SCAFFOLD_LOGGING_SIMPLIFICATION.md

Scaffold Logging Simplification

Summary

Simplified the scaffold progress tracking by removing complex thread spawning, atomic counters, and separate progress meters. Replaced with a simple logging message.

Problems with Previous Implementation

1. Deadlock risk: Complex thread synchronization with atomic counters
2. Overhead: Separate monitoring thread checking progress every 10ms
3. Complexity: Multiple shared pointers, atomic operations, and memory ordering
4. Unnecessary: Scaffold filtering is typically fast enough not to need detailed progress

Changes Made

1. In computeMap.hpp

• Removed: Thread spawning for scaffold progress monitoring
• Removed: Atomic counters (scaffold_total_work, scaffold_completed_work)
• Removed: Separate progress meter for scaffolding
• Added: Simple log message: [wfmash::mashmap] Scaffolding mappings...

2. In mappingFilter.hpp

• Modified: filterByScaffolds function signature to make progress tracking parameters optional with default nullptr
• Removed: All fetch_add calls to update scaffold work counters
• Removed: Progress tracking from the parallel distance computation

Benefits

1. No deadlock risk: No threads or synchronization needed
2. Simpler code: Easier to maintain and debug
3. Same functionality: Scaffold filtering still works exactly the same
4. Better performance: No overhead from progress tracking

Testing

Confirmed that:

• Program compiles successfully
• Scaffold filtering still works (--scaffold-out produces output)
• No deadlocks or hangs
• Simple message "Scaffolding mappings..." is displayed

Example Output

[wfmash::mashmap] mapping  [100.0% complete, 4029392/4029392 units, 0s elapsed]
[wfmash::mashmap] mapping  completed.
[wfmash::mashmap] Scaffolding mappings...
[wfmash::mashmap] Mapped query in 0.4s, results saved to: /dev/stdout