribution.
  • NVIDIA#320: Introduce a new CUB_IGNORE_DEPRECATED_API macro that disables deprecation warnings on Thrust and CUB APIs.

Bug Fixes

  • NVIDIA#277: Fixed sanitizer warnings in RadixSortScanBinsKernels. Thanks to Andy Adinets (@canonizer) for this contribution.
  • NVIDIA#287: cub::DeviceHistogram now correctly handles cases where OffsetT is not an int. Thanks to Dominique LaSalle (@nv-dlasalle) for this contribution.
  • NVIDIA#311: Fixed several bugs and added tests for the cub::BlockShuffle collective operations.
  • NVIDIA#312: Eliminate unnecessary kernel instantiations when compiling cub::DeviceScan. Thanks to Elias Stehle (@elstehle) for this contribution.
  • NVIDIA#319: Fixed out-of-bounds memory access on debugging builds of cub::BlockDiscontinuity::FlagHeadsAndTails.
  • NVIDIA#320: Fixed harmless missing return statement warning in unreachable cub::TexObjInputIterator code path.

Other Enhancements

  • Several documentation fixes are included in this release.
    • NVIDIA#275: Fixed comments describing the cub::If and cub::Equals utilities. Thanks to Rukshan Jayasekara (@rukshan99) for this contribution.
    • NVIDIA#290: Documented that cub::DeviceSegmentedReduce will produce consistent results run-to-run on the same device for pseudo-associated reduction operators. Thanks to Himanshu (@himanshu007-creator) for this contribution.
    • NVIDIA#298: CONTRIBUTING.md now refers to Thrust's build instructions for developer builds, which is the preferred way to build the CUB test harness. Thanks to Xiang Gao (@zasdfgbnm) for contributing.
    • NVIDIA#301: Expand cub::DeviceScan documentation to include in-place support and add tests. Thanks to Xiang Gao (@zasdfgbnm) for this contribution.
    • NVIDIA#307: Expand cub::DeviceRadixSort and cub::BlockRadixSort documentation to clarify stability, in-place support, and type-specific bitwise transformations. Thanks to Himanshu (@himanshu007-creator) for contributing.
    • NVIDIA#316: Move WARP_TIME_SLICING documentation to the correct location. Thanks to Peter Han (@peter9606) for this contribution.
    • NVIDIA#321: Update URLs from deprecated github.com to preferred github.io. Thanks to Lilo Huang (@lilohuang) for this contribution.

CUB 1.12.1 (CUDA Toolkit 11.4)

CUB 1.12.1 is a trivial patch release that slightly changes the phrasing of a deprecation message.

CUB 1.12.0 (NVIDIA HPC SDK 21.3)

Summary

CUB 1.12.0 is a bugfix release accompanying the NVIDIA HPC SDK 21.3 and the CUDA Toolkit 11.4.

Radix sort is now stable when both +0.0 and -0.0 are present in the input (they are treated as equivalent). Many compilation warnings and subtle overflow bugs were fixed in the device algorithms, including a long-standing bug that returned invalid temporary storage requirements when num_items was close to (but not exceeding) INT32_MAX. Support for Clang < 7.0 and MSVC < 2019 (aka 19.20/16.0/14.20) is now deprecated.

Breaking Changes

  • NVIDIA#256: Deprecate Clang < 7 and MSVC < 2019.

New Features

  • NVIDIA#218: Radix sort now treats -0.0 and +0.0 as equivalent for floating point types, which is required for the sort to be stable. Thanks to Andy Adinets for this contribution.

Bug Fixes

  • NVIDIA#247: Suppress newly triggered warnings in Clang. Thanks to Andrew Corrigan for this contribution.
  • NVIDIA#249: Enable stricter warning flags. This fixes a number of outstanding issues:
    • NVIDIA#221: Overflow in temp_storage_bytes when num_items close to (but not over) INT32_MAX.
    • NVIDIA#228: CUB uses non-standard C++ extensions that break strict compilers.
    • NVIDIA#257: Warning when compiling GridEvenShare with unsigned offsets.
  • NVIDIA#258: Use correct OffsetT in DispatchRadixSort::InitPassConfig. Thanks to Felix Kallenborn for this contribution.
  • NVIDIA#259: Remove some problematic __forceinline__ annotations.

Other Enhancements

  • NVIDIA#123: Fix incorrect issue number in changelog. Thanks to Peet Whittaker for this contribution.

CUB 1.11.0 (CUDA Toolkit 11.3)

Summary

CUB 1.11.0 is a major release accompanying the CUDA Toolkit 11.3 release, providing bugfixes and performance enhancements.

It includes a new DeviceRadixSort backend that improves performance by up to 2x on supported keys and hardware.

Our CMake package and build system continue to see improvements with add_subdirectory support, installation rules, status messages, and other features that make CUB easier to use from CMake projects.

The release includes several other bugfixes and modernizations, and received updates from 11 contributors.

Breaking Changes

  • NVIDIA#201: The intermediate accumulator type used when DeviceScan is invoked with different input/output types is now consistent with P0571. This may produce different results for some edge cases when compared with earlier releases of CUB.

New Features

  • NVIDIA#204: Faster DeviceRadixSort, up to 2x performance increase for 32/64-bit keys on Pascal and up (SM60+). Thanks to Andy Adinets for this contribution.
  • Unroll loops in BlockRadixRank to improve performance for 32-bit keys by 1.5-2x on Clang CUDA. Thanks to Justin Lebar for this contribution.
  • NVIDIA#200: Allow CUB to be added to CMake projects via add_subdirectory.
  • NVIDIA#214: Optionally add install rules when included with CMake's add_subdirectory. Thanks to Kai Germaschewski for this contribution.

Bug Fixes

  • NVIDIA#215: Fix integer truncation in AgentReduceByKey, AgentScan, and AgentSegmentFixup. Thanks to Rory Mitchell for this contribution.
  • NVIDIA#225: Fix compile-time regression when defining CUB_NS_PREFIX /CUB_NS_POSTFIX macro. Thanks to Elias Stehle for this contribution.
  • NVIDIA#210: Fix some edge cases in DeviceScan:
    • Use values from the input when padding temporary buffers. This prevents custom functors from getting unexpected values.
    • Prevent integer truncation when using large indices via the DispatchScan layer.
    • Use timesliced reads/writes for types > 128 bytes.
  • NVIDIA#217: Fix and add test for cmake package install rules. Thanks to Keith Kraus and Kai Germaschewski for testing and discussion.
  • NVIDIA#170, NVIDIA#233: Update CUDA version checks to behave on Clang CUDA and nvc++. Thanks to Artem Belevich, Andrew Corrigan, and David Olsen for these contributions.
  • NVIDIA#220, NVIDIA#216: Various fixes for Clang CUDA. Thanks to Andrew Corrigan for these contributions.
  • NVIDIA#231: Fix signedness mismatch warnings in unit tests.
  • NVIDIA#231: Suppress GPU deprecation warnings.
  • NVIDIA#214: Use semantic versioning rules for our CMake package's compatibility checks. Thanks to Kai Germaschewski for this contribution.
  • NVIDIA#214: Use FindPackageHandleStandardArgs to print standard status messages when our CMake package is found. Thanks to Kai Germaschewski for this contribution.
  • NVIDIA#207: Fix CubDebug usage in CachingDeviceAllocator::DeviceAllocate. Thanks to Andreas Hehn for this contribution.
  • Fix documentation for DevicePartition. Thanks to ByteHamster for this contribution.
  • Clean up unused code in DispatchScan. Thanks to ByteHamster for this contribution.

Other Enhancements

  • NVIDIA#213: Remove tuning policies for unsupported hardware (<SM35).
  • References to the old Github repository and branch names were updated.
    • Github's thrust/cub repository is now NVIDIA/cub
    • Development has moved from the master branch to the main branch.

CUB 1.10.0 (NVIDIA HPC SDK 20.9, CUDA Toolkit 11.2)

Summary

CUB 1.10.0 is the major release accompanying the NVIDIA HPC SDK 20.9 release and the CUDA Toolkit 11.2 release. It drops support for C++03, GCC < 5, Clang < 6, and MSVC < 2017. It also overhauls CMake support. Finally, we now have a Code of Conduct for contributors: https://github.com/NVIDIA/cub/blob/main/CODE_OF_CONDUCT.md

Breaking Changes

  • C++03 is no longer supported.
  • GCC < 5, Clang < 6, and MSVC < 2017 are no longer supported.
  • C++11 is deprecated. Using this dialect will generate a compile-time warning. These warnings can be suppressed by defining CUB_IGNORE_DEPRECATED_CPP_DIALECT or CUB_IGNORE_DEPRECATED_CPP_11. Suppression is only a short term solution. We will be dropping support for C++11 in the near future.
  • CMake < 3.15 is no longer supported.
  • The default branch on GitHub is now called main.

Other Enhancements

  • Added install targets to CMake builds.
  • C++17 support.

Bug Fixes

  • NVIDIA/thrust#1244: Check for macro collisions with system headers during header testing.
  • NVIDIA/thrust#1153: Switch to placement new instead of assignment to construct items in uninitialized memory. Thanks to Hugh Winkler for this contribution.
  • NVIDIA#38: Fix cub::DeviceHistogram for size_t OffsetTs. Thanks to Leo Fang for this contribution.
  • NVIDIA#35: Fix GCC-5 maybe-uninitialized warning. Thanks to Rong Ou for this contribution.
  • NVIDIA#36: Qualify namespace for va_printf in _CubLog. Thanks to Andrei Tchouprakov for this contribution.

CUB 1.9.10-1 (NVIDIA HPC SDK 20.7, CUDA Toolkit 11.1)

Summary

CUB 1.9.10-1 is the minor release accompanying the NVIDIA HPC SDK 20.7 release and the CUDA Toolkit 11.1 release.

Bug Fixes

  • NVIDIA/thrust#1217: Move static local in cub::DeviceCount to a separate host-only function because NVC++ doesn't support static locals in host-device functions.

CUB 1.9.10 (NVIDIA HPC SDK 20.5)

Summary

Thrust 1.9.10 is the release accompanying the NVIDIA HPC SDK 20.5 release. It adds CMake find_package support. C++03, C++11, GCC < 5, Clang < 6, and MSVC < 2017 are now deprecated. Starting with the upcoming 1.10.0 release, C++03 support will be dropped entirely.

Breaking Changes

  • Thrust now checks that it is compatible with the version of CUB found in your include path, generating an error if it is not. If you are using your own version of CUB, it may be too old. It is recommended to simply delete your own version of CUB and use the version of CUB that comes with Thrust.
  • C++03 and C++11 are deprecated. Using these dialects will generate a compile-time warning. These warnings can be suppressed by defining CUB_IGNORE_DEPRECATED_CPP_DIALECT (to suppress C++03 and C++11 deprecation warnings) or CUB_IGNORE_DEPRECATED_CPP_11 (to suppress C++11 deprecation warnings). Suppression is only a short term solution. We will be dropping support for C++03 in the 1.10.0 release and C++11 in the near future.
  • GCC < 5, Clang < 6, and MSVC < 2017 are deprecated. Using these compilers will generate a compile-time warning. These warnings can be suppressed by defining CUB_IGNORE_DEPRECATED_COMPILER. Suppression is only a short term solution. We will be dropping support for these compilers in the near future.

New Features

  • CMake find_package support. Just point CMake at the cmake folder in your CUB include directory (ex: cmake -DCUB_DIR=/usr/local/cuda/include/cub/cmake/ .) and then you can add CUB to your CMake project with find_package(CUB REQUIRED CONFIG).

CUB 1.9.9 (CUDA 11.0)

Summary

CUB 1.9.9 is the release accompanying the CUDA Toolkit 11.0 release. It introduces CMake support, version macros, platform detection machinery, and support for NVC++, which uses Thrust (and thus CUB) to implement GPU-accelerated C++17 Parallel Algorithms. Additionally, the scan dispatch layer was refactored and modernized. C++03, C++11, GCC < 5, Clang < 6, and MSVC < 2017 are now deprecated. Starting with the upcoming 1.10.0 release, C++03 support will be dropped entirely.

Breaking Changes

  • Thrust now checks that it is compatible with the version of CUB found in your include path, generating an error if it is not. If you are using your own version of CUB, it may be too old. It is recommended to simply delete your own version of CUB and use the version of CUB that comes with Thrust.
  • C++03 and C++11 are deprecated. Using these dialects will generate a compile-time warning. These warnings can be suppressed by defining CUB_IGNORE_DEPRECATED_CPP_DIALECT (to suppress C++03 and C++11 deprecation warnings) or CUB_IGNORE_DEPRECATED_CPP11 (to suppress C++11 deprecation warnings). Suppression is only a short term solution. We will be dropping support for C++03 in the 1.10.0 release and C++11 in the near future.
  • GCC < 5, Clang < 6, and MSVC < 2017 are deprecated. Using these compilers will generate a compile-time warning. These warnings can be suppressed by defining CUB_IGNORE_DEPRECATED_COMPILER. Suppression is only a short term solution. We will be dropping support for these compilers in the near future.

New Features

  • CMake support. Thanks to Francis Lemaire for this contribution.
  • Refactorized and modernized scan dispatch layer. Thanks to Francis Lemaire for this contribution.
  • Policy hooks for device-wide reduce, scan, and radix sort facilities to simplify tuning and allow users to provide custom policies. Thanks to Francis Lemaire for this contribution.
  • <cub/version.cuh>: CUB_VERSION, CUB_VERSION_MAJOR, CUB_VERSION_MINOR, CUB_VERSION_SUBMINOR, and CUB_PATCH_NUMBER.
  • Platform detection machinery:
    • <cub/util_cpp_dialect.cuh>: Detects the C++ standard dialect.
    • <cub/util_compiler.cuh>: host and device compiler detection.
    • <cub/util_deprecated.cuh>: CUB_DEPRECATED.
    • <cub/config.cuh>: Includes <cub/util_arch.cuh>, <cub/util_compiler.cuh>, <cub/util_cpp_dialect.cuh>, <cub/util_deprecated.cuh>, <cub/util_macro.cuh>, <cub/util_namespace.cuh>`
  • cub::DeviceCount and cub::DeviceCountUncached, caching abstractions for cudaGetDeviceCount.

Other Enhancements

  • Lazily initialize the per-device CUDAattribute caches, because CUDA context creation is expensive and adds up with large CUDA binaries on machines with many GPUs. Thanks to the NVIDIA PyTorch team for bringing this to our attention.
  • Make cub::SwitchDevice avoid setting/resetting the device if the current device is the same as the target device.

Bug Fixes

  • Add explicit failure parameter to CAS in the CUB attribute cache to workaround a GCC 4.8 bug.
  • Revert a change in reductions that changed the signedness of the lane_id variable to suppress a warning, as this introduces a bug in optimized device code.
  • Fix initialization in cub::ExclusiveSum. Thanks to Conor Hoekstra for this contribution.
  • Fix initialization of the std::array in the CUB attribute cache.
  • Fix -Wsign-compare warnings. Thanks to Elias Stehle for this contribution.
  • Fix test_block_reduce.cu to build without parameters. Thanks to Francis Lemaire for this contribution.
  • Add missing includes to grid_even_share.cuh. Thanks to Francis Lemaire for this contribution.
  • Add missing includes to thread_search.cuh. Thanks to Francis Lemaire for this contribution.
  • Add missing includes to cub.cuh. Thanks to Felix Kallenborn for this contribution.

CUB 1.9.8-1 (NVIDIA HPC SDK 20.3)

Summary

CUB 1.9.8-1 is a variant of 1.9.8 accompanying the NVIDIA HPC SDK 20.3 release. It contains modifications necessary to serve as the implementation of NVC++'s GPU-accelerated C++17 Parallel Algorithms.

CUB 1.9.8 (CUDA 11.0 Early Access)

Summary

CUB 1.9.8 is the first release of CUB to be officially supported and included in the CUDA Toolkit. When compiling CUB in C++11 mode, CUB now caches calls to CUDA attribute query APIs, which improves performance of these queries by 20x to 50x when they are called concurrently by multiple host threads.

Enhancements

  • (C++11 or later) Cache calls to cudaFuncGetAttributes and cudaDeviceGetAttribute within cub::PtxVersion and cub::SmVersion. These CUDA APIs acquire locks to CUDA driver/runtime mutex and perform poorly under contention; with the caching, they are 20 to 50x faster when called concurrently. Thanks to Bilge Acun for bringing this issue to our attention.
  • DispatchReduce now takes an OutputT template parameter so that users can specify the intermediate type explicitly.
  • Radix sort tuning policies updates to fix performance issues for element types smaller than 4 bytes.

Bug Fixes

  • Change initialization style from copy initialization to direct initialization (which is more permissive) in AgentReduce to allow a wider range of types to be used with it.
  • Fix bad signed/unsigned comparisons in WarpReduce.
  • Fix computation of valid lanes in warp-level reduction primitive to correctly handle the case where there are 0 input items per warp.

CUB 1.8.0

Summary

CUB 1.8.0 introduces changes to the cub::Shuffle* interfaces.

Breaking Changes

  • The interfaces of cub::ShuffleIndex, cub::ShuffleUp, and cub::ShuffleDown have been changed to allow for better computation of the PTX SHFL control constant for logical warps smaller than 32 threads.

Bug Fixes

  • #112: Fix cub::WarpScan's broadcast of warp-wide aggregate for logical warps smaller than 32 threads.

CUB 1.7.5

Summary

CUB 1.7.5 adds support for radix sorting __half keys and improved sorting performance for 1 byte keys. It was incorporated into Thrust 1.9.2.

Enhancements

  • Radix sort support for __half keys.
  • Radix sort tuning policy updates to improve 1 byte key performance.

Bug Fixes

  • Syntax tweaks to mollify Clang.
  • #127: cub::DeviceRunLengthEncode::Encode returns incorrect results.
  • #128: 7-bit sorting passes fail for SM61 with large values.

CUB 1.7.4

Summary

CUB 1.7.4 is a minor release that was incorporated into Thrust 1.9.1-2.

Bug Fixes

  • #114: Can't pair non-trivially-constructible values in radix sort.
  • #115: cub::WarpReduce segmented reduction is broken in CUDA 9 for logical warp sizes smaller than 32.

CUB 1.7.3

Summary

CUB 1.7.3 is a minor release.

Bug Fixes

  • #110: cub::DeviceHistogram null-pointer exception bug for iterator inputs.

CUB 1.7.2

Summary

CUB 1.7.2 is a minor release.

Bug Fixes

  • #108: Device-wide reduction is now "run-to-run" deterministic for pseudo-associative reduction operators (like floating point addition).

CUB 1.7.1

Summary

CUB 1.7.1 delivers improved radix sort performance on SM7x (Volta) GPUs and a number of bug fixes.

Enhancements

  • Radix sort tuning policies updated for SM7x (Volta).

Bug Fixes

  • #104: uint64_t cub::WarpReduce broken for CUB 1.7.0 on CUDA 8 and older.
  • #103: Can't mix Thrust from CUDA 9.0 and CUB.
  • #102: CUB pulls in windows.h which defines min/max macros that conflict with std::min/std::max.
  • #99: Radix sorting crashes NVCC on Windows 10 for SM52.
  • #98: cuda-memcheck: --tool initcheck failed with lineOfSight.
  • #94: Git clone size.
  • #93: Accept iterators for segment offsets.
  • #87: CUB uses anonymous unions which is not valid C++.
  • #44: Check for C++11 is incorrect for Visual Studio 2013.

CUB 1.7.0

Summary

CUB 1.7.0 brings support for CUDA 9.0 and SM7x (Volta) GPUs. It is compatible with independent thread scheduling. It was incorporated into Thrust 1.9.0-5.

Breaking Changes

  • Remove cub::WarpAll and cub::WarpAny. These functions served to emulate __all and __any functionality for SM1x devices, which did not have those operations. However, SM1x devices are now deprecated in CUDA, and the interfaces of these two functions are now lacking the lane-mask needed for collectives to run on SM7x and newer GPUs which have independent thread scheduling.

Other Enhancements

  • Remove any assumptions of implicit warp synchronization to be compatible with SM7x's (Volta) independent thread scheduling.

Bug Fixes

  • #86: Incorrect results with reduce-by-key.

CUB 1.6.4

Summary

CUB 1.6.4 improves radix sorting performance for SM5x (Maxwell) and SM6x (Pascal) GPUs.

Enhancements

  • Radix sort tuning policies updated for SM5x (Maxwell) and SM6x (Pascal) - 3.5B and 3.4B 32 byte keys/s on TitanX and GTX 1080, respectively.

Bug Fixes

  • Restore fence work-around for scan (reduce-by-key, etc.) hangs in CUDA 8.5.
  • #65: cub::DeviceSegmentedRadixSort should allow inputs to have pointer-to-const type.
  • Mollify Clang device-side warnings.
  • Remove out-dated MSVC project files.

CUB 1.6.3

Summary

CUB 1.6.3 improves support for Windows, changes cub::BlockLoad/cub::BlockStore interface to take the local data type, and enhances radix sort performance for SM6x (Pascal) GPUs.

Breaking Changes

  • cub::BlockLoad and cub::BlockStore are now templated by the local data type, instead of the Iterator type. This allows for output iterators having void as their value_type (e.g. discard iterators).

Other Enhancements

  • Radix sort tuning policies updated for SM6x (Pascal) GPUs - 6.2B 4 byte keys/s on GP100.
  • Improved support for Windows (warnings, alignment, etc).

Bug Fixes

  • #74: cub::WarpReduce executes reduction operator for out-of-bounds items.
  • #72: cub:InequalityWrapper::operator should be non-const.
  • #71: cub::KeyValuePair won't work if Key has non-trivial constructor.
  • #69: cub::BlockStore::Storedoesn't compile ifOutputIteratorT::value_typeisn'tT`.
  • #68: cub::TilePrefixCallbackOp::WarpReduce doesn't permit PTX arch specialization.

CUB 1.6.2 (previously 1.5.5)

Summary

CUB 1.6.2 (previously 1.5.5) improves radix sort performance for SM6x (Pascal) GPUs.

Enhancements

  • Radix sort tuning policies updated for SM6x (Pascal) GPUs.

Bug Fixes

  • Fix AArch64 compilation of cub::CachingDeviceAllocator.

CUB 1.6.1 (previously 1.5.4)

Summary

CUB 1.6.1 (previously 1.5.4) is a minor release.

Bug Fixes

  • Fix radix sorting bug introduced by scan refactorization.

CUB 1.6.0 (previously 1.5.3)

Summary

CUB 1.6.0 changes the scan and reduce interfaces. Exclusive scans now accept an "initial value" instead of an "identity value". Scans and reductions now support differing input and output sequence types. Additionally, many bugs have been fixed.

Breaking Changes

  • Device/block/warp-wide exclusive scans have been revised to now accept an "initial value" (instead of an "identity value") for seeding the computation with an arbitrary prefix.
  • Device-wide reductions and scans can now have input sequence types that are different from output sequence types (as long as they are convertible).

Other Enhancements

  • Reduce repository size by moving the doxygen binary to doc repository.
  • Minor reduction in cub::BlockScan instruction counts.

Bug Fixes

  • Issue #55: Warning in cub/device/dispatch/dispatch_reduce_by_key.cuh.
  • Issue #59: cub::DeviceScan::ExclusiveSum can't prefix sum of float into double.
  • Issue #58: Infinite loop in cub::CachingDeviceAllocator::NearestPowerOf.
  • Issue #47: cub::CachingDeviceAllocator needs to clean up CUDA global error state upon successful retry.
  • Issue #46: Very high amount of needed memory from the cub::DeviceHistogram::HistogramEven.
  • Issue #45: cub::CachingDeviceAllocator fails with debug output enabled

CUB 1.5.2

Summary

CUB 1.5.2 enhances cub::CachingDeviceAllocator and improves scan performance for SM5x (Maxwell).

Enhancements

  • Improved medium-size scan performance on SM5x (Maxwell).
  • Refactored cub::CachingDeviceAllocator:
    • Now spends less time locked.
    • Uses C++11's std::mutex when available.
    • Failure to allocate a block from the runtime will retry once after freeing cached allocations.
    • Now respects max-bin, fixing an issue where blocks in excess of max-bin were still being retained in the free cache.

Bug fixes:

  • Fix for generic-type reduce-by-key cub::WarpScan for SM3x and newer GPUs.

CUB 1.5.1

Summary

CUB 1.5.1 is a minor release.

Bug Fixes

  • Fix for incorrect cub::DeviceRadixSort output for some small problems on SM52 (Mawell) GPUs.
  • Fix for macro redefinition warnings when compiling thrust::sort.

CUB 1.5.0

CUB 1.5.0 introduces segmented sort and reduction primitives.

New Features:

  • Segmented device-wide operations for device-wide sort and reduction primitives.

Bug Fixes:

  • #36: cub::ThreadLoad generates compiler errors when loading from pointer-to-const.
  • #29: cub::DeviceRadixSort::SortKeys<bool> yields compiler errors.
  • #26: Misaligned address after cub::DeviceRadixSort::SortKeys.
  • #25: Fix for incorrect results and crashes when radix sorting 0-length problems.
  • Fix CUDA 7.5 issues on SM52 GPUs with SHFL-based warp-scan and warp-reduction on non-primitive data types (e.g. user-defined structs).
  • Fix small radix sorting problems where 0 temporary bytes were required and users code was invoking malloc(0) on some systems where that returns NULL. CUB assumed the user was asking for the size again and not running the sort.

CUB 1.4.1

Summary

CUB 1.4.1 is a minor release.

Enhancements

  • Allow cub::DeviceRadixSort and cub::BlockRadixSort on bool types.

Bug Fixes

  • Fix minor CUDA 7.0 performance regressions in cub::DeviceScan and cub::DeviceReduceByKey.
  • Remove requirement for callers to define the CUB_CDP macro when invoking CUB device-wide rountines using CUDA dynamic parallelism.
  • Fix headers not being included in the proper order (or missing includes) for some block-wide functions.

CUB 1.4.0

Summary

CUB 1.4.0 adds cub::DeviceSpmv, cub::DeviceRunLength::NonTrivialRuns, improves cub::DeviceHistogram, and introduces support for SM5x (Maxwell) GPUs.

New Features:

  • cub::DeviceSpmv methods for multiplying sparse matrices by dense vectors, load-balanced using a merge-based parallel decomposition.
  • cub::DeviceRadixSort sorting entry-points that always return the sorted output into the specified buffer, as opposed to the cub::DoubleBuffer in which it could end up in either buffer.
  • cub::DeviceRunLengthEncode::NonTrivialRuns for finding the starting offsets and lengths of all non-trivial runs (i.e., length > 1) of keys in a given sequence. Useful for top-down partitioning algorithms like MSD sorting of very-large keys.

Other Enhancements

  • Support and performance tuning for SM5x (Maxwell) GPUs.
  • Updated cub::DeviceHistogram implementation that provides the same "histogram-even" and "histogram-range" functionality as IPP/NPP. Provides extremely fast and, perhaps more importantly, very uniform performance response across diverse real-world datasets, including pathological (homogeneous) sample distributions.

CUB 1.3.2

Summary

CUB 1.3.2 is a minor release.

Bug Fixes

  • Fix cub::DeviceReduce where reductions of small problems (small enough to only dispatch a single thread block) would run in the default stream (stream zero) regardless of whether an alternate stream was specified.

CUB 1.3.1

Summary

CUB 1.3.1 is a minor release.

Bug Fixes

  • Workaround for a benign WAW race warning reported by cuda-memcheck in cub::BlockScan specialized for BLOCK_SCAN_WARP_SCANS algorithm.
  • Fix bug in cub::DeviceRadixSort where the algorithm may sort more key bits than the caller specified (up to the nearest radix digit).
  • Fix for ~3% cub::DeviceRadixSort performance regression on SM2x (Fermi) and SM3x (Kepler) GPUs.

CUB 1.3.0

Summary

CUB 1.3.0 improves how thread blocks are expressed in block- and warp-wide primitives and adds an enhanced version of cub::WarpScan.

Breaking Changes

  • CUB's collective (block-wide, warp-wide) primitives underwent a minor interface refactoring:
    • To provide the appropriate support for multidimensional thread blocks, The interfaces for collective classes are now template-parameterized by X, Y, and Z block dimensions (with BLOCK_DIM_Y and BLOCK_DIM_Z being optional, and BLOCK_DIM_X replacing BLOCK_THREADS). Furthermore, the constructors that accept remapped linear thread-identifiers have been removed: all primitives now assume a row-major thread-ranking for multidimensional thread blocks.
    • To allow the host program (compiled by the host-pass) to accurately determine the device-specific storage requirements for a given collective (compiled for each device-pass), the interfaces for collective classes are now (optionally) template-parameterized by the desired PTX compute capability. This is useful when aliasing collective storage to shared memory that has been allocated dynamically by the host at the kernel call site.
    • Most CUB programs having typical 1D usage should not require any changes to accomodate these updates.

New Features

  • Added "combination" cub::WarpScan methods for efficiently computing both inclusive and exclusive prefix scans (and sums).

Bug Fixes

  • Fix for bug in cub::WarpScan (which affected cub::BlockScan and cub::DeviceScan) where incorrect results (e.g., NAN) would often be returned when parameterized for floating-point types (fp32, fp64).
  • Workaround for ptxas error when compiling with with -G flag on Linux (for debug instrumentation).
  • Fixes for certain scan scenarios using custom scan operators where code compiled for SM1x is run on newer GPUs of higher compute-capability: the compiler could not tell which memory space was being used collective operations and was mistakenly using global ops instead of shared ops.

CUB 1.2.3

Summary

CUB 1.2.3 is a minor release.

Bug Fixes

  • Fixed access violation bug in cub::DeviceReduce::ReduceByKey for non-primitive value types.
  • Fixed code-snippet bug in ArgIndexInputIteratorT documentation.

CUB 1.2.2

Summary

CUB 1.2.2 adds a new variant of cub::BlockReduce and MSVC project solections for examples.

New Features

  • MSVC project solutions for device-wide and block-wide examples
  • New algorithmic variant of cub::BlockReduce for improved performance when using commutative operators (e.g., numeric addition).

Bug Fixes

  • Inclusion of Thrust headers in a certain order prevented CUB device-wide primitives from working properly.

CUB 1.2.0

Summary

CUB 1.2.0 adds cub::DeviceReduce::ReduceByKey and cub::DeviceReduce::RunLengthEncode and support for CUDA 6.0.

New Features

  • cub::DeviceReduce::ReduceByKey.
  • cub::DeviceReduce::RunLengthEncode.

Other Enhancements

  • Improved cub::DeviceScan, cub::DeviceSelect, cub::DevicePartition performance.
  • Documentation and testing:
    • Added performance-portability plots for many device-wide primitives.
    • Explain that iterator (in)compatibilities with CUDA 5.0 (and older) and Thrust 1.6 (and older).
  • Revised the operation of temporary tile status bookkeeping for cub::DeviceScan (and similar) to be safe for current code run on future platforms (now uses proper fences).

Bug Fixes

  • Fix cub::DeviceScan bug where Windows alignment disagreements between host and device regarding user-defined data types would corrupt tile status.
  • Fix cub::BlockScan bug where certain exclusive scans on custom data types for the BLOCK_SCAN_WARP_SCANS variant would return incorrect results for the first thread in the block.
  • Added workaround to make cub::TexRefInputIteratorT work with CUDA 6.0.

CUB 1.1.1

Summary

CUB 1.1.1 introduces texture and cache modifier iterators, descending sorting, cub::DeviceSelect, cub::DevicePartition, cub::Shuffle*, and cub::MaxSMOccupancy. Additionally, scan and sort performance for older GPUs has been improved and many bugs have been fixed.

Breaking Changes

  • Refactored block-wide I/O (cub::BlockLoad and cub::BlockStore), removing cache-modifiers from their interfaces. cub::CacheModifiedInputIterator and cub::CacheModifiedOutputIterator should now be used with cub::BlockLoad and cub::BlockStore to effect that behavior.

New Features

  • cub::TexObjInputIterator, cub::TexRefInputIterator, cub::CacheModifiedInputIterator, and cub::CacheModifiedOutputIterator types for loading & storing arbitrary types through the cache hierarchy. They are compatible with Thrust.
  • Descending sorting for cub::DeviceRadixSort and cub::BlockRadixSort.
  • Min, max, arg-min, and arg-max operators for cub::DeviceReduce.
  • cub::DeviceSelect (select-unique, select-if, and select-flagged).
  • cub::DevicePartition (partition-if, partition-flagged).
  • Generic cub::ShuffleUp, cub::ShuffleDown, and cub::ShuffleIndex for warp-wide communication of arbitrary data types (SM3x and up).
  • cub::MaxSmOccupancy for accurately determining SM occupancy for any given kernel function pointer.

Other Enhancements

  • Improved cub::DeviceScan and cub::DeviceRadixSort performance for older GPUs (SM1x to SM3x).
  • Renamed device-wide stream_synchronous param to debug_synchronous to avoid confusion about usage.
  • Documentation improvements:
    • Added simple examples of device-wide methods.
    • Improved doxygen documentation and example snippets.
  • Improved test coverege to include up to 21,000 kernel variants and 851,000 unit tests (per architecture, per platform).

Bug Fixes

  • Fix misc `cub::DeviceScan, BlockScan, DeviceReduce, and BlockReduce bugs when operating on non-primitive types for older architectures SM1x.
  • SHFL-based scans and reductions produced incorrect results for multi-word types (size > 4B) on Linux.
  • For cub::WarpScan-based scans, not all threads in the first warp were entering the prefix callback functor.
  • cub::DeviceRadixSort had a race condition with key-value pairs for pre-SM35 architectures.
  • cub::DeviceRadixSor bitfield-extract behavior with long keys on 64-bit Linux was incorrect.
  • cub::BlockDiscontinuity failed to compile for types other than int32_t/uint32_t.
  • CUDA Dynamic Parallelism (CDP, e.g. device-callable) versions of device-wide methods now report the same temporary storage allocation size requirement as their host-callable counterparts.

CUB 1.0.2

Summary

CUB 1.0.2 is a minor release.

Bug Fixes

  • Corrections to code snippet examples for cub::BlockLoad, cub::BlockStore, and cub::BlockDiscontinuity.
  • Cleaned up unnecessary/missing header includes. You can now safely include a specific .cuh (instead of cub.cuh).
  • Bug/compilation fixes for cub::BlockHistogram.

CUB 1.0.1

Summary

CUB 1.0.1 adds cub::DeviceRadixSort and cub::DeviceScan. Numerous other performance and correctness fixes and included.

Breaking Changes

  • New collective interface idiom (specialize/construct/invoke).

New Features

  • cub::DeviceRadixSort. Implements short-circuiting for homogenous digit passes.
  • cub::DeviceScan. Implements single-pass "adaptive-lookback" strategy.

Other Enhancements

  • Significantly improved documentation (with example code snippets).
  • More extensive regression test suit for aggressively testing collective variants.
  • Allow non-trially-constructed types (previously unions had prevented aliasing temporary storage of those types).
  • Improved support for SM3x SHFL (collective ops now use SHFL for types larger than 32 bits).
  • Better code generation for 64-bit addressing within cub::BlockLoad/cub::BlockStore.
  • cub::DeviceHistogram now supports histograms of arbitrary bins.
  • Updates to accommodate CUDA 5.5 dynamic parallelism.

Bug Fixes

  • Workarounds for SM10 codegen issues in uncommonly-used cub::WarpScan/cub::WarpReduce specializations.

CUB 0.9.4

Summary

CUB 0.9.3 is a minor release.

Enhancements

  • Various documentation updates and corrections.

Bug Fixes

  • Fixed compilation errors for SM1x.
  • Fixed compilation errors for some WarpScan entrypoints on SM3x and up.

CUB 0.9.3

Summary

CUB 0.9.3 adds histogram algorithms and work management utility descriptors.

New Features

  • cub::DevicHistogram256.
  • cub::BlockHistogram256.
  • cub::BlockScan algorithm variant BLOCK_SCAN_RAKING_MEMOIZE, which trades more register consumption for less shared memory I/O.
  • cub::GridQueue, cub::GridEvenShare, work management utility descriptors.

Other Enhancements

  • Updates to cub::BlockRadixRank to use cub::BlockScan, which improves performance on SM3x by using SHFL.
  • Allow types other than builtin types to be used in cub::WarpScan::*Sum methods if they only have operator+ overloaded. Previously they also required to support assignment from int(0).
  • Update cub::BlockReduce's BLOCK_REDUCE_WARP_REDUCTIONS algorithm to work even when block size is not an even multiple of warp size.
  • Refactoring of cub::DeviceAllocator interface and cub::CachingDeviceAllocator implementation.

CUB 0.9.2

Summary

CUB 0.9.2 adds cub::WarpReduce.

New Features

  • cub::WarpReduce, which uses the SHFL instruction when applicable. cub::BlockReduce now uses this cub::WarpReduce instead of implementing its own.

Enhancements

  • Documentation updates and corrections.

Bug Fixes

  • Fixes for 64-bit Linux compilation warnings and errors.

CUB 0.9.1

Summary

CUB 0.9.1 is a minor release.

Bug Fixes

  • Fix for ambiguity in cub::BlockScan::Reduce between generic reduction and summation. Summation entrypoints are now called ::Sum(), similar to the convention in cub::BlockScan.
  • Small edits to documentation and download tracking.

CUB 0.9.0

Summary

Initial preview release. CUB is the first durable, high-performance library of cooperative block-level, warp-level, and thread-level primitives for CUDA kernel programming.