Computer arch
Download as PPTX, PDF1 like464 views
The document describes the Von Neumann architecture of modern computers including the central processing unit (CPU), main memory, and input/output system. It discusses how programs are stored in memory and executed by the CPU through sequential instruction processing. Machine instructions, machine languages, and common instruction types are defined. Examples of instruction execution including data transfer and arithmetic operations are provided. The document also discusses techniques for improving performance including pipelining and parallel processing.
Computer arch
- 1. Computer Architecture and Data Manipulation Chapter 3
- 2. Von Neumann Architecture • Today’s stored-program computers have the following characteristics: – Three hardware systems: • A central processing unit (CPU) – Arithmetic and Logic Unit (ALU) – Control Unit – Registers • A main memory system • An I/O system – The capacity to carry out sequential instruction processing. – A single data path between the CPU and main memory.
- 3. CPU and main memory connected via a bus
- 4. Stored Program Concept A program can be encoded as bit patterns and stored in main memory. From there, the CPU can then extract the instructions and execute them. In turn, the program to be executed can be altered easily.
- 5. Terminology • Machine instruction: An instruction (or command) encoded as a bit pattern recognizable by the CPU • Machine language: The set of all instructions recognized by a machine
- 6. Machine Language Philosophies • Reduced Instruction Set Computing (RISC) – Few, simple, efficient, and fast instructions – Examples: PowerPC from Apple/IBM/Motorola and SPARC from Sun Microsystems • Complex Instruction Set Computing (CISC) – Many, convenient, and powerful instructions – Example: Pentium from Intel
- 7. Machine Instruction Types • Data Transfer: copy data from one location to another • Arithmetic/Logic: use existing bit patterns to compute a new bit patterns • Control: direct the execution of the program
- 8. Example - Adding values stored in memory
- 9. Example - Dividing values stored in memory
- 10. Program Execution • Controlled by two special-purpose registers – Program Counter: address of next instruction – Instruction Register: current instruction • Machine Cycle – Fetch – Decode – Execute
- 12. Program Execution • Controlled by two special-purpose registers – Program counter: address of next instruction – Instruction register: current instruction • Machine Cycle – Fetch – Decode – Execute
- 13. The architecture of the machine described in Appendix C
- 14. Start of the Fetch Execute cycle All of the instructions were fetched and executed as part of the machine cycle
- 15. Performing the fetch step of the machine cycle
- 16. Performing the fetch step of the machine cycle (cont’d)
- 17. Parts of a Machine Instruction • Op-code: Specifies which operation to execute • Operand: Gives more detailed information about the operation – Interpretation of operand varies depending on op- code
- 18. The composition of an instruction for the machine in Appendix C 3 means to From register To memory address A7 store the contents 5 of a register to memory
- 19. Appendix C: A Simple Machine Language Op-code Operand Description 1 RXY LOAD reg. R from cell XY. 2 RXY LOAD reg. R with XY. 3 RXY STORE reg. R at XY. 4 0RS MOVE R to S. 5 RST ADD S and T into R. (2’s comp.) 6 RST ADD S and T into R. (floating pt.) 7 RST OR S and T into R. 8 RST AND S and T into R. 9 RST XOR S and T into R. A R0X ROTATE reg. R X times. B RXY JUMP to XY if R = reg. 0. C 000 HALT. D 0XY JUMP to XY always
- 20. Sample Machine Program • PC = 0 • Mem Address Contents 0 1506 1 1607 2 5056 3 3008 4 C000 5 0001 6 0002 7 0003 8 0000
- 21. Exercise • PC = 0 • Write a program that subtracts 1 from the value in memory address FF
- 22. Another Program – What’s it do? • PC = 0 Address Contents 0 20FF 1 2102 2 2200 3 130A 4 5223 5 5110 6 B108 7 D004 8 320A 9 C000 A 0003
- 23. Exercise • Write a program that computes the opposite of the value in memory address FF – E.g. if the value is +5 then it becomes -5
- 24. Communicating with Other Devices • Controller: An intermediary apparatus that handles communication between the computer and a device – Specialized controllers for each type of device – General purpose controllers (USB and FireWire) • Port: The point at which a device connects to a computer • Memory-mapped I/O: CPU communicates with peripheral devices as though they were memory cells
- 25. Controllers attached to a machine’s bus
- 26. A conceptual representation of memory-mapped I/O
- 27. Communicating with Other Devices (continued) • Direct memory access (DMA): Main memory access by a controller over the bus • Von Neumann Bottleneck: Insufficient bus speed impedes performance • Handshaking: The process of coordinating the transfer of data between components
- 28. Communicating with Other Devices (continued) • Parallel Communication: Several communication paths transfer bits simultaneously. • Serial Communication: Bits are transferred one after the other over a single communication path.
- 29. Data Communication Rates • Measurement units – Bps: Bits per second – Kbps: Kilo-bps (1,000 bps) – Mbps: Mega-bps (1,000,000 bps) – Gbps: Giga-bps (1,000,000,000 bps) • Bandwidth: Maximum available rate
- 30. Increasing Performance • Technologies to increase throughput: – Faster clock speed – Bigger word size – Larger cache memory – Pipelining: Overlap steps of the machine cycle
- 31. Pipelining • Why not start fetching the next instruction while we’re decoding the current instruction? • Why not decode the next instruction while we’re executing the current instruction? Time 0 1 2 3 4 5 6 7 8 9 10 11 12 Instruction 1 FFFDDDEEE Instruction 2 FFFDDDEEE Instruction 3 FFFDDDEEE … What if Instruction 1 is the JUMP to XY if R = reg. 0 instruction and we JUMP?
- 32. Increasing Performance – Parallel Processing: Use multiple processors simultaneously • SISD: No parallel processing • MIMD: Different programs, different data – Dual core, quad core • SIMD: Same program, different data – SSE, MMX