In what cases is using multiple processors inefficient or undesirable?

Using multiple processors (multiprocessing) can be inefficient or undesirable in the following cases:

• Tasks with strong dependence on sequential execution. If the algorithm cannot be parallelized (for example, due to sequential dependencies), the overhead of context switching and synchronization will outweigh the benefits.

• Small tasks with low computational load. Parallelizing small operations can lead to significant overhead in creating and managing processes.

• Resource limitations. If the system is limited in memory or other resources, launching multiple processes can degrade performance.

• Interaction with global resources. Frequent access to shared resources (such as files, databases) can become a bottleneck due to synchronization.

• **In Python, due to GIL (Global Interpreter Lock), multithreading does not provide a gain in CPU-bound tasks, so multiprocessing is preferred, but it also has overheads.

Example: if the task is simply to sum a small list of numbers, launching multiple processes will be redundant and slow down execution.