Computer architecture is made up of rules, methods, and procedures that describe how computer systems implement and function. Architecture meets the demands of the user while also taking into account economic and financial constraints. Previously, architecture was on paper and then realized using hardware.
The architecture is created, tested, and formed in hardware form once the transistor-transistor logic is built in. The performance, efficiency, dependability, and cost of a computer system can all characterize computer architecture. It is with technology standards for software and hardware. The CPU, memory, I/O devices, and communication channels that connect to it make up the computer system.
1. Von-Neumann Architecture
John von Neumann proposed this architecture. The architecture of today’s computers is based on von Neumann architecture. It is based on a few ideas.
For read and write instructions and data, we have a single read/write memory available. When we talk about memory, we’re talking about a single area that’s for reading and writing data and instructions. Within the computer system, data and instructions store in a single read/write memory.
Every memory has numerous places, each with its own address. We can read or write any data and instructions by addressing the contents of memory by its position, regardless of what form of data and instructions contained in the memory. Unless a change is required, execution is always done in a sequential order. If, for example, we’re running an instruction from line 1 to line 10, but we need to run line 50 instead of line 11, we’ll jump to instruction 50 and run it.
The instruction and data code execute on a bus (address bus/data bus/control bus). The input device receives data or instructions, and the central processing unit (CPU) executes one action at a time, either fetching data or instructions into or out of memory. The process is completed and then communicated to the output device. The central processing unit contains the control and logic units for processing processes.
2. Harvard Architecture
When data and code are stored in distinct memory blocks, Harvard architecture is employed. For data and instructions, a separate memory block is required. One memory location can access data, whereas another memory location can access instructions. It has all of its data storage within the central processor unit (CPU). It’s just necessary to use a single set of clock cycles. It is possible to build the pipeline. It is difficult to design. The CPU is capable of reading and writing instructions as well as processing data access. Harvard design contains distinct access codes and data address spaces, i.e., instruction address zero and data address zero are not the same. Data address zero identifies an 8-byte value that is not part of the 24-byte value, whereas instruction address zero identifies a 24-byte value.
The modified Harvard architecture machine is similar to a Harvard architecture machine in that it contains a shared address space for the data and instruction caches. It has digital signal processors that can run simple or complex audio or video algorithms, and it can replicate. Microcontrollers have a limited number of programs and data memory, and they speed up processing by running concurrent instructions and accessing data.
As seen in the figure below, there is separate data and instruction memory, as well as a bus for performing operations. It’s totally within the Central Processing Unit (CPU). It contains a separate arithmetic and logic unit and can execute simultaneous input/output operations.
3. Instruction Set Architecture
Instruction set architecture must complete the architecture since it contains a set of instructions that the processor understands. It has two instruction sets: RISC (reduced instruction set computer) and CISC (common instruction set computer) (complex instruction set computer).
IBM pioneered reduced instruction set computer architecture in the 1990s. Multiple address modes are available in instruction, but programs do not use all of them, which is why the number of address modes have limits. This makes it easier for the compiler to write the instructions, and the number of instructions written increases.
Because early compilers did not allow programmers to build programs, complex instruction set architecture make programming easier. The best results obtain by using ISA’s straightforward instructions.
4. Microarchitecture
When the instruction set architecture is a built-in processor, microarchitecture is famous as computer organizations. Instruction set architecture implements using diverse microarchitecture, which differs due to technological advancements.
Microarchitecture has a specific function. It receives and decodes the instruction, finds parallel data to process the instruction, and finally processes the instruction and generates output.
Microprocessors and microcontrollers use it. While some designs overlap numerous instructions while executing, microarchitecture does not. Execution units, such as arithmetic logic units, floating-point units, load units, and so on, are must-have for the processor’s operation. Within the system, microarchitecture decisions such as memory size, latency.
5. System Design
The name says it all: the design will meet user needs for a system’s architecture, modules, interfaces, and data, and it will link to product development. It’s the process of gathering marketing data and turning it into a product design. Hardware and software create modular systems.
Click here to read more useful articles.