From: Matt J on
"Joel" <esposito(a)usna.edu> wrote in message <hljitv$rg8$1(a)fred.mathworks.com>...
> I used 1 million points and N=20 in my testing.
>
> Any other possible ideas?

I think you're already using the fastest method (assuming again that small N is what you'll always be using).
From: Joel on
How large do you think N should be before BSXFUN is faster?
are you talking 30 or 3000?
From: Matt J on
"Joel" <esposito(a)usna.edu> wrote in message <hljkad$se8$1(a)fred.mathworks.com>...
> How large do you think N should be before BSXFUN is faster?
> are you talking 30 or 3000?

you'd have to test it, but I would guess that by N=100 BSXFUN will start to pull into the lead.