What is the computer assembly line

What is the computer assembly line

The computer assembly line is Intel's first use in the 486 chip. The assembly line works like an assembly line in industrial production. In the CPU, an instruction processing assembly line is composed of 5-6 different functional circuit units, and then an X86 instruction is divided into 5-6 steps and then executed by these circuit units separately, so that one instruction can be completed in one CPU clock cycle. , so improve the CPU's computing speed. Classic Pentium each integer assembly line is divided into four levels, that is instruction prefetch, decoding, execution, write back the results, floating point is divided into eight levels of assembly line.
Assembly line technology is a key technology widely used in microprocessor chips (CPU). Computer automation line technology refers to the description of the execution mode of each instruction in the CPU. To understand it, it must be First understand the instructions and their execution.

1, computer instructions
Computer instructions are a set of specific binary sets that tell the CPU what to do. If we compare a CPU to a processing plant, then one instruction is like an order, which triggers a series of actions of the CPU_processing plant, and finally obtains the operation results and products. So how do they work? First, there must be a department that receives orders - the CPU's instruction fetching mechanism; and secondly, there must be a shop that completes the order - the CPU's execution instruction mechanism. In the factory, the product on an order is divided into many processes, and the instructions are also converted into many corresponding micro-operations in the CPU. After completing them in order, the entire instruction is executed.

2, the implementation of instructions
In the low-end CPU, the execution of the instructions is serial. In simple terms, after the completion of an instruction, the next instruction is executed. For example, the above-mentioned processing plant at the beginning of the venture has only a small workshop. And the boss who is fighting alone, then, when he receives an order, he must be busy completing the first order, and not able to pick up the second order. In this way, orders are accepted, orders are completed, orders are received, orders are fetched, instructions are executed, instructions are fetched, and instructions are taken. This is a serial process. Later, the boss found that it took a lot of time to accept the order, and he had a helper who could work independently of one another. In this way, the owner accepted the next order while completing the last order. This is manifested in the CPU is the instruction fetching mechanism and the execution of the command mechanism are separated, so that from the CPU as a whole, the CPU in the implementation of the last instruction at the same time in parallel to remove the instructions. This is a qualitative leap in the CPU technology, it makes the CPU from the serial work into parallel work, which has a streamlined prototype.

After the CPU has completed this step, what remains is how to improve the parallel processing capability. The CPU designers are inspired by the assembly line of the processing plant and break down the execution of one instruction into many different ones. The process_microinstruction greatly simplifies the complexity of the instruction, simplifies the logic design, and increases the speed. In a CPU with an assembly line technology, the first instruction has just executed the first "process", and immediately the second instruction has been added to the pipeline to begin execution. Obviously, this kind of  automation conveyor technology requires multiple execution units, which are implemented in the X86 chip.

Through the above introduction, we have learned what pipeline technology is. Although this is not an innovation, it is a major obstacle in the realization of technology. It is a major contribution of CPU designers to the development of computers.

So, what's going on with the P6 chip's super-assembly line?

3, P6 super-assembly line profile
The Super assembly line is still essentially an assembly line  technology, but it does the following improvements.

A. The number of automation line has increased from the two Pentiums to three, and there are eleven independent execution units supporting them in parallel.

B. Out-of-order processing techniques are used in the execution. That is, when an instruction requires some data and fails to execute it immediately, it will be stripped out of the pipeline and automation system for data. The CPU will immediately execute the next instruction, which is like finding out that some product is not qualified on the assembly line and it is eliminated. , waiting for rework a reason. In this way, it is possible to prevent one instruction from being executed and affect the efficiency of the entire conveyor line.

C. The instructions are divided into finer stages in P6, which makes the logic design, processes, etc. easier and speeds up. In the 486 chip, an instruction is generally divided into five standard parts, as is Pentium. In P6, due to the adoption of RISC-like techniques, one instruction was divided into fourteen record periods. This greatly increases the speed of the assembly line.