perf: optimize strpos by eliminating double iteration for UTF-8 #19572
+31
−8
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
For non-ASCII strings, the original implementation used string.find() to get the byte index, then counted characters up to that byte index. This required two passes through the string. This optimization uses char_indices() to find the substring while simultaneously tracking character positions, completing the search in a single pass. Benchmark results (UTF-8 strings): - str_len_8: 188.98 µs → 140.54 µs (25.4% faster) - str_len_32: 615.69 µs → 294.15 µs (52.2% faster) - str_len_128: 2.2707 ms → 1.2462 ms (45.1% faster) - str_len_4096: 74.328 ms → 36.538 ms (50.9% faster) ASCII performance unchanged (already optimized with fast path).
viirya
force-pushed
the
strpos_optimize
branch
from
d190f85 to
5ed05f7
Compare
Contributor
|
run benchmark character_length |
This comment was marked as outdated.
This comment was marked as outdated.
Contributor
(I just added this to the whitelist and will run now) |
Contributor
|
run benchmark character_length |
|
🤖 |
Contributor
|
run benchmark strpos |
|
🤖: Benchmark completed Details
|
|
🤖 |
|
🤖: Benchmark completed Details
|
Contributor
|
run benchmark strpos |
|
🤖 |
alamb
approved these changes
Jan 1, 2026
| for (byte_idx, _) in string.char_indices() { | ||
| char_pos += 1; | ||
| if byte_idx + substring_bytes.len() <= string_bytes.len() | ||
| && &string_bytes[byte_idx..byte_idx + substring_bytes.len()] |
Contributor
There was a problem hiding this comment.
you could potentially use string_bytes.get_unchecked here if it makes any difference as you validate the bounds check immediately above.
Member
Author
There was a problem hiding this comment.
Use get_unchecked to avoid redundant bounds checking now when comparing substring slices.
|
🤖: Benchmark completed Details
|
Use `get_unchecked` to avoid redundant bounds checking when comparing substring slices, as we already validate the bounds in the if condition. Benchmark results show modest improvements, especially for shorter strings: - strpos_StringArray_ascii_str_len_8: 3.7% faster - strpos_StringViewArray_ascii_str_len_8: 2.0% faster - strpos_StringViewArray_utf8_str_len_8: 1.9% faster Other benchmarks show no significant change (within noise threshold).
viirya
force-pushed
the
strpos_optimize
branch
from
37c0509 to
4c6c95d
Compare
Member
Author
|
Thanks @alamb |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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.
Which issue does this PR close?
Rationale for this change
What changes are included in this PR?
For non-ASCII strings, the original implementation used string.find() to get the byte index, then counted characters up to that byte index. This required two passes through the string.
This optimization uses char_indices() to find the substring while simultaneously tracking character positions, completing the search in a single pass.
Benchmark results (UTF-8 strings):
ASCII performance unchanged (already optimized with fast path).
Are these changes tested?
Are there any user-facing changes?