What is the most performant configuration to run my Polyspace verification analysis?

I am running a Polyspace Code Prover verification analysis and I would like it to run faster. I want to make use of the parallelization capabilities of Polyspace and I tried using the option -max-processes

I intended to make use of the 4 physical + 4 virtual cores on my machine.

However it seems that Code Prover only recognizes 4 cores per the message below:

*Remark: 4 core(s) detected. The verification uses 4 core(s). 
Value 8 of -max-processes option has been ignored.*

Is there any way to use virtual cores and more generally, how can I improve the performance of my verification analysis?

Answer 1

[MathWorks Support Team]

Using Polyspace Code Prover, it is not possible to increase number of cores unless having physical cores on the machine. So -max-processes 8 will work only and only if you have 8 real cores (hyper-threading is not taken into account).

Now, we have experimented that increasing number of cores does not help too much for a Polyspace Code Prover analysis.

Attached to this answer is our unofficial Hardware Guidelines for Polyspace Server. This shows in particular the speedup gained from the number of physical cores used, as you can see, going from 4 to 8 does not improve much.

You will also find some information that may help you make the best use out of Polyspace and Code Prover in particular.