Computer Architecture ECE 4801 Berk Sunar
Outline Brief Overview How is a computer program executed? Computer organization Roadmap for this class
Things You Learn in this Course How computers work; the basic foundation How to analyze their performance (and how not to) Key technologies determining the performance of modern processors Datapath Design Pipelining Cache Systems Memory Hierarchy I/O Multiprocessors
Instruction Set Architecture Important abstraction Interface between hardware and low-level software Or features available to programmers instructions set architecture (ISA) e.g. does the processor have an multiply instruction? instruction encoding Data representation I/O mechanism. addressing mechanism Modern instruction set architectures: 80x86/Pentium/K6, PowerPC, DEC Alpha, MIPS, SPARC, HP, ARM.
Computer Organization Computer Organization is how features are implemented in hardware Transparent to programmers Different implementations are possible for the same architecture (affects performance/price) Determines how memory, CPU, peripherals, busses are interconnected and how control signals routed. Has HUGE impact on performance. Performance of the organization is usually application dependent. (e.g. I/O intensive, computation intensive, memory bound etc.)
How to Program a Computer? A simple but universal interface Machine Code (binary images) Assembly language Uses mnemonics that map directly to ISA e.g. addw, lb, jmp etc. More readable than machine languages Error prone but excellent for low-level optimization High-level languages E.g. C/C++, Pascal, Fortran, Java, C# Much easier to use and program Promotes code portability Not as efficient as custom assembly
Processing a C Program High-level language program (in C) swap: muli $2, $5, 4 add $2, $4, $2 lw $15, 0($2) lw $16, 4($2) sw $16, 0($2) sw $15, 4($2) jr $31 Assembly language program for MIPS swap (int v[], int k){ int temp; temp = v[k]; v[k] = v[k+1]; v[k+1] = temp; } C compiler 00000000101000010000000000011000000000001000111000011000001000011000110001100010000000000000000010001100111100100000000000000100101011001111001000000000000000001010110001100010000000000000010000000011111000000000000000001000 Binary machine language program for MIPS Assembler
Functions of a Computer Data processing , e.g. sort entries of a spreadsheet Data storage, e.g. personal files, applications, movies, music etc. Data movement, e.g. play a music file, display a picture Control, (applies to all examples above)
Functions of a Computer source & destination of data data movements apparatus Control mechanism Data processing facility Data storage facility
Five Classic Components Computer Processor Memory Input Datapath Control Output System Interconnection
Motherboard USB 2.0 SIMM Sockets Processor PCI Card Slots IDE Sound Parallel/Serial PS/2 connectors SIMM Sockets Processor PCI Card Slots IDE Connectors
Inside the Processor Chip Instruction Cache Control branch prediction Data Cache Bus integer datapath floating-point datapath
CPU Cache Memory Registers ALU Control Unit Internal CPU interconnection CPU
Memory Nonvolatile: Volatile How about solid state drives? ROM Hard disk, floppy disk, magnetic tape, CDROM, USB Memory Flash memory Volatile DRAM used usually for main memory SRAM used mainly for on-chip memory such as register and cache DRAM is much cheaper than SRAM SRAM is much faster than DRAM How about solid state drives?
DRAM and Processor Characteristics
Solutions to Memory Problems Increase number of bits retrieved at one time Make DRAM “wider” rather than “deeper” Change DRAM interface Cache Reduce frequency of memory access More complex cache and cache on chip Increase interconnection bandwidth High speed buses Hierarchy of buses
Computer Networks Very essential aspect of computer systems Communication Resource sharing Remote access Ethernet is the most popular LAN Range is limited to 1 kilometer 10/100 Mbit/s Wide Area Networks (WAN) Cross continents and backbone of the Internet
Roadmap Performance issues Instruction set of MIPS Arithmetic and ALU Constructing a processor to execute our instructions (datapath design) Pipelining Memory hierarchy: caches and virtual memory I/O