⚡️ Slightly faster SequenceSet#xor
#567
Merged
+12
−6
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The results are highly dependant on data, YJIT, etc... but in my tests
this implementation seems to _consistently_ run faster.
```
Comparison:
builtin: L ^ R
local: 967.0 i/s
before #567 (559b86a): 917.5 i/s - 1.05x slower
before #562 (a228ee1): 901.2 i/s - 1.07x slower
v0.5.12: 840.0 i/s - 1.15x slower
v0.5.0: 830.7 i/s - 1.16x slower
v0.4.21: 797.0 i/s - 1.21x slower
v0.5.9: 790.4 i/s - 1.22x slower
builtin: L.xor! R
local: 960.9 i/s
before #562 (a228ee1): 954.3 i/s - 1.01x slower
before #567 (559b86a): 940.0 i/s - 1.02x slower
v0.5.12: 796.4 i/s - 1.21x slower
v0.5.0: 633.9 i/s - 1.52x slower
v0.4.21: 620.1 i/s - 1.55x slower
v0.5.9: 619.0 i/s - 1.55x slower
```
When I first wrote this code, the `master` branch was consistently
1.3x-1.5x slower. The speedup is much less now, so other implementation
details are probably more important. When the internal set data
implementation is changed (see #484), this and other operations should
be re-evaluated. I suspect an iterative "merge" may be the fastest
overall approach, but that code is more complex than simply relying on
the core add/subtract/complement methods with simple boolean algebra.
nevans
force-pushed
the
sequence_set/faster-xor
branch
from
131997d to
a42108a
Compare
nevans
changed the title
SequenceSet#xor ~1.4x fasterSequenceSet#xor
nevans
added
the
sequence-set
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The results are highly dependant on data, YJIT, etc... but in my tests this implementation seems to consistently run faster.
When I first wrote this code, the
masterbranch was consistently 1.3x-1.5x slower. The speedup is much less now, so other implementation details are probably more important. Specifically, most of the speedup probably came from the change in #550, which coercedrhsto a SequenceSet before using it.When the internal set data implementation is changed (see #484), this and other operations should be re-evaluated. I suspect an iterative "merge" may be the fastest overall approach, but that code is more complex than simply relying on the core add/subtract/complement methods with simple boolean algebra.