Skip to content

Latest commit

 

History

History
42 lines (30 loc) · 1.96 KB

Troubleshooting-Guide.md

File metadata and controls

42 lines (30 loc) · 1.96 KB

Troubleshooting Guide

General Troubleshooting Steps

  • Check your geometry (ascent/samples/sample_index/models/model_index/debug_geom.mph)
  • Check your sample (ascent/samples/sample_index/slides/cassette_index/slide_index/masks)

Installation Issues

  • NEURON
    • Issue: Sometimes NEURON installation will complete but NEURON will not be installed
    • Solution: Run the installer in Windows 8 compatibility mode

Pipeline Issues

  • Sample
    • Issue: Small white pixel islands cause the pipeline to fail, or the pipeline does not fail but very small elements are present as fascicles that reflect segmentation errors
    • Solution: increase value of image_preprocessing>object_removal_area in sample.json
  • Model
    • Issue: fascicle borders appear ragged/wavy in model geometry
    • Solutions:
      • Increase inner_interp_tol and/or outer_interp_tol in sample.json (depending on whether the Issue exists for outers or inners)
      • If your mask is very pixelated, increase smoothing>fascicle distance in sample.json
    • Issue: The pipeline solves one basis, then errors.
    • Solution: In COMSOL, disable automatic saving of recovery files.

COMSOL Issues

  • Issue: fiber xy-location(s) fall outside the solution, which results in error while extracting potentials. -Error log: - Exception: com.comsol.util.exceptions.FlException: Internal numerical error - Messages: Internal error in numerical routines. -Output log: - Intel MKL Error: Parameter 7 was incorrect on entry to DGELS.
  • Solution: Increase parameter in Sim -> "fibers" -> "xy_trace_buffer". Note: this is assuming that "endo_only_solution" in Run config is true.

NEURON Issues

  • Issue: Compiling NEURON files with python submit.py results in an error, mpicc: command not found
  • Solution: Install open-mpi 2.0 and add to path

Python Issues

  • Issue: python output prints out of order with java output
  • Solution: pass the -u flag (i.e. python -u run pipeline ), which turns off output buffering