Stochastic Floating-Point Types

This package defines a new SFloat64 type which represents a regular double-precision floating-point number, except every operation gets randomly rounded upwards or downwards. It should be possible tu use SFloat64 numbers everywhere Float64 would be supported. For example:

julia> A = Float64[1 2 3;
                   4 5 6;
                   7 8 10];
julia> B = Float64[1 ; 2 ; 3];

julia> A \ B
3-element Array{Float64,1}:
 -0.3333333333333333
  0.6666666666666666
 -0.0

julia> SFloat64.(A) \ SFloat64.(B)
3-element Array{SFloat64,1}:
 SFloat64(-0.3333333333333325)
 SFloat64(0.666666666666665)
 SFloat64(8.881784197001221e-16)

Reference