Computer System Architecture

• Computer system architecture refers to the conceptual structure and fundamental operational organization of a computer system.
• It encompasses the arrangement of hardware and software elements and how they interact to execute tasks.
• It considers instruction set architecture, microarchitecture, logic design, and implementation.

Instruction Set Architecture (ISA)

• ISA is the interface between hardware and software, defining instructions the CPU can execute.
• It includes instruction formats, addressing modes, registers, data types, and the instruction set.
• Examples of ISAs are x86, ARM, RISC-V, and MIPS.
• Different ISAs lead to different performance characteristics, power consumption, and complexity in hardware design.

Microarchitecture

• Microarchitecture is the implementation of the ISA, detailing how the ISA is realized in hardware.
• It involves components like the CPU's internal organization, cache hierarchy, pipelining, and branch prediction.
• Different microarchitectures can implement the same ISA, trading off performance, power, and cost.

Memory Hierarchy

• Memory hierarchy is the organization of memory components in terms of speed and cost.
• Typically includes registers (fastest, most expensive), cache (L1, L2, L3), RAM (main memory), and secondary storage (slowest, least expensive).
• Caches store frequently accessed data for quick retrieval, reducing average memory access time.
• Cache coherence protocols ensure data consistency across multiple caches in multi-processor systems.

Input/Output (I/O) System

• I/O system manages communication between the CPU and peripheral devices.
• It includes device controllers, buses, and interfaces (e.g., USB, PCIe).
• Direct Memory Access (DMA) allows devices to transfer data directly to/from memory without CPU intervention, improving efficiency.
• Interrupts signal the CPU when a device requires attention, enabling asynchronous operation.

Multiprocessing and Parallelism

• Multiprocessing involves multiple CPUs in a single system, enabling parallel execution of tasks.
• Symmetric Multiprocessing (SMP) provides equal access to system resources for all processors.
• Parallel processing divides a task into multiple subtasks that can be executed concurrently.
• Techniques include multi-threading, vector processing, and distributed computing.

Pipelining

• Pipelining is a technique that overlaps the execution of multiple instructions.
• Instructions are divided into stages (e.g., fetch, decode, execute, memory access, write back), allowing multiple instructions to be in different stages simultaneously.
• Hazards (data, structural, control) can disrupt the pipeline, requiring stalling or forwarding to maintain correctness.

Branch Prediction

• Branch prediction attempts to predict whether a conditional branch will be taken or not taken.
• Accurate branch prediction reduces pipeline stalls caused by control hazards.
• Branch prediction techniques include static prediction (based on branch type) and dynamic prediction (based on past behavior).

Memory Management

• Memory management involves allocating and deallocating memory space for programs.
• Virtual memory allows programs to access more memory than physically available by swapping data between RAM and secondary storage.
• Address translation maps virtual addresses to physical addresses, enabling memory protection and isolation.

System Bus

• System bus connects the CPU, memory, and I/O devices.
• Front-side bus (FSB) traditionally connected the CPU to the northbridge chipset.
• Modern systems use faster interconnects like Intel's QuickPath Interconnect (QPI) or AMD's HyperTransport.

Cache Coherence

• Cache coherence ensures that multiple caches in a multiprocessor system have a consistent view of shared memory.
• Protocols like MESI (Modified, Exclusive, Shared, Invalid) maintain cache coherence by tracking the state of cache lines.
• Snooping protocols monitor bus activity to detect cache updates, while directory-based protocols use a central directory to track cache line ownership.

Interrupts and Exceptions

• Interrupts are signals that cause the CPU to suspend its current execution and handle a specific event.
• Hardware interrupts are triggered by external devices, while software interrupts (system calls) are initiated by programs.
• Exceptions are abnormal conditions that occur during program execution, such as division by zero or memory access violations.
• Interrupt handlers (ISRs) are routines that respond to interrupts and exceptions.

Virtualization

• Virtualization allows multiple operating systems to run concurrently on a single physical machine.
• A hypervisor (virtual machine monitor) manages the virtual machines and allocates resources.
• Hardware virtualization extensions (e.g., Intel VT-x, AMD-V) improve virtualization performance.

Power Management

• Power management techniques reduce energy consumption in computer systems.
• Clock gating disables the clock signal to inactive components.
• Voltage scaling adjusts the voltage supplied to the CPU based on workload.
• Power-gating completely shuts off power to unused components.

Multicore Architecture

• Multicore processors integrate multiple CPU cores on a single chip.
• Cores can share resources like cache and memory controllers.
• Multicore architectures improve performance by enabling parallel execution of tasks.

GPU Architecture (Graphics Processing Unit)

• GPUs are specialized processors designed for parallel processing of graphics and compute-intensive tasks.
• Consist of many cores that execute the same instruction on different data (SIMD).
• Used in gaming, machine learning, and scientific simulations.

Heterogeneous Computing

• Heterogeneous computing systems combine different types of processors (e.g., CPU, GPU, FPGA) to optimize performance and energy efficiency.
• Offloads tasks to the most suitable processor for the given workload.
• Common in mobile devices, embedded systems, and high-performance computing.

Embedded Systems Architecture

• Embedded systems are computer systems designed for specific applications within larger devices or systems.
• Often have real-time constraints and low power requirements.
• Examples include microcontrollers in consumer electronics, automotive systems, and industrial control systems.

Digital Logic

• Digital logic is the fundamental building block of computer systems.
• Uses logic gates (AND, OR, NOT, XOR) to perform boolean operations.
• Combinational logic circuits produce outputs based on current inputs.
• Sequential logic circuits use memory elements (flip-flops) to store state.

Floating-Point Representation

• Floating-point numbers represent real numbers using a sign, exponent, and mantissa.
• IEEE 754 standard defines formats for single-precision (32-bit) and double-precision (64-bit) floating-point numbers.
• Operations can result in rounding errors due to the finite precision of floating-point representation.

RISC vs. CISC

• Reduced Instruction Set Computing (RISC) architectures use a small set of simple instructions.
• Complex Instruction Set Computing (CISC) architectures use a large set of complex instructions.
• RISC designs typically have simpler hardware and execute instructions faster, while CISC designs aim for code density and ease of programming (historically).
• Modern processors often incorporate features from both RISC and CISC architectures

Interconnection Networks

• Interconnection networks provide communication pathways between processors and memory modules in parallel computing systems.
• Common topologies include buses, crossbars, meshes, and hypercubes.
• Network topology affects performance, cost, and scalability.

Fault Tolerance

• Fault tolerance ensures that a computer system continues to operate correctly even in the presence of faults.
• Techniques include redundancy (hardware and software), error detection and correction codes, and fault-tolerant algorithms.
• Important for critical applications where downtime is unacceptable.