Minutes_2019_08_27
Siu Kwan Lam edited this page Oct 14, 2019
·
1 revision
Attendees: Stuart, Pearu, Val, Todd, Siu, Stan, Ehsan, Aaron
- **** #4484 - Nested parallel function will never return and weird concurrency
- TOFIX: try to detect threadpool lockup
- SUGGEST: TBB
- **** #4483 - Incorrect results for np.dot with zero-sized arrays
- What should it do?
-
#4481 - function accelerated by jit can not use lambda
-
sorted
doesn't support key
-
-
#4479 - np.argsort and ndarray.argsort only on 1-D data
- need to look into this
-
#4478 - Support for new const arrays / ndarray.setflags
-
setflags
usage can be problematic
-
-
#4477 - How to express bool arguments to a @guvectorise function?
- no action required
- **** #4473 - Type hints for Numba code base
-
#4471 - Array order check is too strict
- good idea, but requires a lot of changes in implementations of NumPy calls
-
#4470 - Can't create a numpy array from a numpy array
- PR in flight
-
#4469 - Issue using differently shaped NumPy array
- needs more investigation
-
#4466 - Support passing scalars by reference to ctypes/cffi methods
- good suggestion
-
#4465 - Overload with kwargs provides inconsistent type arguments
- need more helpers to make life easier with kwargs
- related to declarative typing
- #4482 - Fix example in guvectorize docstring.
- #4480 - [WIP] Reimplementing np.mean with overload and adding axis parameter
- #4476 - [WIP] Another attempt at fixing frame injection in the dispatcher tracing path
- #4475 - Overload np.array to accept arrays
- **** #4474 - Fix liveness for remove dead of parfors (and other IR extensions)
- #4472 - [WIP] allow dtype input argument in np.sum
- #4468 - Prevent CUDA kernel launch without a specified launch config.
- **** #4467 - [WIP] Literal dispatch
- #4464 - Add a newline in patched errors
- Finish pending rewrite passes
- Python 3.8 support
- Document best practices for constructing new compiler pipelines
- Define autodiscovery system for Numba extensions (like numba_scipy or HPAT) that don't need direct user import
- Allow opt-in dispatching of functions by literal value
- Making caching aware of transitive dependencies
- Define declarative typing system for @overload (to be used in future releases)
- Numba Runtime C API for extensions to register reference-counted memory with the runtime.
- Start using new CI system in parallel with existing one
- Priority bug fixes:
- Low performance of JIT method calls (requested by Pandas devs)
- Others TBD