Performance Optimization: Memory Pool & Early Termination

[mgyoo86]

Summary

Optimize PAM's pellet ablation routine (ablate!) with significant memory reduction and improved computational efficiency through allocation pooling and early termination logic.

Performance Improvements

Memory Allocation Reduction

ablate! function optimization:

Metric	Before	After	Improvement
Allocations	48.27k	6.34k	86.9% ↓
Memory	11.065 MiB	326.953 KiB	97.1% ↓
Execution time	15.9 ms	3.0 ms	81.2% ↓

---

Key Changes

1. Allocation Pool Pattern

Pre-allocate and reuse buffers for pellet_density calculations across iterations.

Implementation:

mutable struct AllocationPool{T<:Real}
    nsource_buffer::Vector{T}
    nsource_max_len::Int
end

Impact:

• Eliminates repeated allocations in hot loop
• Reuses single buffer across all flux surfaces
• Reduces memory fragmentation

2. Early Termination Logic

Detect pellet full ablation and terminate simulation immediately.

Implementation:

if pelt.radius[k] <= 0.0
    @info "Pellet fully ablated at t = $(pelt.time[k]) s"
    return
end

Impact:

• Avoids unnecessary iterations after pellet disappears
• Faster completion for high ablation rate scenarios
• Prevents NaN propagation

3. Test Infrastructure

• Add OMFIT PAM comparison tests for validation
• Switch to HDF5 format for reduced memory footprint
• Verify JSON/HDF5 format compatibility

---

Performance Profiling Note

While ablate! showed significant improvement (97% memory reduction), overall runtime is bottlenecked by IMAS.trace_surfaces, which consumes the majority of execution time in run_PAM.

Profiling breakdown:

• IMAS.trace_surfaces: ~85-90% of total runtime
• ablate! (optimized): ~10-15% of total runtime

Future optimization opportunities lie in the IMAS.trace_surfaces implementation.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 70.95%. Comparing base (b33c3b0) to head (110b41e).
⚠️ Report is 5 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #10      +/-   ##
==========================================
+ Coverage   70.82%   70.95%   +0.13%     
==========================================
  Files           1        1              
  Lines         329      334       +5     
==========================================
+ Hits          233      237       +4     
- Misses         96       97       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[avdeevag]

works well, thanks!

[orso82]

you should have merge permissions @avdeevag ;)