Why do I see performance decrease when building my Simulink model or performing a matrix multiplication on a new and powerful Xeon processor?

The Xeon CPU takes much longer to build the model than the other CPU for the same project on the same operating system.

I also perform a matrix multiplication and use "timeit" to measure performance:

mat1 = rand(1e4);

mat2 = rand(1e4);

f = @() mat1 * mat2;
t = timeit;

The new Xeon processor takes much longer time for the above operation as well than an AMD processor on the same operating system.

Answer 1

[MathWorks Support Team]

Please check the energy settings of your system. We have seen cases where the energy option (power option) was set to "balanced" and on a Xeon processor this had more impact than on another processor.

Switching to energy option "high performance" for the first CPU hardly changed anything, whereas for the Xeon processor it was about 40% faster.