About Computing sequence.

If i Set a matrix A = (1, 23, 17; -2, 1, 94)

and if there is a matrix B = (0.5; 1.6; -0.3)

and if i want to multiply A and B then put the result in C( C = A * B).

Does the computing sequence occurs simultaneously or sequentially[(1, 23, 17)*(0.5; 1.6; -0.3) first occurs and then (-2, 1, 94)*(0.5; 1.6; -0.3) occurs so that computting time take 2 times compared to simultaneous computting]?

Answer 1

[Guillaume]

matrices operations in matlab are very optimised, so it's likely that they are parallelised as much as possible. However, implementation details are not documented so you can't know for sure. In any case, whether or not the execution at the processor level occurs in parallel is going to depend on the version of matlab, the type of processor you have and probably the load on said processor.

Why do you care anyway?

[Jay Chung]

I wantted to know if there exists an time cost to code my algorithm with for loop. Anyway in some degree you mean if my code is vectorized, i can run it faster than using for loop?

[Stephen23]

"...if my code is vectorized, i can run it faster than using for loop?"

Vectorized operations are generally highly optimized, so in many cases vectorized code may be faster than a loop... in other cases the loop is faster. It depends on the data, the algorithm, how the code was written, the sizes of intermediate arrays, the MATLAB version, etc. etc.

The only way to know for sure is to test it.

But you should be concentrating first on writing clear, simple, and understandable code (which will in turn be easier to check and have fewer bugs). Unless you have written your code very carefully and used the profiler to determine that this particular operation is causing a major bottleneck and your code is going to run millions/billions/trillions/... of iterations, then there is likely little point in premature optimization of your code.