Performance- Probleme auf modernen Intel-Prozessoren

The new computer is likely equipped with a 12th generation Intel processor or newer.

Since the 12th generation (Alder Lake) of Core i processors, Intel has made a fundamental change to the CPU architecture.

Until this generation, a processor contained several identical CPU cores. Starting with the 12th generation, this is now different.

There are two different types of cores. The performance cores (P-cores) are tailored for high computing power. For this, they consume a lot of power. A high power consumption means that in a notebook the battery drains faster and that more heat is generated in the processor, which needs to be dissipated by cooling.

Besides the P-cores, there are efficiency cores (E-cores). These handle the many smaller tasks in the background where speed is not crucial. The E-cores are considerably slower but consume much less power and consequently produce less heat.

Processors with different types of cores are not a new invention by Intel. For smartphones and other mobile devices, this has already been standard for a long time. Here, the demand for low energy consumption and coverage of short-term performance peaks is even more stringent than for notebooks or even desktop PCs. Apple also uses this technology in devices with M processors.

The operating system's task is now to allocate the running processes to the cores. This task is handled by the scheduler. It is not a trivial task to determine whether a process needs to be assigned to the P-cores or if the E-cores suffice. The scheduler, among other things, considers user permissions and can be found here:

If we now look at the architecture of RFEM 6, it becomes clear that problems can arise.

The program architecture of RFEM 6 is dual. On one side, there is the program with the graphical user interface. This takes user inputs and displays results graphically. On the other side, there is the solver. It reads the input data, performs the calculations, and then returns the results back to RFEM.

In RFEM, one (RFEM 5) or several (RFEM 6) solver processes start and then run in parallel on multiple cores to perform the calculations. These are, of course, computation-intensive tasks.

Due to the fact that the processes are started by the main program as invisible subprocesses, it can happen that these processes are executed on the slower E-cores.

As a result, RFEM on a processor of the 12th generation suddenly computes significantly slower than on an older 11th generation processor.

A quick remedy is to start RFEM with admin rights. However, this is not always possible for security reasons.

Another solution is to influence the rule that subprocesses are executed on E-cores. Proceed as follows:

1. Open the Registry.
2. Search for "Heterogeneous thread scheduling policy".
3. Edit the key "Attributes" and set it to "0".

Registry

Power Settings