This provides a Julia interface to some of OS X's Accelerate framework. At the moment, the package consists mostly of an interface to the array-oriented functions, which provide a vectorised form of many common mathematical functions, however the package does provide access to several other vectorized operations and more are being added on a regular basis. In general, the performance is significantly better than using standard libm functions, though there does appear to be some reduced accuracy.
The following functions are supported:
Note there are some slight differences from behaviour in Base:
DomainErrors are raised, instead
NaNvalues are returned.
roundbreaks ties (values with a fractional part of 0.5) by choosing the nearest even value.
exponentreturns a floating point value of the same type (instead of an
Some additional functions that are also available:
rec(x): reciprocal (
1.0 ./ x)
rsqrt(x): reciprocal square-root (
1.0 ./ sqrt(x))
pow(x,y): power (
x .^ yin Base)
fdiv(x,y): divide (
x ./ yin Base)
To avoid naming conflicts with Base, methods are not exported and so need to be accessed via the namespace:
using AppleAccelerate X = randn(1_000_000) @time Y = exp(X) # standard libm function @time Y = AppleAccelerate.exp(X) # Accelerate array-oriented function
@replaceBase macro replaces the relevant Base methods directly
AppleAccelerate.@replaceBase sin cos tan AppleAccelerate.@replaceBase(.^, ./) # use parenthesised form for infix ops @time sin(X) # will use AppleAccelerate methods for vectorised operations
Output arrays can be specified as first arguments of the functions suffixed
out = Array(Float64,1_000_000) @time AppleAccelerate.exp!(out, X)
Warning: no dimension checks are performed on the
! functions, so ensure
your input and output arrays are of the same length.
Operations can be performed in-place by specifying the output array as the
input array (e.g.
AppleAccelerate.exp!(X,X)). This is not mentioned in the
Accelerate docs, but this comment by one of the authors indicates that it is safe.