% Reductions with fused loaders, 300 times over 10 million elements:
% dot, norm and max each read their operands exactly once per pass --
% sum(a.*b) is one sweep of memory, not two.
n = 10000000;
reps = 300;
a = rand(n, 1);
b = rand(n, 1);
tic;
for k = 1:reps
  d = dot(a, b);
  nrm = norm(a - b);
  mx = max(abs(a - b));
end
t = toc;
fprintf('dot %.0f   norm %.2f   max %.6f\n', d, nrm, mx);
fprintf('%.1f GB/s effective\n', reps*6*n*4/1e9/t);