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.
- Documentation.
- Add support for dpctl.dparray.
- Support NumPy functions via DPNP: random, linalg, transcendental, array ops, array creation.
- Wheels building.
- Using Bandit for finding common security issues in Python code.
- Start using black code style formatter.
- Build SPIRV code in setup.py.
- Start using pytest for running tests.
- Start using Apache 2.0 license.
- Consistency of file headers.
- Updated to Numba 0.52, dpCtl 0.6 and dpNP 0.5.1.
- Don't create a new copy of a usm shared array data pointers for kernel call.
- Modify test cases and examples to use Level Zero queue.
- Fix incorrect import in examples.
- numba-dppy is a standalone package now. Added setup.py and conda recipe.
- Offload diagnostics.
- Controllable fallback.
- Add flags to generate debug symbols.
- Implementation of
np.linalg.eig
,np.ndarray.sum
,np.ndarray.max
,np.ndarray.min
,np.ndarray.mean
. - Two new re-write passes to convert NumPy calls into a pseudo
numba_dppy
call site to allow target-specific overload of NumPy functions. The rewrite passes is a temporary fix till Numba gains support for target-specific overlaods. - Updated to dpCtl 0.5.* and dpNP 0.4.*
- The
dpnp
interface now uses Numba's@overload
functionality as opposed to the previous@lower_builtin
method. - Rename
DPPL
toDPPY
. - Cleaned test code.
DPPLTestCase
replaced withunittest.TestCase
.- All tests and examples use
with device_context
. - Config environment variables starts with
NUMBA_DPPY_
(i.e. NUMBA_DPPY_SAVE_IR_FILES and NUMBA_DPPY_SPIRV_VAL) - Remove nested folder
dppl
intests
. - No dependency on
cffi
.
- The old backup file.
This release includes:
- Caching of dppy.kernels which will improve performance.
- Addition of support for Intel Advisor which will help in profiling applications.