Simulation errorsΒΆ

Occasionally, a simulation will encounter a numerical error. Some simulation types attempt to catch these errors and raise a SimulationError. These provide detailed error messages that can help diagnose the source of the error.

A common source of numerical errors is division by some term (x - c), where x is a variable and c is a constant. Using double precision, this should only happen very rarely and try-catch blocks around a simulation are not typically necessary. This error is much more common when using single precision (for example during a GPU enabled simulation) or when applying a step protocol (for example, a term (V - 40) will not cause problems in most simulations, but can trigger errors if V is stepped to exactly 40 mV. In these scenarios, it is advised to add a try-catch block around a simulation:

import myokit
m,p,x = myokit.load('model.mmt')
s = myokit.SimulationOpenCL(m, p, 64)
except myokit.SimulationError as e:

Example result:

Encountered numerical error at t=411.054992676 in cell (55) when membrane.V=-77.0.
Obtained 4 previous state(s).
State before:
membrane.V = -77.0
ina.m      =  5.76317822560667992e-03
ina.h      =  3.50723683834075928e-01
ina.j      =  1.33311852812767029e-01
ica.d      =  6.89173266291618347e-02
ica.f      =  7.21855103969573975e-01
ik.x       =  3.83078485727310181e-01
ica.Ca_i   =  2.27514933794736862e-03
State during:
membrane.V = nan
ina.m      =  5.75887179002165794e-03
ina.h      =  3.50991368293762207e-01
ina.j      =  1.33416950702667236e-01
ica.d      =  6.88845962285995483e-02
ica.f      =  7.21879780292510986e-01
ik.x       =  3.83071571588516235e-01
ica.Ca_i   =  2.27477238513529301e-03
Evaluating derivatives at state before...
invalid value encountered in float_scalars
Encountered when evaluating
  if(membrane.V < -100.0, 1.0, 2.837 * (exp(0.04 * (membrane.V + 77.0))

        - 1.0) / ((membrane.V + 77.0) * exp(0.04 * (membrane.V + 35.0))))

With the following operands:
  (1) 0.0
  (2) 0.0
And the following variables:
  membrane.V = -77.0
             = -77.0