Reduce overhead for cache hits in specialized opcodes.

[markshannon]

BPO	45527
Nosy	@ambv, @markshannon, @Fidget-Spinner
PRs	bpo-45527: Don't count cache hits, just misses. #29092

^{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/markshannon'
closed_at = <Date 2021-10-20.18:54:51.464>
created_at = <Date 2021-10-19.17:24:13.601>
labels = ['interpreter-core', '3.11', 'performance']
title = 'Reduce overhead for cache hits in specialized opcodes.'
updated_at = <Date 2021-10-20.18:54:51.463>
user = 'https://github.com/markshannon'

bugs.python.org fields:

activity = <Date 2021-10-20.18:54:51.463>
actor = 'lukasz.langa'
assignee = 'Mark.Shannon'
closed = True
closed_date = <Date 2021-10-20.18:54:51.464>
closer = 'lukasz.langa'
components = ['Interpreter Core']
creation = <Date 2021-10-19.17:24:13.601>
creator = 'Mark.Shannon'
dependencies = []
files = []
hgrepos = []
issue_num = 45527
keywords = ['patch']
message_count = 4.0
messages = ['404320', '404423', '404521', '404522']
nosy_count = 3.0
nosy_names = ['lukasz.langa', 'Mark.Shannon', 'kj']
pr_nums = ['29092']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'performance'
url = 'https://bugs.python.org/issue45527'
versions = ['Python 3.11']

[markshannon]

Every time we get a cache hit in, e.g. LOAD_ATTR_CACHED, we increment the saturating counting. Takes a dependent load and a store, as well as the shift. For fast instructions like BINARY_ADD_FLOAT, this represents a significant portion of work done in the instruction.

If we don't bother to record the hit, we reduce the overhead of fast, specialized instructions.

The cost is that may have re-optimize more often.
For those instructions with high hit-to-miss ratios, which is most, this be barely measurable.
The cost for type unstable and un-optimizable instruction shouldn't be much changed.

Initial experiments show ~1% speedup.

[Fidget-Spinner]

Strong +1 from me.

Not to mention some instructions don't even need to read the _PyAdaptiveEntry apart from recording cache hits, so that's one more dependent load and store too.

Extremely cheap instructions off the top of my head:

• BINARY_SUBSCR
• LOAD_METHOD
• LOAD_GLOBAL
• LOAD_ATTR

[ambv]

New changeset bc85eb7 by Mark Shannon in branch 'main':
bpo-45527: Don't count cache hits, just misses. (GH-29092)
bc85eb7

[ambv]

Thanks, Mark! ✨ 🍰 ✨