1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
|
from __future__ import annotations
import platform
import numpy as np
import pytest
from numpy.testing import assert_array_equal
import boost_histogram as bh
STORAGES = (bh.storage.Int64, bh.storage.Double, bh.storage.Unlimited)
DTYPES = (np.float64, np.float32, np.int64, np.int32)
# Casting is broken for numpy on ppc64le: https://github.com/numpy/numpy/issues/21062
if platform.machine() == "ppc64le":
DTYPES = (np.float64, np.int64)
bins = (100, 100)
ranges = ((-1, 1), (-1, 1))
bins = np.asarray(bins).astype(np.int64)
ranges = np.asarray(ranges).astype(np.float64)
edges = (
np.linspace(ranges[0, 0], ranges[0, 1], bins[0] + 1),
np.linspace(ranges[1, 0], ranges[1, 1], bins[1] + 1),
)
np.random.seed(42)
vals_core = np.random.normal(size=[2, 100000])
vals = {t: vals_core.astype(t) for t in DTYPES}
answer = {t: np.histogram2d(*vals[t], bins=bins, range=ranges)[0] for t in DTYPES}
@pytest.mark.benchmark(group="2d-fills")
@pytest.mark.parametrize("dtype", vals)
def test_numpy_perf_2d(benchmark, dtype):
result, _, _ = benchmark(np.histogram2d, *vals[dtype], bins=bins, range=ranges)
assert_array_equal(result, answer[dtype])
def make_and_run_hist(flow, storage, vals):
histo = bh.Histogram(
bh.axis.Regular(bins[0], *ranges[0], underflow=flow, overflow=flow),
bh.axis.Regular(bins[1], *ranges[1], underflow=flow, overflow=flow),
storage=storage(),
)
histo.fill(*vals)
return histo.view()
@pytest.mark.benchmark(group="2d-fills")
@pytest.mark.parametrize("dtype", vals)
@pytest.mark.parametrize("storage", STORAGES)
def test_2d(benchmark, flow, storage, dtype):
result = benchmark(make_and_run_hist, flow, storage, vals[dtype])
assert_array_equal(result[:-1, :-1], answer[dtype][:-1, :-1])
|
