Skip to content

Latest commit

 

History

History
691 lines (566 loc) · 52.8 KB

CHANGELOG.md

File metadata and controls

691 lines (566 loc) · 52.8 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[dev]

Added

Changed

  • Removed dpctl.tensor.numpy_usm_shared obsolete class and associated tests which were being skipped

Fixed

[0.14.5] - 07/17/2023

Added

  • Added dpctl.tensor.log2 and dpctl.tensor.log10: #1267
  • Added dpctl.tensor.negative, dpctl.tensor.positive, dpctl.tensor.square #1268
  • Added dpctl.tensor.logical_not, dpctl.tensor.logical_and, dpctl.tensor.logical_or, dpctl.tensor.logical_xor #1270

Changed

  • dpctl.tensor.astype behavior for newdtype=None changes #1261
  • dpctl.tensor.usm_ndaray constructor default value of dtype keyword argument changed to None: #1265
  • Support for out arguments that overlap with inputs for unary elementwise functions#1281
  • Copying from one array to another a no-op if both arrays view into the same memory #1284

[0.14.4] - 06/14/2023

Added

  • Added dpctl.tensor.less_equal, dpctl.tensor.greater, dpctl.tensor.greater_equal: #1239

Changed

  • Optimized in-place arithmetic operations for updating matrix with rows/columns via broadcasting: #1244

Fixed

  • Fixed handling of 0d arrays in dpctl.tensor.sum: #1238

[0.14.3] - 06/13/2023

Added

  • Added support of axis=None in dpctl.tensor.concat #1125
  • Added caching for dpctl.SyclDevice.filter_string property #1127
  • Added dpctl.tensor.isdtype from array API #1133
  • Added dpctl.tensor.unstack, dpctl.tensor.moveaxis, dpctl.tensor.swapaxes #1137, #1174
  • Allow for mutation of dpctl.tensor.usm_ndarray.flags.writable #1141
  • Added dpctl.tensor.where from array API #1147
  • Include libtensor headers in dpctl installation layout #1185
  • Added new properties of dpctl.tensor.usm_ndarray object #1199
  • Added a list of unary and binary elementwise functions from array API:
    • #1203: dpctl.tensor.add, dpctl.tensor.divide, dpctl.tensor.isnan, dpctl.tensor.isinf, dpctl.tensor.isfinite, dpctl.tensor.cos, dpctl.tensor.abs, dpctl.tensor.equal
    • #1205: dpctl.tensor.sqrt
    • #1209: implements out keyword argument
    • #1211: dpctl.tensor.multiply, dpctl.tensor.subtract
    • #1214: dpctl.tensor.not_equal
    • #1216: dpctl.tensor.exp, dpctl.tensor.sin
    • #1217: dpctl.tensor.real, dpctl.tensor.imag, dpctl.tensor.proj
    • #1218: dpctl.tensor.log, dpctl.tensor.log1p, dpctl.tensor.expm1
    • #1221: dpctl.tensor.floor_divide
    • #1235: dpctl.tensor.less
    • #1237: in-place support for addition, multiplication and subtraction
  • Added dpctl.tensor.all and dpctl.tensor.any #1204
  • Added dpctl.tensor.sum #1210

Changed

  • Updated examples of native Python extensions built using dpctl #1108
  • Used security flags to compile and link native extensions of dpctl #1109
  • Changed types of dpctl.tensor.finfo and dpctl.tensor.iinfo output structure per array API spec #1110
  • Consolidated multiple USM temporaries life-time management host_tasks to improve test suite stability #1111
  • MAINT: Improved cmake target dependency tracking #1112
  • MAINT: Improved docstrings for existing dpctl.tensor functions #1123
  • Changed default value of mode keyword in dpctl.tensor.take and dpctl.take.put from clip to wrap #1132
  • Added support for (nested) sequence of dpctl.tensor.usm_ndarray objects in dpctl.tensor.asarray #1139
  • Improved exception handling in dpctl.tensor.usm_ndarray.__setitem__ special method #1146
  • Simplified implementation of copy-and-cast kernels and removed special casing for 2D arrays to conserve binary size #1165
  • Improved speed of dpctl.tensor.usm_ndarray printing functionality #1187
  • Require DPC++ RT 2023.1 to build and run dpctl #1195
  • Compile offloading native extensions with -fno-sycl-id-queries-fit-in-int fixing gh-1184, #1200
  • Transition to conda-forge ecosystem #1213

Fixed

  • Fix to add empty values check for dpctl.tensor.place #1105, #1106
  • Fixed gh-1089 by improving dpctl.tensor.asarray handling of NumPy arrays viewing into host-accessible USM allocation objects.
  • MAINT: Fixed build break with newer GCC and SYCLOS #1118
  • Fixed a bug in basic indexing of dpctl.tensor.usm_ndarray #1136

[0.14.2] - 03/07/2023

Fixed

  • Fixed a bug with boolean advanced indexing #1103

[0.14.1] - 03/06/2023

Added

  • Added dpctl.SyclDevice.partition_max_sub_devices property #1005
  • Added dpctl.program.SyclKernel.max_sub_group_size property #1028
  • Implemented printing of usm_ndarray #1013, #1043, #1060
  • Implemented support for advanced indexing for dpctl.tensor.usm_ndarray #1095, #1097, #1099, #1101
  • Implemented support for platform listing in dpctl.__main__ script #1014
  • Improved performance of dpctl.tensor.asnumpy #1026
  • Added UsmNDArray_Make* C-API for constructing dpctl.tensor.usm_ndarray from native allocations #1050, #1067
  • Added support for dpctl.SyclDevice.native_vector_width_* device descriptors #1075
  • Added dpctl::tensor::usm_ndarray::get_shape_vector and dpctl::tensor::usm_ndarray::get_strides_vector methods #1090

Changed

  • Removed dpctl.select_host_device, dpctl.has_host_device, dpctl.SyclDevice.is_host, and dpctl.SyclDevice.has_aspect_host since support for host device has been removed in DPC++ 2023 and from SYCL 2020 spec #1028

  • usm_ndarrayis made writable by default #1012, and writable flag is now checked by __setitem__.

  • Added convenience signature for C++ utility function in "dpctl4pybind11.hpp" #1016

  • Improved error reported when attempting to submit kernel that uses a data-type unsupported by target device #1018, #1040

  • Updated C++ code to require DPC++ 2023.0.0 or newer #1028, #1066

  • The dpctl.tensor.Device class supports print_device_info method #1029, equality comparison, and hashing #1048

  • Updated version of pybind11 used to 2.10.2 #1031

  • Improved internal utility responsible for reduction of iteration space dimensionality #1044, #1054

  • Changed return type of DCPCTLUSM_GetPointerType function in SyclInterface library #1061, #1065

  • Updated supported version of DLPack to 0.8 #1073

  • Implemented queue cache per context/device pair and deployed it in dpctl.memory, dpctl.tensor.from_dlpack and dpctl.tensor array creation functions #1076, #1079

  • Maintainance, CI work: #1001, #1009, #1011, #1024, #1030, #1032, #1035, #1037, #1039, #1041, #1045, #1047, #1055, #1057, #1059, #1068, #1070, #1074, #1077, #1078, #1081, #1084, #1085, #1088, #1086, #1092, #1093

Fixed

  • Fixed error gh-998 in forming Python exception, #999.
  • A small memory leak fixed, #1000
  • Improved dtype support in dpctl.tensor.full, PR #1002
  • Added missing header file #1008 fixing gh-1007
  • Fixed a typo in device-specific dtype mapping #1015
  • Fixed default device integer type to align with NumPy's behavior on Windows #1017
  • Fixed unexpected overflow in dpctl.tensor.linspace when one of the parameters is the largest floating point value #1034
  • Constructors dpctl.tensor.empty, dpctl.tensor.zeros, and usm_ndarray constructor itself no longer allow to create array with data-types not supported by targeted device #1042
  • Fixed parameter validation in dpctl.SyclQueue constructor #1052
  • Fixed usm_type of the resulting array in dpctl.tensor.tril and dpctl.tensor.triu functions #1062
  • Used DPC++ configuration files to ensure correct use of conda compiler toolchain on Linux #1072
  • Fixed issue with empty argument of dpctl.tensor.meshgrid function #1080
  • Fixed linking problem on Windows enabling dpctl to be functional on Windows for devices not supporting some data types #1083

[0.14.0] - 11/18/2022

Added

  • Implemented dpctl.tensor.linspace function from array-API #875.
  • Implemented dpctl.tensor.eye function from array-API #896.
  • Implemented dpctl.tensor.tril and dpctl.tensor.triu functions from array-API #910.
  • Added data type objects to dpctl.tensor namespace, finfo, iinfo, can_cast, and result_type functions #913.
  • Implemented dpctl.tensor.meshgrid creation function from array-API #920.
  • Implemented convenience class to represent output of dpctl.tensor.usm_ndarray.flags property #921.
  • Added new device attributes and kernel's device-specific attributes #894.
  • Added dpctl.utils.onetrace_enabled context manager for targeted trace collection #903.
  • Added support for stream keyword in __dlpack__ method, enabling support for sending usm_ndarray using mpi4py #906.
  • dpctl.tensor.asarray can now transition data between incompatible devices, #951.
  • Introduced "syclinterface/dpctl_sycl_types_casters.hpp" header file with declaration of conversion routines between SYCL type pointers and SyclInterface library opaque pointers #960.
  • Added C-API to dpctl.program.SyclKernel and dpctl.program.SyclProgram. Added type casters for new types to "dpctl4pybind11" and added an example demonstrating its use #970.
  • Introduced "dpctl/sycl.pxd" Cython declaration file to streamline use of SYCL functions from Cython, and added an example demonstrating its use #981.
  • Added experimental support for sharing data allocated on sub-devices via dlpack #984.
  • Added dpctl.SyclDevice.sub_group_sizes property to retrieve supported sizes of sub-group by the device #985.

Changed

  • Improved queue compatibility testing in dpctl.tensor's implementation module #900.
  • Added automatic measurement of array-API conformance test suite in CI #901.
  • Improved performance of array metadata transfer from host to device #912.
  • Used os.add_dll_directory on Windows to ensure that DPCTLSyclInterface library can be found #918.
  • Refactored dpctl.tensor's implementation module #941 to streamline adding new functionality. Streamlined dpctl::tensor::usm_ndarray class implementation.
  • Added debugging messaging in case when DPCTLDynamicLib::getSymbol encounters errors #956.
  • Updated code base according to changes in DPC++ compiler #952, #957, #958.
  • Changed dpctl to use pybind11 2.10.1 #967.
  • Extended dpctl.tensor.full to accept 0d and higher dimensional arrays for fill-value parameter #982 and #995.

Fixed

  • Improved SyclDevice constructor error message #893.
  • Fixed issue gh-890 about dpctl.tensor.reshape function #915.
  • Fixed unexpected UnboundLocalError exception in #922.
  • Fixed bugs in dpctl.tensor.arange in #945.
  • Fixed issue with type inferencing in dpctl.tensor.asarray in #949.
  • Added missing docstrings for dpctl.SyclDevice properties #964.

[0.13.0] - 07/28/2022

Added

  • Implemented and deployed dedicated kernels for copying with casting #781, used in __setitem__, implementaion of asarray, dpctl.tensor.copy functions.

  • Implemented dedicated copying kernel for dpctl.tensor.reshape function #810, added support for copy keyword #807.

  • Implemented dedicated kernel to copy with casting from numpy.ndarray into dpctl.tensor.usm_ndarray #817.

  • Implemented dpctl.tensor.permute_dims function from array-API #787.

  • Implemented dpctl.tensor.expand_dims function from array-API #788.

  • Implemented dpctl.tensor.squeeze function from array-API #790.

  • Implemented dpctl.tensor.broadcast_to function from array-API #791.

  • Implemented dpctl.tensor.broadcast_arrays function from array-API #798.

  • Implemented dpctl.tensor.flip function from array-API #801.

  • Implemented dpctl.tensor.usm_ndarray.mT property per array-API #805.

  • Implemented dpctl.tensor.roll function from array-API #809.

  • Implemented dpctl.tensor.arange function from array-API #814.

  • Implemented dpctl.tensor.zeros function from array-API #816.

  • Implemented dpctl.tensor.zeros function from array-API #816.

  • Implemented dpctl.tensor.ones, dpctl.tensor.full, dpctl.tensor.empty_like, dpctl.tensor.zeros_like, dpctl.tensor.ones_like, dpctl.tensor.full_like functions from array-API #822.

  • Implemented DPCTLQueue_Memset function in SyclInterface library #812, and exposed it for dpctl.memory.MemoryUSM* classes #815.

  • Implemented dpctl.utils.get_coerced_usm_type to deduced usm type of the output array from types of input arrays in compute-follows-data execution model #797.

  • Added dpctl.SyclDevice.profiling_timer_resolution property #825.

  • Added dpctl.SyclDevice.platform and dpctl.SyclPlatform.default_context properties #827.

  • Provided pybind11 example for functions working on dpctl.tensor.usm_ndarray container applying oneMKL functions #780, #793, #819. The example was expanded to demonstrate implementing iterative linear solvers (Chebyshev solver, and Conjugate-Gradient solver) by asynchronously submitting individual SYCL kernels from Python #821, #833, #838.

  • Wrote manual page about working with dpctl.SyclQueue #829.

  • Added cmake scripts to dpctl package layout and a way to query the location #853.

  • Implemented dpctl.tensor.concat function from array-API #867.

  • Implemented dpctl.tensor.stack function from array-API #872.

Changed

  • Enhanced coverage collection for SyclInterface library by also collecting it during pytest run and combining traces with those collected during C-test run #818. This change also allows to not rebuild SyclInterface library when building C-test executable.
  • Exported keep_args_alive utility in dpctl4pybind11.hpp header #820. The utility uses sycl::handler::host_task to keep given Python arguments alive until eac sycl::event from the given vector of events is complete. The host task is scheduled on the SYCL queue provided as the first argument.
  • Changed the size of struct underlying dpctl.SyclEvent to avoid storing Python object previously used to keep kernel arguments scheduled with dpctl.SyclQueue.submit #823.
  • Fixed docstring for dpctl.SyclTimer #824.
  • Changed type of exceptions raised on failure to create dpctl.SyclDevice from ValueError to dpctl.SyclDeviceCreationError #826.
  • Improved performance of pybind11 type casters #837.
  • Changed implementation of dpctl.SyclProgram from using deprecated sycl::program to sycl::kernel_bundle #845.
  • Removed deprecated device aspects, added new supported aspects #844.
  • Updated vendored dlpack.h to version 0.7 #847.

Fixed

  • Fixed dpctl.lsplatform() to work correctly when used from within Jupyter notebook #800.
  • Fixed script to drive debug build #835 and fixed code to compile in debug mode #836.
  • Fixed filter selector string produced in outputs of dpctl.lsplatform(verbosity=2) and dpctl.SyclDevice.print_device_info #866.
  • Fixed issue with slicing reported in gh-870 in #871.

[0.12.0] - 03/01/2022

Added

  • Properties added to MemoryUSM* objects. #647
  • Added dpctl.tensor.asarray #646
  • Implemented DLPack support for usm_ndarray #682
  • Exported dpctl.tensor.Device class #708 #718
  • Added testing of examples in CI #722
  • Added user manuals to dpctl documentation #712 #773

Changed

  • Folder dpctl-capi/ renamed to libsyclinterface/ in sources and documentation. #666 #768
  • Added workflow to publish rendered documentation on PRs #673 #753 #726
  • Synchronization functions and USM allocation functions release GIL #736 #766
  • dpctl.SyclEvent destructor is made non-blocking #751

Fixed

  • Fixed for issue in code of dpctl.tensor.usm_ndarray.T #653
  • Fixed issue with dpctl.tensor.reshape's affect on contiguity flags of usm_ndarray #695
  • Fixed handling of empty list by dpctl.tensor.asarray #694
  • Fixed type inference with array of empty arrays in dpctl.tensor.asarray #697
  • Fixed issue gh-698 with dpctl.tensr.asarray #709
  • Fixed performance of item assignment from numpy array #724
  • DPCTLDeviceMgr_GetNumDevices should not operate on rejected devices #737
  • Fixed issue gh-729 for dpctl.tensor.reshape applied to 0-element usm_ndarray #756
  • Fixed issue gh-728 with dpctl.tensor.astype #757
  • Fixed type in memory overlapping test #770
  • Fixed issue with operator.pos for dpctl.tensor.usm_ndarray #783
  • Only call PyThread_Ensure from host_task if the main-thread interpreter is initialized and not finalizing #776 #778 #721

Full Changelog: https://github.com/IntelPython/dpctl/compare/0.11.4...0.12.0

[0.11.4] - 12/03/2021

Fixed

  • Fix tests for nested context factories expecting for integration environment by @PokhodenkoSA in IntelPython#705

[0.11.3] - 11/30/2021

Fixed

  • Set the last byte in allocated char array to zero [cherry picked from #650] #699

[0.11.2] - 11/29/2021

Added

  • Extending dpctl.device_context with nested contexts #678

Fixed

  • Fixed issue #649 about incorrect behavior of .T method on sliced arrays #653

[0.11.1] - 11/10/2021

Changed

  • Replaced uses of clang compiler with icx executable #665

[0.11.0] - 11/01/2021

Added

  • Use Python 3.9 in public CI #599
  • Add a new C API utility function (DPCTLDeviceMgr_GetDeviceInfoStr) to return the device info as a C string object #620
  • New Github workflow to build dpclt with nightly Intel llvm/sycl + drivers #621
  • Always raise SubDeviceCreationError even when sub-device counts are zero #622
  • Updated OpenCL interoprability code to fix build with Intel llvm/sycl bundle #625
  • Enabled use of default platform context extension in SYCL compilers that implement this extension #627
  • Implemented dpctl.utils.get_execution_queue(queue_seq) utility to help implementing "compute-follows data" convention for offload target #632 #631

Changed

  • Replaced host_device device type with host in tests #616
  • Rework the logic in dpctl.memory's copy_from_device method to work correctly with host device #618
  • Use dpctl.device_type.host instead of dpctl.device_type.host_device #626
  • Reinstate deprecated sycl::program and that was conditionally removed from open source DPC++ toolchain #633
  • Use LoadLibraryExA instead of LoadLibraryA to mitigate a possible DLL injection issue when we load the Level zero DLL on windows #636
  • Github coverage workflow is changed to use oneAPI 2021.3 instead of latest to work around broken profiling instrumentation in DPC++ 2021.4 #614
  • Update build dependencies for NumPy #641
  • Use "readelf" on SYCL's pi_level_zero library to find out and use the exact name of ze_loader.so in SyclInterface library #617

Removed

  • Removed use of DPC++ features deprecated in 2021.4 and open source Intel llvm/sycl compiler #603

Fixed

  • Suppress errant CMake log #610
  • Fixes to compile dpctl using Intel llvm/sycl compiler #603
  • Fix for the hang is to avoid passing nullptr argument to sycl::queue::prefetch #612
  • Fixed the logic to return device count #623
  • Enabled building of C extensions with dpctl by including header defining bool type for C compilers #604

[0.10.0] - 09/28/2021

Added

  • Added methods bool, float, int, index, and complex to usm_ndarray #578
  • Added data-API required special methods to usm_ndarray class, as well as to_numpy/from_numpy, astype, reshape functions #586
  • Added methods to query dpctl.SyclDevice for size of global/local memory #589
  • Added tests for constructors with invalid capsules #577
  • Improved test coverage of dpctl.SyclQueue implementation #574
  • Added a test to exercise API exported function (get_event_ref). #570
  • Expanded tests in test_sycl_context to improve coverage #571
  • Tweaks to test_sycl_event to improve coverage #567
  • Improved coverage of dpctl.init file and other service functions #563
  • Added test for repr and test for default argument to constructor #565
  • Added some tests to involve capsule #564
  • Added workflow for Public CI on Windows #534
  • DPCTLQueue_Memcpy, _Prefetch, _Memadvise become asynchronous #557
  • Added device aspect selector, dpctl.select_device_with_aspects #558
  • Added test based on example from #583

Changed

  • Parametrized tests for executing OpenCL kernels compiled from source in types of arguments #581
  • Temporary disabled self-hosted CI jobs runner #559
  • Changed static method SyclQueue._create_from_context_and_device #579
  • Transitioned all Python API to use pytest over unittest, improved coverage in dpctl/memory #575
  • Changed dpctl.SyclEvent.profiling_info_submit from method to a property #573
  • Simplified arg parsing in SyclDevice constructor #572
  • Used tag with alignment attribute set in README #562
  • Moved sycl timer into dpctl.SyclTimer #555
  • Used clang-format off, clang-format on to avoid include reordering in pybind11 example #588

Fixed

  • Implemented a workaround for running conda-build using Klocwork #566
  • Separated pipelines for Linux and Windows #582
  • Fixed inconsistency in __sycl_usm_array_interface__ of usm_ndarray instance #584
  • Fixed memory leak: Capsule deleters now free resources for renamed capsules too #568
  • Fixed version test to allow for semantic versioning #569
  • Improved coverage of _types.pxi #556
  • Fixed UnboundLocalError when default queue could not be created #554

[0.9.0] - 08/25/2021

Added

  • Improvements to logic for working with custom DPC++ toolchain #481
  • Add SyclContext unit test cases #488
  • Consolidate configurations of tools that support PEP 518 into pyproject.toml #486
  • Added C-API hash function, used them in Python interface #491
  • Add missing extra checks to ensure unwrapped pointer is not Null
  • Add error messages to L0 program creation routine
  • Improve test coverage for dpctl_sycl_queue_interface #492
  • Use pytest.warns in test_lsplatform3 #495
  • Added test class to test DRef=nullptr case #496
  • Extend parameterized test in test_sycl_queue_interface #497
  • Use Memcpy, memadvise in tests
  • Expanded types tests by TestQueueSubmitRange
  • Added a test that retrieved DPCPP compiled kernel and submits them via DPCTLQueue_SubmitRange #499 , DPCTLEvent_GetCommandExecutionStatus #516, , DPCTLEvent_GetWaitList #510 functions
  • Propagate compile flags #512
  • Add conda package CI pipeline on GitHub Actions #515
  • Run tests on GPU #518
  • Add 3 wrapper func for event::get_profiling_info #519
  • Changes to build_backend.py to enable sycl-compiler-prefix on Windows
  • dtype keyword of usm_ndarray now supports np.double and other types #526
  • Implemented DPCTLQueue_SubmitBarrier, DPCTLQueue_SubmitBarrierForEvents, SyclQueue.submit_barrier #524
  • Added C-API DPCTLQueue_HasEnableProfiling
  • Added Python API SyclQueue.has_enable_profiling
  • Use public for data owning class definitions
  • Queue has enable profiling #531
  • Use public for data owning class definitions #533
  • Added logic to verify that all bits of property integer were recognized and used #494
  • Added support for some properties/methods of underluing device
  • A test for properties, method of q mirroring that of device
  • Conda build scripts should build wheels in the same setup invocation as install #538
  • Added install_requires keyword to setup call
  • Added requirements.txt files in dpctl/ and in dpctl/docs #540
  • Improved C-API for dpctl Cython classes, added example of using them in Pybind11 extension. #550
  • dpctl.SyclEvent acquired ability to get command status and get profiling information. #553

Changed

  • Moved DPCLSyclInterface library from MANIFEST.in #482
  • Refactored tests
  • Use dpcpp compiler package for Linux #514
  • Update conda-package.yml
  • Static methods _init_helper made into functions and removed from PXD files #532

Removed

  • Remove imports from future #485

Fixed

  • Fix sub devices #479
  • Fix addressof_ref function in SyclContext #488
  • Follow DPCTLDevice_CreateFromSelector which passes the check #487
  • Fix a typo in the pytest configuration #490
  • Fixed dbg_build.sh script for Linux to use L0
  • Reuse IntelSycl_LIBRARY_DIR variable in cmake
  • CXX, dpcpp used on Windows too
  • Update conda-recipe/bld.bat
  • Change to SyclQueue.repr to reflect properties #531
  • Static methods _init_helper made into functions and removed from PXD files #532
  • Fixed typo in pip installation instruction #536
  • Fixed dpctl_config.h, added dpctl_service.h, .cpp #539
  • Fixed __sycl_usm_array_interface__ output for 0d arrays #547

[0.8.0] - 05/26/2021

Added

  • Implemented support for constructing MemoryUSM* from object with sycl_usm_array_interface when array-info is not contiguous #400
  • Print the backend as part of SyclDevice.print_device_info function #409
  • Added dpctl/tensor/_usmarray submodule #427
  • Added arg checking to functions in dpctl_sycl_usm_interface.cpp #430
  • A static method of _Memory to create from external allocation #430
  • Added usm_ndarray accessors #435
  • Added Device class representing Data-API notion of device #440
  • Added free Python function as_usm_memory(obj) #443 and associated unit tests #449
  • Dependency for numpy 1.17 #445
  • Add a flag to make doxygen HTML generation optional #450
  • Added a feature to get the filter string for a device from Python using the new dpctl.SyclDevice.get_filter_string method. Also added the corresponding DPCTLDeviceMgr_GetPositionInDevices(DRef, device_mask) C API function #453
  • New options to setup.py to specify which dpcpp compiler to use, if L0 program creation is to be supported, and to generate code coverage #426
  • Github action to check Python code quality #422
  • Github action to auto-publish Sphinx docs for master #446
  • Github action to generate coverage report and publish to coveralls.io #459

Changed

  • Rename dpctl.dptensor to dpctl.tensor #407
  • Changed repr for Memory objects #442
  • Used dpctl.SyclQueue instead of manager and get current queue in tests for SyclProgram #448

Fixed

  • Issue #189 dpctl.memory.MemoryUSMShared(np.int64(16)) should work #392
  • Use size_t instead of Py_ssize_t to fit device USM pointer #405
  • Various code quality issues identified by flake8 (#417, #419, #420, #422)
  • Fixed issues in slicing and array construction #441
  • Fixed an issue #447 where dpctl.get_devices does not return devices in the same order as sycl::device::get_devices #451
  • L0 program creation support on Windows #319

Removed

  • Removing public keyword to get_current_queue Cython declaration #437

[0.7.0] - 05/03/2021

Added

  • Complete support for sycl::ONEAPI::filter_selector in dpctl. , and sycl::platform #298 creation using opaque pointers.
  • A DPCTLDeviceMgr module in C API that caches a default context for root devices #277.
  • DPCTLSyclBackendType and DPCTLSyclDeviceType have a new member ALL #287.
  • C API now provides helper functions to convert between dpctl and SYCL enum values #296.
  • Macros to help create opaque vector classes for opaque SYCL types #297. , SyclContext #334, SyclPlatform (#336, #298), SyclQueue #323 have constructors that recognize filter selectors and closely follow DPC++ interface.
  • Add API to get a PyCapsule from SyclQueue, SyclContext instances #350.
  • Added get_queue_ref_from_ptr_and_syclobj(ptr, syclobj) that creates DPCTLSyclQueueRef from a USM pointer and Python object syclobj from __sycl_usm_array_interface__ #380.
  • Support for SYCL sub-devices, including sub-device creation, queue, and context creation using sub-devices #343.
  • SyclDevice.parent_device property to indicate if an instance has a parent device #366.
  • Several new getter functions for device info descriptors to device interface (#300, #335, #318, #315, #308).
  • Support for SYCL device aspects #307.
  • Properties for every sycl::device info and aspect that we support in SyclDevice #324.
  • Support handling async errors inside SylQueue instances #346.
  • get_backend, get_platform, get_device_type to Python SyclDevice class #300
  • A _sycl_device_factory.pyx module providing SyclDevice constructors using standard sycl::device_selector classes (previously in _sycl_device.pyx) and a new get_devices #277 function to enumerate all devices.
  • _sycl_device_factory.pyx implements get_num_devices and has_*_device(s) functions #320.
  • Enable Python coverage in CI for Linux #369.
  • Use public keyword in _sycl_*.pxd to generate header files allowing non-Cython centric native extensions to work with dpctl's Python objects #218.
  • Documentation improvements #341.

Changed

  • Rename dpCtl to dpctl in all comments, license headers, and docs. #342
  • dpctl.memory.MemoryUSM* constructors now use dpctl.SyclQueue() instead of dpctl.get_current_queue() when the queue keyword argument is None (default) #382.
  • dpctl.set_default_queue has been renamed to dpctl.set_global_queue() #323.
  • Changed dpctl.dump to dpctl.lsplatform #336.
  • Various SyclDevice methods related to querying sycl::info::device were converted to properties #324.
  • Various C API functions names were changed.

Fixed

  • Possible crashes when a SYCL platform is not available #349.
  • Fix tests which fail if GPU is not available (only CPU is available) #359.
  • Fix breaking C API tests #358.
  • Bandit warning about "subprocess.check_call(shell=True)" for Windows #306.

Removed

  • Removed get_num_platforms, has_cpu_queues, has_gpu_queues, get_num_queues, has_sycl_platforms #320.

[0.6.1] - 2021-03-01

Fixed

  • Do not use POP_FRONT in FindDPCPP.cmake so that we can use a cmake version older that 3.15.

[0.6.0] - 2021-03-01

Added

  • Documentation improvements.
  • Cmake improvements and Coverage for C API, Cython and Python.
  • Added support for Level Zero devices and queues.
  • Added support for SYCL standard device_selector classes.
  • SyclDevice instances can now be constructed using filter selector strings.
  • Code of conduct.
  • Building wheels.
  • Queue manager improvements.
  • Adding __array_function__ so that Numpy calls with dparrays work.
  • Using clang-format for C/C++ code formatting.
  • Using pytest for running tests.
  • Add python and cython file coverage.
  • Using Bandit for finding common security issues in Python code.
  • Add instructions about file headers formats.

Changed

  • Changed compiler name usage from clang++ to dpcpp.
  • Reformat backend.pxd to be closer to black style.

Fixed

  • Remove cython from install_requires. It allows use dpCtl in numba extensions.
  • Incorrect import in example.
  • Consistency of file headers.
  • Klocwork issues.

[0.5.0] - 2020-12-17

Added

  • _Memory.get_pointer_type static method which returns kind of USM pointer.
  • Utility functions to transform string to device type and back.
  • New dpctl.dptensor.numpy_usm_shared module containing USM array. USM array extends NumPy ndarray.
  • A lot of new examples. Including examples of building Cython extensions with DPC++ compiler that interoperate with dpCtl.
  • Mechanism for registering a callback function to look and see if the object supports USM.

Changed

  • setup.py builds C++ backend for develop and install commands.
  • Building wheels.
  • Use DPC++ runtime from package dpcpp_cpp_rt.
  • All usage of DPPL in C-API functions was changed to DPCTL, e.g., DPPLQueueMgr_GetCurrentQueue to DPCTLQueueMgr_GetCurrentQueue.
  • Renamed the C-API directory is now called dpctl-capi instead of backends.
  • Refactoring the dpctl-capi functions to prepare for changes to add Level Zero program creation.
  • SyclProgram and SyclKernel classes were moved out of dpctl into the dpctl.program sub-module.

Fixed

  • Klockwork static code analysis warnings.

[0.4.0] - 2020-11-04

Added

  • Device descriptors "max_compute_units", "max_work_item_dimensions", "max_work_item_sizes", "max_work_group_size", "max_num_sub_groups" and "aspects" for int64 atomics inside dpctl C API and inside the dpctl.SyclDevice class.
  • MemoryUSM* classes moved to dpctl.memory module, added support for aligned allocation, added support for prefetch and mem_advise (sychronous) methods, implemented copy_to_host, copy_from_host and copy_from_device methods, pickling support, and zero-copy interoperability with Python objects which implement __sycl_usm_array_inerface__ protocol.
  • Helper scripts to generate API documentation for both C API and Python.

Fixed

  • Compiler warnings when building libDPPLSyclInterface and the Cython extensions.

Removed

  • The Legacy OpenCL interface.

[0.3.8] - 2020-10-08

Changed

  • How the initial active queue is populated inside DPPLQueueMgr.
  • dpctl.SyclQueueManager only reports the number of non-host platform.
  • dpctl.SyclQueueManager now raises an exception if DPCTL C API returns a nullptr instead of a valid Sycl queue.

Fixed

  • Several crashes in cases where an OpenCL or Level Zero platform is not available.
  • Fix failing platform test case. #116
  • Properly skip tests when no OpenCL devices are available.
  • Add skip tests to test_sycl_usm.py
  • Fix Gtests configuration.

[0.3.7] - 2020-10-08

Fixed

  • A crash on Windows due a Level Zero driver problem. Each device was getting enumerated twice. To handle the issue, we added a temporary fix to use only first device for each device type and backend #118.

[0.3.6] - 2020-10-06

Added

  • Changelog was added for dpctl.

Fixed

  • Windows build was fixed.

[0.3.5] - 2020-10-06

Added

  • Add a helper function to all Python SyclXXX classes to get the address of the base C API pointer as a long.

Changed

  • Rename PyDPPL to dpCtl in comments (function name renaming to come later)

Fixed

  • Fix bugs highlighted by tools.
  • Various code clean ups.

[0.3.4] - 2020-10-05

Added

  • Dump functions were enhanced to print back-end information.
  • dpctl gained support for unint_8 and unsigned long data types.
  • oneAPI Beta 10 tool chain support was added.

Changed

  • dpctl is now aware of DPC++ Sycl PI back-ends. The functionality is now exposed via the context interface.
  • C API's queue manager was refactored to require back-end.
  • dpct's device_context now requires back-end, device-type, and device-id to be provided in a string format, e.g. opencl:gpu:0.

Fixed

  • Fixed some important bugs found by static analysis.

[0.3.3] - 2020-10-02

Added

  • Add dpctl.get_curent_device_type().

[0.3.2] - 2020-09-29

Changed

  • Set _cpu_device and _gpu_device to None by default.

[0.3.1] - 2020-09-28

Added

  • Add get include and include headers.

Changed

  • DPPL shared objects are installed into dpctl.

Fixed

  • Refactor unit tests.

[0.3.0] - 2020-09-23

Added

  • Adds C and Cython API for portions of Sycl queue, device, context interfaces.
  • Implementing USM memory management.

Changed

  • Refactored API to expose a minimal sycl::queue interface.
  • Modify cpu_queues, gpu_queues and active_queues to functions.
  • Change static vectors to static pointers to verctors. It disables call for destructors. Destructors are also call in undefined order.
  • Rename package PyDPPL to dpCtl.
  • Use dpcpp.exe on Windows instead of dpcpp-cl.exe deleted in oneAPI beta08.

Fixed

  • Correct use ERRORLEVEL in conda scripts for Windows.
  • Fix using dppl.has_sycl_platforms() and dppl.has_gpu_queues() functions in skipIf