Computer function-and-interconnection 3

Top Level View of Computer Function and Interconnection

Program ConceptHardwired systems are inflexibleGeneral purpose hardware can do different

tasks, given correct control signalsInstead of re-wiring, supply a new set of

control signals

What is a program?A sequence of stepsFor each step, an arithmetic or logical

operation is doneFor each operation, a different set of control

signals is needed

Function of Control UnitFor each operation a unique code is provided

e.g. ADD, MOVEA hardware segment accepts the code and

issues the control signals

We have a computer!

ComponentsThe Control Unit and the Arithmetic and

Logic Unit constitute the Central Processing Unit

Data and instructions need to get into the system and results outInput/output

Temporary storage of code and results is neededMain memory

Computer Components:Top Level View

Instruction Cycle (Processing for a single instruction)Two steps:

FetchExecuteHalt if machine is turned off, unrecoverable

error

Fetch CycleProgram Counter (PC) holds address of next

instruction to fetchProcessor fetches instruction from memory

location pointed to by PCIncrement PC

Unless told otherwiseInstruction loaded into Instruction Register

(IR)Processor interprets instruction and performs

required actions

Execute CycleProcessor-memory

Data transfer between CPU and main memoryProcessor I/O

Data transfer between CPU and I/O moduleData processing

Some arithmetic or logical operation on dataControl

Alteration of sequence of operationse.g. jump

Combination of above

Example of Program Execution

Instruction Cycle State Diagram

InterruptsMechanism by which other modules (e.g. I/O)

may interrupt normal sequence of processingProgram

e.g. overflow, division by zeroTimer

Generated by internal processor timerUsed in pre-emptive multi-tasking

I/Ofrom I/O controller

Task completion, variety of errorsHardware failure

e.g. Power failure, memory parity error

InterruptsA way to improve the processing efficiency Slow peripherals can not comprehend speedy processor

Processor remains in stall state until device catches upDifferent tasks interleaved with WRITE

Without interrupts, program would wait for IO to complete

May Periodically pole the device.With interrupts, user can execute other instructions

while IO is performedOS and Processor manage the suspension and

resumption of program

Program Flow Control

Interrupt CycleAdded to instruction cycleProcessor checks for interrupt

Indicated by an interrupt signalIf no interrupt, fetch next instructionIf interrupt pending:

Suspend execution of current program Save context (address of next instruction to be

executed)Set PC to start address of interrupt handler routineProcess interruptRestore context and continue interrupted program

In the Interrupt handler routineFetch interrupt handler instructions from

memoryInterrupt handler routines are part of OS Extra instructions are executed but still save

processing power

Transfer of Control via Interrupts

Instruction Cycle with Interrupts

Hanggan d2 lng report ko mwaah

Program TimingShort I/O Wait

Program TimingLong I/O Wait

Instruction Cycle (with Interrupts) - State Diagram

Multiple Interrupts-IWe may have more than one interrupts

A program receiving data from communication line and printing data to a printing device

Option to handle multiple interrupts

Multiple Interrupts-IIDisable interrupts

Processor will ignore further interrupts whilst processing one interrupt

Interrupts remain pending and are checked after first interrupt has been processed

Interrupts handled in sequence as they occur (No Priority task) Data arriving from communication line should be absorbed

immediatelyDefine priorities

Low priority interrupts can be interrupted by higher priority interrupts

When higher priority interrupt has been processed, processor returns to previous interrupt E.g. Printer 2, disk 4, communication line 5

Multiple Interrupts - Sequential

Multiple Interrupts – Nested

Time Sequence of Multiple Interrupts

ConnectingAll the units must be connectedDifferent type of connection for different type

of unitMemoryInput/OutputCPU

Collection of paths connecting various structures are called interconnection structures.

Computer Modules

Memory ConnectionReceives and sends dataReceives addresses (of locations)Receives control signals

ReadWriteTiming

Input/Output Connection(1)Similar to memory from computer’s

viewpointOutput

Receive data from computerSend data to peripheral

InputReceive data from peripheralSend data to computer

Input/Output Connection(2)Receive control signals from computerSend control signals to peripherals

e.g. spin diskReceive addresses from computer

e.g. port number to identify peripheralSend interrupt signals (control)

CPU ConnectionReads instruction and dataWrites out data (after processing)Sends control signals to other unitsReceives (& acts on) interrupts

BusesThere are a number of possible

interconnection systemsSingle and multiple BUS structures are most

commone.g. Control/Address/Data bus (PC)e.g. Unibus (DEC-PDP)

What is a Bus?A communication pathway connecting two or

more devicesOften grouped

A number of channels in one buse.g. 32 bit data bus is 32 separate single bit

channelsPower lines may not be shown

Data BusCarries data

Remember that there is no difference between “data” and “instruction” at this level

Width is a key determinant of performance8, 16, 32, 64 bit

Address busIdentify the source or destination of datae.g. CPU needs to read an instruction (data)

from a given location in memoryBus width determines maximum memory

capacity of systeme.g. 8080 has 16 bit address bus giving 64k

address space

Control BusControl and timing information

Memory read/write signalIO Read/WriteTransfer Ack.Interrupt request (indicates interrupt is pending)Interrupt Ack. (pending intrrupt acknowledged)Sending and requesting data requires:

Request use of bus Transfer of data via bus

Bus Interconnection Scheme

ArrangementWhat do buses look like?

Parallel lines on circuit boardsRibbon cablesStrip connectors on mother boardsOn Chip and board wires

Physical Realization of Bus Architecture

Single Bus ProblemsLots of devices on one bus leads to:

Propagation delays Long data paths mean that co-ordination of bus use

can adversely affect performance Bus may become a bottleneck as aggregate

aggregate data transfer approaches bus capacity Increase the data rate (32-bit,64-bit) Still growing application demands can t be met

Most systems use multiple buses to overcome these problems

•Devices with high speed demand are closer to processor•But independent of processor ( Processor’s Architectural changes have no affect on high speed bus)

Bus TypesDedicated

Separate data & address linesMultiplexed (time multiplexing)

Shared linesAddress valid or data valid control lineAdvantage - fewer lines, cost benefit Disadvantages

More complex control Ultimate performance

Physical Dedication (IO bus connects IO modules only)

Bus ArbitrationMore than one module controlling the buse.g. CPU and DMA controllerOnly one module may control bus at one timeArbitration may be centralised or distributed

Centralised or Distributed ArbitrationCentralised

Single hardware device controlling bus access Bus Controller Arbiter

DistributedEach module may claim the busControl logic on all modules

TimingCo-ordination of events on busSynchronous

Events determined by clock signalsControl Bus includes clock lineA single 1-0 is a bus cycleAll devices can read clock lineUsually sync on leading edgeUsually a single cycle for an event

PCI BusPeripheral Component InterconnectionIntel released to public domain32 or 64 bit50 lines

PCI Bus Lines (required)Systems lines

Including clock and resetAddress & Data

32 time mux lines for address/dataInterrupt & validate lines

Interface ControlArbitration

Not sharedDirect connection to PCI bus arbiter

Error lines

PCI Bus Lines (Optional)Interrupt lines

Not sharedCache support64-bit Bus Extension

Additional 32 linesTime multiplexed2 lines to enable devices to agree to use 64-bit

transferJTAG/Boundary Scan

For testing procedures

PCI CommandsTransaction between initiator (master) and

targetMaster claims busDetermine type of transaction

e.g. I/O read/writeAddress phaseOne or more data phases

Foreground ReadingStallings, chapter 3 (all of it)www.pcguide.com/ref/mbsys/buses/

In fact, read the whole site!www.pcguide.com/