[Util] Optimize conversion functions for Cortex-M0+ #41
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
Use shift-based operations instead of software division. Performance comparison (1000 iterations x test values): | Function | Original | Optimized | vs stdlib | |-----------|------------|------------|------------------| | utilItoa | 89,795 us | 33,784 us | 2.0x faster | | utilAtoi | 33,768 us | 12,643 us | 5.1x faster | | utilFtoa | 84,628 us | 63,126 us | N/A | | utilAtof | 159,728 us | 95,508 us | ~4% slower* | *utilAtof supports comma as decimal separator (European locales) Key optimizations: - fastDiv10() using shifts/adds instead of hardware divide - Process strings left-to-right to eliminate reversal - Multiply by 10 using (x<<3)+(x<<1) instead of *10 - utilAtoi/utilAtof no longer modify input buffer Also adds benchmark test for measuring performance on target. Co-Authored-By: Claude <[email protected]>
Add const correctness to utilAtof() and utilAtoi() functions since they only read from their string buffer parameters and never modify them. This improves API safety by preventing accidental modifications and allows these functions to accept string literals and const char* arguments without requiring casts. Co-Authored-By: Claude <[email protected]>
awjlogan
reviewed
Jan 8, 2026
SAMD21 has a single-cycle hardware multiplier, so the compiler optimizes `* 10` to a single `muls` instruction. The shift-based approach ((x << 3) + (x << 1)) was not faster and hurt readability. Co-Authored-By: Claude <[email protected]>
Contributor
Author
|
Hi @awjlogan |
awjlogan
reviewed
Jan 11, 2026
Collaborator
Almost certainly - just need to focus a bit on the CM part at the moment :) |
Replace shift-based *100 with direct multiplication. SAMD has single-cycle 32b multiplier making shifts unnecessary. Co-Authored-By: Claude <[email protected]>
Contributor
Author
|
@awjlogan |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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.
Summary
Performance Results
Measured on emon32 hardware (SAMD21, 1000 iterations per test value):
*
utilAtofsupports comma as decimal separator (European locales), which stdlibatofdoes not.Key Optimizations
fastDiv10()using shifts/adds only - avoids expensive software divide on M0+Multiply by 10 using(x<<3)+(x<<1)instead of*10utilAtoiandutilAtofno longer modify their input buffersTest plan
🤖 Generated with Claude Code