# NAME

fpmath – a module for Tensile

# DESCRIPTION

fpmath adds floating-point support to Tensile.

# FUNCTIONS

## isfloat(num)

 RETURNS: true if num is a legal floating-point number

 RETURNS: the sum of all its argument

## fpostincr(var val)

 RETURNS: old value of var

## fincr(var val)

 RETURNS: new value of var

## fpostdecr(var val)

subtracts val from var
 RETURNS: old value of var

## fdecr(var val)

subtracts val from var
 RETURNS: old value of var

## fmul(val1 val2 ...)

 RETURNS: a multiply of vals

## fsub(val1 val2)

 RETURNS: difference between val1 and val2

## fdiv(val1 val2)

 RETURNS: quotient of val1 and val2

## fneg(val)

 RETURNS: a negated val

## fabs(val)

 RETURNS: the absolute value of val

## fint(val)

 RETURNS: integral part of val

## frac(val)

 RETURNS: fractional part of val

## fsqrt(val)

 RETURNS: a square root of val

## fln(val)

 RETURNS: a natural logarithm of val

## fsin(val)

 RETURNS: a sine of val (in radians)

## fcos(val)

 RETURNS: a cosine of val (in radians)

## fexp(val)

 RETURNS: e^val

## fpower(val1 val2)

 RETURNS: val1^val2

## fprecision([prec])

if prec is present, sets the floating point precision to it
 RETURNS: the old value of precision

## Comparison functions

eq, neq, less, greater, le, ge accept float as their first argument, resulting in floating point comparison