Repository navigation
Inline cache for slots #87093
Description
Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.
Show more details
GitHub fields:
assignee = None
closed_at = <Date 2021-01-30.02:03:05.447>
created_at = <Date 2021-01-14.04:45:09.122>
labels = ['3.10', 'performance']
title = 'Inline cache for slots'
updated_at = <Date 2021-01-31.22:55:58.106>
user = 'https://github.com/gvanrossum'bugs.python.org fields:
activity = <Date 2021-01-31.22:55:58.106>
actor = 'pablogsal'
assignee = 'none'
closed = True
closed_date = <Date 2021-01-30.02:03:05.447>
closer = 'gvanrossum'
components = []
creation = <Date 2021-01-14.04:45:09.122>
creator = 'gvanrossum'
dependencies = []
files = ['49744']
hgrepos = []
issue_num = 42927
keywords = ['patch']
message_count = 15.0
messages = ['385063', '385064', '385066', '385076', '385077', '385078', '385079', '385080', '385081', '385091', '385095', '385103', '385124', '385965', '386046']
nosy_count = 7.0
nosy_names = ['gvanrossum', 'rhettinger', 'methane', 'serhiy.storchaka', 'yselivanov', 'pablogsal', 'BTaskaya']
pr_nums = ['24216', '24383']
priority = 'normal'
resolution = 'fixed'
stage = 'commit review'
status = 'closed'
superseder = None
type = 'performance'
url = 'https://bugs.python.org/issue42927'
versions = ['Python 3.10']
Activity
gvanrossum commentedon Jan 14, 2021
I've been thinking about Python performance improvements, and I played around with an inline cache enhancement that supports slots. The results on a very simple benchmark look promising (30% speedup) but I'm terrible with our benchmarking tools, and this should be considered *very* carefully before we go ahead with it, since it could potentially pessimize the inline cache for standard attributes.
I'll attach a PR in a minute.
serhiy-storchaka commentedon Jan 14, 2021
Slot means so many different things in Python. Here it's about data descriptors created when you set __slots__ in the class definition.
It is amazing that so large speed up can be achieved for such base operation.
pablogsal commentedon Jan 14, 2021
I will try to run today the pyperformance test suite to see if there is any impact on standard attribute access
pablogsal commentedon Jan 14, 2021
These are the results:
venv ❯ python -m pyperf compare_to json_old/* -G --min-speed=2 --table
+---------------------+--------------------------------------+------------------------------------------+
| Benchmark | 2021-01-14_09-40-master-14cfa325c298 | 2021-01-14_10-39-slotscache-9123a84ad8d4 |
+=====================+======================================+==========================================+
| float | 148 ms | 133 ms: 1.11x faster |
+---------------------+--------------------------------------+------------------------------------------+
| pidigits | 265 ms | 252 ms: 1.05x faster |
+---------------------+--------------------------------------+------------------------------------------+
| scimark_monte_carlo | 130 ms | 126 ms: 1.03x faster |
+---------------------+--------------------------------------+------------------------------------------+
| spectral_norm | 168 ms | 164 ms: 1.03x faster |
+---------------------+--------------------------------------+------------------------------------------+
| logging_silent | 221 ns | 216 ns: 1.02x faster |
+---------------------+--------------------------------------+------------------------------------------+
| sympy_expand | 598 ms | 585 ms: 1.02x faster |
+---------------------+--------------------------------------+------------------------------------------+
| regex_v8 | 31.9 ms | 32.6 ms: 1.02x slower |
+---------------------+--------------------------------------+------------------------------------------+
| deltablue | 8.93 ms | 9.16 ms: 1.03x slower |
+---------------------+--------------------------------------+------------------------------------------+
| unpickle | 17.0 us | 17.7 us: 1.04x slower |
+---------------------+--------------------------------------+------------------------------------------+
| Geometric mean | (ref) | 1.00x faster |
+---------------------+--------------------------------------+------------------------------------------+
Benchmark hidden because not significant (50): sympy_sum, sqlalchemy_imperative, sympy_str, sympy_integrate, dulwich_log, scimark_fft, genshi_text, tornado_http, regex_dna, sqlalchemy_declarative, mako, meteor_contest, unpickle_pure_python, xml_etree_parse, genshi_xml, scimark_sor, sqlite_synth, pickle_pure_python, nbody, pickle_dict, pyflate, regex_effbot, xml_etree_process, logging_simple, python_startup, 2to3, fannkuch, python_startup_no_site, raytrace, go, hexiom, scimark_lu, json_dumps, richards, logging_format, xml_etree_generate, chaos, telco, pickle_list, unpack_sequence, regex_compile, django_template, json_loads, crypto_pyaes, xml_etree_iterparse, unpickle_list, nqueens, chameleon, scimark_sparse_mat_mult, pickle
Ignored benchmarks (1) of json_old/2021-01-14_09-40-master-14cfa325c298.json.gz: pathlib
I uploaded the pyperf result data to bpo as well
pablogsal commentedon Jan 14, 2021
So it seems that everything is in the noise range except the "float" benchmark that is 1.11x faster
pablogsal commentedon Jan 14, 2021
Some microbenchmarks:
CURRENT MASTER:
Variable and attribute read access:
4.5 ns read_local
5.8 ns read_nonlocal
7.8 ns read_global
7.8 ns read_builtin
21.2 ns read_classvar_from_class
19.1 ns read_classvar_from_instance
16.4 ns read_instancevar
23.6 ns read_instancevar_slots
20.2 ns read_namedtuple
40.5 ns read_boundmethod
PR 24216
Variable and attribute read access:
4.5 ns read_local
5.8 ns read_nonlocal
7.8 ns read_global
8.2 ns read_builtin
21.4 ns read_classvar_from_class
19.0 ns read_classvar_from_instance
16.6 ns read_instancevar
10.2 ns read_instancevar_slots
20.5 ns read_namedtuple
42.9 ns read_boundmethod
1st1 commentedon Jan 14, 2021
Yeah, this is why. https://github.com/python/pyperformance/blob/master/pyperformance/benchmarks/bm_float.py#L12
This is a great result, IMO. I'm +1 to merge this.
1st1 commentedon Jan 14, 2021
Can you add a new one
read_slots?pablogsal commentedon Jan 14, 2021
Same here. It would be good if Inada-san could confirm the benchmarks.
That should be "read_instancevar_slots". Or you refer to some other check?
rhettinger commentedon Jan 15, 2021
+1 Thanks for this.
gvanrossum commentedon Jan 15, 2021
Thanks for all the positive feedback! What is the next step?
pablogsal commentedon Jan 15, 2021
I would say just finishing the PR (making sure that we do not miss some arcane edge case) and updating the what's new for 3.10 :)
gvanrossum commentedon Jan 15, 2021
I created a benchmark for this, see python/pyperformance#86.
Next I'll add a blurb and then it's off to reviewers.
gvanrossum commentedon Jan 30, 2021
New changeset 5c5a938 by Guido van Rossum in branch 'master':
bpo-42927: Inline cache for attributes defined with '__slots__' (bpo-24216)
5c5a938
pablogsal commentedon Jan 31, 2021
New changeset a776da9 by Pablo Galindo in branch 'master':
bpo-42927: Update the What's new entry for LOAD_ATTR optimizations (GH-24383)
a776da9