Harmonize STORE_DEREF with STORE_FAST and LOAD_DEREF

[rhettinger]

BPO	28665
Nosy	@rhettinger, @serhiy-storchaka
PRs	[Do Not Merge] Convert Misc/NEWS so that it is managed by towncrier #552
Files	fastcell.diff delete_deref.diff: Fix-up DELETE_DEREF concat_deref.diff: Fix-up unicode_concat issue28665.py

^{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 = 'https://github.com/rhettinger'
closed_at = <Date 2016-11-12.09:12:16.093>
created_at = <Date 2016-11-11.12:10:41.742>
labels = ['interpreter-core', '3.7', 'performance']
title = 'Harmonize STORE_DEREF with STORE_FAST and LOAD_DEREF'
updated_at = <Date 2017-03-31.16:36:26.072>
user = 'https://github.com/rhettinger'

bugs.python.org fields:

activity = <Date 2017-03-31.16:36:26.072>
actor = 'dstufft'
assignee = 'rhettinger'
closed = True
closed_date = <Date 2016-11-12.09:12:16.093>
closer = 'rhettinger'
components = ['Interpreter Core']
creation = <Date 2016-11-11.12:10:41.742>
creator = 'rhettinger'
dependencies = []
files = ['45442', '45454', '45455', '45457']
hgrepos = []
issue_num = 28665
keywords = ['patch']
message_count = 11.0
messages = ['280574', '280576', '280577', '280578', '280584', '280585', '280635', '280642', '280643', '280645', '280646']
nosy_count = 3.0
nosy_names = ['rhettinger', 'python-dev', 'serhiy.storchaka']
pr_nums = ['552']
priority = 'low'
resolution = 'fixed'
stage = 'commit review'
status = 'closed'
superseder = None
type = 'performance'
url = 'https://bugs.python.org/issue28665'
versions = ['Python 3.7']

[rhettinger]

The STORE_FAST, LOAD_FAST, and LOAD_DEREF opcodes all use fast macros for variable access. This patch harmonizes STORE_DEREF to follow the same pattern.

Both the C code and the generated assembly look nicer. Gives an approx 40% speed-up (using both Clang and GCC-6) on the "store_nonlocal" portion of the variable access benchmark at http://code.activestate.com/recipes/577834
The eliminates the nonlocal speed penalty, making cell variable updates run nearly as fast as updates to locals.

[serhiy-storchaka]

LGTM. This saves function call and INCREF/DECREF pair.

What about DELETE_DEREF? PyCell_Set() also is used in unicode_concatenate().

[python-dev]

New changeset d78d45436753 by Raymond Hettinger in branch '3.6':
Issue bpo-28665: Harmonize STORE_DEREF with STORE_FAST and LOAD_DEREF giving a 40% speedup.
https://hg.python.org/cpython/rev/d78d45436753

[rhettinger]

Thanks for the quick review. I'll look at the other two cases when I get a chance.

[python-dev]

New changeset 7ec45e7d2194 by Serhiy Storchaka in branch 'default':
Merge from 3.6 (issue bpo-28665).
https://hg.python.org/cpython/rev/7ec45e7d2194

[serhiy-storchaka]

You forgot to merge and created new head.

[serhiy-storchaka]

Could you please measure the performance effect of these changes?