How is parallel computing implemented in multi-core processors?

Question

Programming

Guys, tell a non-programmer, do I understand everything correctly? The classical von Neumann architecture is strictly sequential calculations: one operation at one particular moment in time. But a multi-core processor allows you to move away from the sequence and implement parallelism, simultaneously executing the number of operations equal to the number of cores. That is, one core - one operation at a particular point in time.

Answer 1

Multi-core itself has nothing to do with the computational pipeline. Each core has its own pipeline and its operation does not depend on other cores.
The essence of multi-core is that each core executes its own computational thread . On an N-core processor, N threads are running in parallel at any given time.
Therefore, the number of processor cores can only affect the performance of multi- threaded programs. Any program that does not have multi-threading and load parallelization between threads will not receive any speed advantage when upgrading a single-core processor to a 16-core processor (at the same clock speed).

Answer 2

In fact, it looks like the truth, adjusted for the presence of a computational pipeline , due to which more than one operation is performed per cycle. Here is an overview article on the subject: Journey Through the Processor Computing Pipeline

Answer 3

Here is an interesting article habrahabr.ru/post/180595/
I would do this: I would
place it on the left server (some kind of file hosting service) and provide a link to download the book.
Or I posted it at my place, but I would leave a note for the copyright holder, saying that I will remove it at the first request. It’s not a fact, of course, that it will work in court if it comes to it, but the probability of being there is already lower.