# How to specify precision level of a compiled function? [closed]

In order to speed up the memoization of data generated via NIntegrate, a possible approach is to use Compile on both NIntegrate function and the function itself separately. How can one set the WorkingPrecision, PrecisionGoal and AccuracyGoal of a compiled function like the example presented below?

cf = Compile[{{x, _Real}}, Sin[x] + x^2 - 1/(1 + x)]

• Compile only works with machine number reals, so there is no notion of WorkingPrecision. – Carl Woll Sep 21 '17 at 19:28