Minutes_2019_08_13
Siu Kwan Lam edited this page Oct 14, 2019
·
1 revision
Attendees: Aaron, Ehsan, Pearu, Stuart, Todd, Stan, Siu, James
- SSA https://github.com/numba/numba/pull/4445
- status on tests
- will array analysis effectiveness be impacted by SSA?
- array analysis is merging vars info from incoming blocks currently. Introducing PHI node will require changing this merges.
- blocks some inlining
- branch-pruning, constant propagation, etc...
- affects
.get_definitions()
- PHI node is preferred
- How to customize LLVM passes at codegen.py?
- Intel LLVM specific passes
- Polly passes
- Adjust threadpool size https://github.com/numba/numba/issues/2713
- does it satisfy the requirement if numba initialized max number of threads but can mask off threads? (James will gather more info)
- #4438 - paralled numpy array error:
- #4437 - Buffer dtype cannot be buffer, have dtype: array(float64, 2d, C)
- #4436 - njitted array getitem accepts out-of-bounds array indices
- **** #4435 - Calling jitted function in unboxing fails
- callconv changes
- having high-level API for (un)boxing is preferred
- likely require a set of unsafe intrinsic for Python CAPI
- #4434 - Adding Jitclass as Identifier for other Jitclasses/function types
- #4430 - remove dead comment in azure script
- #4429 - np.random.random_sample(int64,) with NumPy 1.17/Numba 0.45.1
- #4428 - Add array tuple index from array unsafe function
- #4424 - Compilation/Type Error with opaque pointers in cffi
- **** #4423 - omitted types forcing additional compilation?
- related to value-based specialisation
- #4421 - numba.errors.InternalError: only one advanced index supported
- **** #4420 - slowing down the functions
- need FAQ
- multithreaded numpy vs numba?
- dragracing small functions
- need FAQ
- #4419 - @njit numpy mean incorrect result
- #4417 - 'NoneType' object has no attribute 'module'
- #4416 - Lowering UnicodeCharSeq fails
-
#4414 - TypingError:
np.max(..., axis=1)
-
#4413 - Accessing array fields works via getattr while it should not.
- deprecate first because we allowed it for so long.
- #4422 - TypingError: Failed at nopython (nopython frontend)
- #4433 - Removes stale azure comment
- #4432 - [WIP] Bounds checking
- #4431 - Fix and test numpy.random.random_sample(n) for np117
- **** #4427 - [WIP] Entrypoints for numba extensions
- #4426 - Fixed two module teardown races in py2.
- #4425 - [WIP] Unicode array support
- ** #4418 - [WIP] Directed graph task roadmap
- #4415 - Make 32bit incremental builds on linux not use free channel
- #4412 - Fix #4387. Limit the number of return types for recursive functions
- #4411 - Name of type should not contain user facing description for debug.
- 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