On Thursday, January 8, 2015 at 3:29:01 PM UTC-5, Joshua Adelman wrote:
>
> You're reading it correctly. I'm not sure exactly why numpy is performing
> better for the larger matrices, but I suspect that numpy's accumulate
> function that np.vander uses may be taking advantage of simd, sse or mkl
> optimizations. I'd have to do a bit of digging though to confirm that. I
> also haven't experimented with Julia's @inbounds or @simd operators, but
> that might help. I also believe that some of the subarray stuff might be
> better optimized in Julia 0.4 vs 0.3.4.
>
Josh, I think you can do better in Julia just by avoiding subarrays,
powers, and cumprod entirely. Here is a version that seems significantly
faster than any of the ones in your notebook for large arrays:
function myvand{T}(p::Vector{T}, n::Int)
# n: number of rows
# p: a vector
m = length(p)
V = Array(T, m, n)
for j = 1:m
@inbounds V[j, 1] = 1
end
for i = 2:n
for j = 1:m
@inbounds V[j,i] = p[j] * V[j,i-1]
end
end
return V
end
And I think you can do better still by more unrolling and rearrangement,
even without using SIMD.
