Fix(macos): Use frequency-weighted GPU usage calculation#1293
Merged
Conversation
The previous GPUPH-based calculation reported raw P-state residency (time in any active state), which significantly overestimated GPU usage (e.g. 50-60% at idle) because macOS keeps the GPU clocked at P1 for display compositing. Now reads the DVFS frequency table from the PMGR IOKit node and applies the formula: (avg_active_freq × active_ratio) / max_freq, matching the approach used by macmon. Falls back to unweighted calculation if the frequency table is unavailable.
github-actions
Bot
added
bug
Contributor
There was a problem hiding this comment.
Pull request overview
Updates macOS (Apple Silicon) GPU usage reporting to avoid inflated idle readings by weighting GPUPH P-state residency with the GPU DVFS frequency table read from the PMGR IOKit node, aligning the metric with macmon.
Changes:
- Add PMGR (
pmgr/voltage-states9) DVFS frequency-table reader and expose it to the IOReport sampler. - Compute GPU usage using frequency-weighted residency when frequencies are available; otherwise fall back to unweighted active-time ratio.
- Add architecture documentation describing the data sources and formula used for macOS GPU usage.
Reviewed changes
Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
| src-tauri/src/infrastructure/providers/macos/io_kit/iokit_info.rs | Adds PMGR DVFS table parsing (voltage-states9) for GPU P-state frequencies. |
| src-tauri/src/infrastructure/providers/macos/io_kit/io_report.rs | Stores DVFS frequencies at init and applies frequency-weighted GPU usage computation from IOReport deltas. |
| docs/ARCHITECTURE/macos-gpu-usage-algorithm.md | Documents the macOS GPU usage algorithm, inputs, formula, and fallback behavior. |
Contributor
Rust Backend Coverage ReportCoverage Details |
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
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.
Summary
The previous GPUPH-based calculation reported raw P-state residency (time in any active state), which significantly overestimated GPU usage (e.g. 50-60% at idle) because macOS keeps the GPU clocked at P1 for display compositing.
Now reads the DVFS frequency table from the PMGR IOKit node and applies the formula: (avg_active_freq × active_ratio) / max_freq, matching the approach used by macmon. Falls back to unweighted calculation if the frequency table is unavailable.
Related Issues
Type of Change
fix/branch)feat/branch)refactor/branch)docs/branch)chore/branch)Screenshots / Videos
Test Plan
Checklist
npm run lint && npm run format/cargo tauri-lint && cargo tauri-fmt)npm test/cargo tauri-test)