Exclude from Project vs Power Save Mode for PyCharm Indexer CPU Spikes

Direct answer

Exclude from Project is the only option that actually reduces indexer CPU load. Power Save Mode does not throttle or pause the indexer; it only disables editor-level inspections, auto-import, and UI animations. If your goal is specifically lowering indexer CPU, exclusion is the correct lever.

However: excluding directories removes them from the project model entirely. Code inspections, refactoring, navigation, and "Find Usages" will not work on those files. Since you stated you need inspection and refactoring support for the excluded paths, you cannot use Exclude from Project on directories you still need to analyze.

Why Power Save Mode doesn't help the indexer

• Power Save Mode (File → Power Save Mode) stops background editor inspections, on-the-fly error highlighting, and auto-import.
• The indexer runs on a separate background thread and continues at normal priority. Its CPU usage appears as "Indexing" in the Background Processes tool window (View → Tool Windows → Background Processes).
• Power Save Mode masks overall IDE CPU but leaves indexing spikes untouched.

What happens when you exclude directories

1. Mark directories as Excluded (right-click → Mark Directory as → Excluded). The change is immediate in the UI.
2. PyCharm schedules a re-index. During that cycle the indexer processes the removal, so CPU may still spike for several minutes on large projects.
3. Only after the "Indexing" progress bar disappears are the directories fully dropped from the index. CPU measurements in Background Processes are reliable after that point.

Alternatives that preserve inspections

If the generated directories are build outputs (e.g., build/, dist/, target/), exclude them—they rarely need inspections. For vendored dependencies or generated code you do want analyzed:

• Library roots: Move the directory outside the project root, then add it as a library (File → Project Structure → Libraries → + → Java/Kotlin/Python). Library roots are indexed once, read-only, and fully support inspections/refactoring without re-indexing on every project change.
• Custom inspection scope: Keep the directory in the project but create a scope (File → Settings → Appearance & Behavior → Scopes) that excludes those paths from analysis while leaving them indexed. Use that scope in Inspect Code or inspection profiles.

Verification steps

1. Open Background Processes and note the "Indexing" entry.
2. Apply your chosen exclusion or library-root change.
3. Wait until the indexing progress bar finishes and the entry disappears.
4. Observe CPU during a normal editing session (no active indexing).
5. Test inspection coverage: right-click a file in the affected directory → Analyze → Inspect Code. Excluded files will report "File is not included in project"; library-root files will inspect normally.

One diagnostic question

Do you need inspections/refactoring to work inside the generated directories themselves, or only on project code that references them? If the former, use library roots or a custom scope. If the latter, simple exclusion is fine and gives the largest CPU reduction.