Skip to content

Micro-optimizations for the statistics module #152618

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
rhettinger merged 2 commits into from
Jun 29, 2026
+4 −7
Merged

Micro-optimizations for the statistics module #152618

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
0f12359
Remove boundmethod optimization which is no longer performant
rhettinger Jun 29, 2026
2801a30
Improve speed and accuracy for the quartic kernel pdf.
rhettinger Jun 29, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 4 additions & 7 deletions Lib/statistics.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -894,7 +894,7 @@ def _quartic_invcdf_estimate(p):

@register('quartic', 'biweight')
def quartic_kernel():
pdf = lambda t: 15/16 * (1.0 - t * t) ** 2
pdf = lambda t: 15/16 * (u := 1.0 - t * t) * u

@eendebakpt eendebakpt Jun 29, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This does not improve readability. Do you have a bench ark for the gain of this change?

@rhettinger rhettinger Jun 29, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes. The timings were in the individual commit messages. I just moved them to the top message in Conversation.

It is not a beautiful edit, but it isn't terrible either. Note, besides giving a nice speed-up, the reduction in strength from x ** 2 to x * x also improves accuracy.

@eendebakpt eendebakpt Jun 29, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@rhettinger If it improves accuracy, it changes behavior. Can you add a news entry and a test?

(since you already merged the PR, this should be a followup PR)

cdf = lambda t: sumprod((3/16, -5/8, 15/16, 1/2),
(t**5, t**3, t, 1.0))
invcdf = _newton_raphson(_quartic_invcdf_estimate, f=cdf, f_prime=pdf)
Expand Down Expand Up @@ -1486,15 +1486,13 @@ def _sum(data):
"""
count = 0
types = set()
types_add = types.add
partials = {}
partials_get = partials.get

for typ, values in groupby(data, type):
types_add(typ)
types.add(typ)
for n, d in map(_exact_ratio, values):
count += 1
partials[d] = partials_get(d, 0) + n
partials[d] = partials.get(d, 0) + n

if None in partials:
# The sum will be a NAN or INF. We can ignore all the finite
Expand Down Expand Up @@ -1524,12 +1522,11 @@ def _ss(data, c=None):

count = 0
types = set()
types_add = types.add
sx_partials = defaultdict(int)
sxx_partials = defaultdict(int)

for typ, values in groupby(data, type):
types_add(typ)
types.add(typ)
for n, d in map(_exact_ratio, values):
count += 1
sx_partials[d] += n
Expand Down
Loading