Graham Markall
Graham Markall
From: https://numba.discourse.group/t/cannot-gettypeinfo-on-a-type-that-is-unsized-error-on-jitclass-with-typed-dict-variable-that-has-deferred-type-as-value-type/1191 ```python import numba as nb from collections import OrderedDict from numba import deferred_type, optional from numba.experimental import jitclass from numba import types node_type = deferred_type() @jitclass([ ('parent', optional(node_type)),...
The following: ```python from numba import njit @njit('int32(int32, int32)') def f(x, y): return x + y f.inspect_types() f.inspect_types(pretty=True) ``` prints, in a terminal: ``` $ python repro.py reducer_override for f...
This warning applied to imports that followed SVML initialization, but now there are none. Related question - do we have a test that ensures that nothing is `@njit`-ed during Numba...
Originally noted in https://numba.discourse.group/t/type-inference-for-dispatching-of-function-arguments-in-njit-with-multiple-signatures/1536/2 When function compilation is disabled (e.g. due to eager typing), checks whether one function typing can cast to another typing of the function cause the whole...
As noted in https://groups.google.com/a/continuum.io/forum/?utm_medium=email&utm_source=footer#!msg/numba-users/S0Vrk2ZfyQg/nNcxcnzzEAAJ it's quite fiddly to set the cache configuration for a kernel - support is exposed in the driver API, but there seems to be no way...
Support for multiple outputs from CUDA gufuncs. To-do list: - [ ] Additional test coverage. - [ ] Fix simulator test failures. - [ ] Fix the use of `_need_device_conversion`...
For debugging / testing #9432, and anything else that happens to be wrong on `main` at present.
Noted whilst investigating https://numba.discourse.group/t/avoid-multiple-copies-of-large-numpy-array-in-closure/2017 In many cases Numba will copy global and closure variables as constant arrays inside jitted functions and kernels (and will always attempt to do this for...
The following: ```python from numba import cuda import numpy as np @cuda.jit('void()') def f(): x = cuda.shared.array(0, dtype=np.int32) x.reshape(2, 2) ``` Gives: ``` Traceback (most recent call last): File "/home/gmarkall/numbadev/numba/numba/cuda/cudadrv/driver.py",...
**EDIT**: All the changes in this PR excepting actually making indices unsigned are in PR #6127 for a separate, simpler review. **Original comments:** The main purpose of this PR is...