functions
abs erf log tanh
atan erfc log10
atan2 exp sin
cos int sqrt

The above math functions return a double precision value and take a double
precision argument. The exception is atan2 with two double precision arguments.
## DIAGNOSTICS

Arguments which are out of range give an argument domain diagnostic.
These functions call the library routines supplied by the compiler.

math

trigonometric function of radian argument.

see sin for the Vector class.

math

trigonometric function of radian argument.
see sin

math

returns the arc-tangent of y/x in the range -PI/2 to PI/2. (x > 0)

math

## SYNTAX

`radians = atan2(y, x)`

## DESCRIPTION

returns the arc-tangent of y/x in the range -PI < radians <= PI. y and x
can be any double precision value, including 0. If both are 0 the value
returned is 0.
Imagine a right triangle with base x and height y. The result
is the angle in radians between the base and hypotenuse
## EXAMPLES

execute following example
atan2(0,0)
for i=-1,1 { print atan2(i*1e-6, 10) }
for i=-1,1 { print atan2(i*1e-6, -10) }
for i=-1,1 { print atan2(10, i*1e-6) }
for i=-1,1 { print atan2(-10, i*1e-6) }
atan2(10,10)
atan2(10,-10)
atan2(-10,10)
atan2(-10,-10)

math

hyperbolic tangent.
see tanh for the Vector class.

math

logarithm to the base e
see log for the Vector class.

math

logarithm to the base 10
see log10 for the Vector class.

math

## DESCRIPTION

returns the exponential function to the base e
When exp is used in model descriptions, it is often the
case that the cvode variable step integrator extrapolates
voltages to values which return out of range values for the exp (often used
in rate functions). There were so many of these false warnings that it was
deemed better to turn off the warning message when Cvode is active.
In any case the return value is exp(700). This message is not turned off
at the interpreter level or when cvode is not active.

for i=690, 710 print i, exp(i)

math

square root
see sqrt for the Vector class.

math

returns the integer part of its argument (truncates toward 0).

math

absolute value
see abs for the Vector class.

math

normalized error function
@eqn
\[{\rm erf}(z) = \frac{2}{\sqrt{\pi}} \int_{0}^{z} e^{-t^2} dt\]
@endeqn

math

returns 1.0 - `erf(z)`

but on sun machines computed by other methods
that avoid cancellation for large z.

neuron/general/function/sin.hel : May 13 2012