Skip to content

Commit a694299

Browse files
committed
Add gh-2335 to the changelog
1 parent 3575f77 commit a694299

1 file changed

Lines changed: 10 additions & 0 deletions

File tree

‎CHANGELOG.md‎

Lines changed: 10 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -6,9 +6,19 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
66

77
## [dev] - XXX. XX, XXXX
88

9+
**NOTE**: This release contains breaking changes:
10+
11+
* The value of the `DPCTL_UNSUPPORTED_KERNEL_ARG` enumerator of `DPCTLKernelArgType` increased by one, as `DPCTL_BYTES16_T` was added before it [gh-2335](https://github.com/IntelPython/dpctl/pull/2335)
12+
913
### Added
14+
* `dpctl.SyclQueue.submit` and `dpctl.SyclQueue.submit_async` accept NumPy scalars of any data type whose values are plain bytes, including structured data types and those of `ml_dtypes` such as `bfloat16` and `complex32`, as well as Python `bool`, `int`, `float`, and `complex` scalars, which are passed as `numpy.bool`, `numpy.int64`, `numpy.float64`, and `numpy.complex128` respectively [gh-2335](https://github.com/IntelPython/dpctl/pull/2335)
15+
* Added `DPCTL_BYTES16_T` enumerator to `DPCTLKernelArgType` for 16-byte scalar and local accessor kernel arguments [gh-2335](https://github.com/IntelPython/dpctl/pull/2335)
16+
* Added a `sycl::ext::oneapi::bfloat16` type caster to `dpctl4pybind11.hpp`, defined when the compiler provides the `sycl_ext_oneapi_bfloat16` extension [gh-2335](https://github.com/IntelPython/dpctl/pull/2335)
1017

1118
### Changed
19+
* `dpctl.SyclQueue.copy`, `dpctl.SyclQueue.copy_async`, `dpctl.SyclQueue.fill`, `dpctl.SyclQueue.fill_async`, `dpctl.LocalAccessor`, `dpctl.WorkGroupMemory`, and `dpctl.compiler.SpecializationConstant` accept a data type in any form `numpy.dtype` accepts, for any data type whose values are plain bytes, rather than a fixed set of type strings [gh-2335](https://github.com/IntelPython/dpctl/pull/2335)
20+
* `dpctl.compiler.SpecializationConstant` accepts NumPy scalars and Python `bool`, `int`, `float`, and `complex` scalars [gh-2335](https://github.com/IntelPython/dpctl/pull/2335)
21+
* Added `DPCTL_BYTES16_T` to `DPCTLKernelArgType` before `DPCTL_UNSUPPORTED_KERNEL_ARG`, which shifts the value of the latter [gh-2335](https://github.com/IntelPython/dpctl/pull/2335)
1222

1323
### Deprecated
1424

0 commit comments

Comments
 (0)