Consider fast_double_parser for faster str->double

[pitrou]

BPO	42081
Nosy	@tim-one, @mdickinson, @pitrou, @ericvsmith
Superseder	bpo-41310: micro-optimization: increase our float parsing speed by Nx

^{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 2020-10-19.17:20:42.966>
created_at = <Date 2020-10-19.16:56:46.019>
labels = ['interpreter-core', 'type-feature', '3.10']
title = 'Consider fast_double_parser for faster str->double'
updated_at = <Date 2020-10-19.17:20:42.965>
user = 'https://github.com/pitrou'

bugs.python.org fields:

activity = <Date 2020-10-19.17:20:42.965>
actor = 'pitrou'
assignee = 'none'
closed = True
closed_date = <Date 2020-10-19.17:20:42.966>
closer = 'pitrou'
components = ['Interpreter Core']
creation = <Date 2020-10-19.16:56:46.019>
creator = 'pitrou'
dependencies = []
files = []
hgrepos = []
issue_num = 42081
keywords = []
message_count = 3.0
messages = ['378960', '378964', '378967']
nosy_count = 4.0
nosy_names = ['tim.peters', 'mark.dickinson', 'pitrou', 'eric.smith']
pr_nums = []
priority = 'low'
resolution = 'duplicate'
stage = 'resolved'
status = 'closed'
superseder = '41310'
type = 'enhancement'
url = 'https://bugs.python.org/issue42081'
versions = ['Python 3.10']

[pitrou]

This is more of a "could be nice" entry if there are any interested core developers.

A new algorithm and library was recently published for faster parsing of double numbers. It claims rigorous accuracy and the performance is impressive. The implementation is in C++ but it's really decorated C. It also seems simple enough that it can be reimplemented independently.

https://github.com/lemire/fast_double_parser/

[ericvsmith]

This was proposed in bpo-41310.

[pitrou]

Oh, thank you.