gmarkall

import itertools

import numba as nb
from numba.experimental import jitclass
from typing import List, Tuple, Dict
from heapq import heappush, heappop


# @jitclass
class PurePythonPriorityQueue:

gmarkall

import math
from numba import cuda, njit, objmode
from time import perf_counter
import numpy as np

import cupy as cp



@njit

gmarkall

Iterating
Stream <CUDA stream 93950825260944 on <CUDA context c_void_p(93950819544272) of device 0>> done
Stream <CUDA stream 93950825975488 on <CUDA context c_void_p(93950819544272) of device 0>> done

gmarkall

# Works in conjunction with https://github.com/numba/numba/pull/7453

from numba import cuda
import asyncio


async def f():
    s1 = cuda.stream()
    s2 = cuda.stream()

gmarkall

$ PYTHONMALLOC=malloc valgrind-numba python -m numba.runtests numba.tests.test_ufuncs
==7578== Memcheck, a memory error detector
==7578== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==7578== Using Valgrind-3.17.0 and LibVEX; rerun with -h for copyright info
==7578== Command: python -m numba.runtests numba.tests.test_ufuncs
==7578== 
==7579== Warning: invalid file descriptor 1024 in syscall close()
==7579== Warning: invalid file descriptor 1025 in syscall close()
==7579== Warning: invalid file descriptor 1026 in syscall close()
==7579== Warning: invalid file descriptor 1027 in syscall close()

gmarkall

gmarkall

from numba import njit, f8
from numba.typed import List
from numba.extending import models, register_model



class Interval(object):
    """
    A half-open interval on the real number line.
    """

gmarkall

diff --git a/numba/core/extending.py b/numba/core/extending.py
index 9d005fe74..b42442a38 100644
--- a/numba/core/extending.py
+++ b/numba/core/extending.py
@@ -155,8 +155,10 @@ def register_jitable(*args, **kwargs):
     def wrap(fn):
         # It is just a wrapper for @overload
         inline = kwargs.pop('inline', 'never')
+        target = kwargs.pop('target', 'cpu')
 

gmarkall

# Implements unicode equality for the CUDA target

from numba import cuda, types
from numba.core.extending import overload
from numba.core.pythonapi import (PY_UNICODE_1BYTE_KIND,
                                  PY_UNICODE_2BYTE_KIND,
                                  PY_UNICODE_4BYTE_KIND)
from numba.cpython.unicode import deref_uint8, deref_uint16, deref_uint32
import numpy as np
import operator

gmarkall

# Use with https://github.com/gmarkall/numba/tree/cuda-linker-options

from numba import cuda, float32, void


def axpy(r, a, x, y):
    start = cuda.grid(1)
    step = cuda.gridsize(1)

    for i in range(start, len(r), step):