(COM OMPU PUTE TER R SCIENC IENCE) E) PROCESSING DATA - - PowerPoint PPT Presentation
, (COM OMPU PUTE TER R SCIENC IENCE) E) PROCESSING DATA
Τμήμα Ξένων Γλωσσών, Μετάφρασης και Διερμηνείας
ΠΡΟΓΡΑΜΜΑ ΠΡΟΠΤΥΧΙΑΚΩΝ ΣΠΟΥΔΩΝ
ΔΙΔΑΣΚΟΝΤΕΣ: Κωνσταντίνος Οικονόμου, Αν. Καθηγητής Γεώργιος Κουφουδάκης, Διδάκτορας Αθανάσιος Τσίπης, Υπ. Διδάκτορας Αστέριος Παπαμιχαήλ, Μεταπτυχιακός Φοιτητής
■ The “brain” of the computer ■ Place where data is manipulated ■ In average microprocessors, the entire CPU is a single unit ■ Today we find multi-core processors (i.e. dual-core, quad-core, six-core, octa-core, …) ■ Also we use the term many-core or massively multi-core – Example: Intel has produced a 48-core processor for research in cloud computing ■ In large computer systems (e.g. supercomputers, mainframes, etc.) multiple processing chips to handle data and tasks – It may consist of hundreds (or even thousands) of separate processor chips ■ All CPUs has at least two basic parts: – The Control Unit – The Arithmetic Logic Unit
■ Manages all computer’s resources ■ The “traffic signal” directing the flow of data through the CPU and to/from other devices ■ The logical hub of the computer ■ Holds all instructions for carrying out commands ■ Instructions or Instruction Set: – List of all operations the CPU can perform – Each instruction expressed by a microcode – Microcode:
■ A series of basic directions that tell CPU how to execute more complex operations
■ All data is stored as numbers ➔ Much processing involves comparing numbers or executing mathematical operations ■ Two types of operations: – Arithmetic operations
■ Addition, Subtraction, Multiplication, Division
– Logical operations
■ Comparisons: one number is equal to, greater than, less than another number ■ Every comparison has an opposite – Use of word not – Not equal to, not greater than, not less than
■ CU performs tasks such as moving data from one place to another, e.g. from memory to storage or from memory to printer ■ When an instruction demands arithmetic of logic operation CU passes it to ALU ■ ALU performs the operations listed in instruction ■ ALU includes a group of registers ■ Registers: – high speed memory locations – Used to hold data currently being processed or used for calculation by ALU
■ CU loads two numbers from memory to ALU’s registers ■ CU tells ALU to divide the two numbers (arithmetic operation)
CU tells ALU to see whether the two numbers are equal (logical operation) ■ The number/result of the above calculations is then stored in another register ■ It is then being sent out of the CPU
■ Each time CPU executes an instruction it takes a series of steps ■ The complete series of steps is called machine cycle ■ The machine cycle can be further broken down into – The instruction cycle – The execution cycle
■ At the beginning of the machine cycle ■ CPU takes two actions: – Fetching: Before CPU can execute an instruction, the CU must retrieve (or fetch) a command of data from computer’s memory – Decoding: Before a command can be executed, the CU must break down (or decode) the command into instructions corresponding to CPU’s instruction set
■ After the step of decoding the execution may begin ■ CPU takes two actions: – Executing: when CPU executes a command, it carries out the instructions in
– Storing: the CPU may be required to store the results of an action into memory
■ The process is complex ■ Nevertheless, it is accomplished at an incredible speed ■ Translation of multiple instructions every second ■ CPU is measured in millions of instructions per second (MIPS) – Newer CPUs are measured in billions of instructions per second (BIPS) – Supercomputers can perform trillions of instructions per second (TIPS)
■ Microprocessors execute instructions rapidly ■ To speed up the process newer processors use pipelining (or pipeline processing) ■ Instruction-level parallelism within a single processor ■ The CU initiates an new machine cycle before current is over ■ Instruction executions are performed in stages: – When first instruction’s fetching stage is over it moves to the decoding stage – At that point an new instruction is fetched – Think pipeline as an assembly line
■ Each instruction is broken up in parts ■ Once on part is over it is passed to the second ■ The first step in the line is now idle ■ The pipeline feeds a new step one
■ Microprocessors may execute up to 20 instructions simultaneously
Instr structi ucti
Fetch Decode Execute Store
Fetch Decode Execute Store
Fetch Decode Execute Store
Fetch Decode Execute Store
Fetch Decode Execute Store
Fetch Decode Execute Store Time Step 1 2 3 4 5 6 7 8 9
■ Support for running many programs / tasks (multitasking) ■ CPU performs tasks for more than one program ■ CPU creates threads ■ Thread: – One instruction from a program – Execution of one thread at one time – CPU performs multiple threads very quickly – The users thinks programs are executed at the same time ■ Newer processors (multi-core) support hyper-threading – Execution of multiple threads at the same time
■ Peter Norton, Introduction to Computers, Sixth Edition, pp. 130 - 131 ■ Intro ch 05_a Peter Norton http:// ://www ww.slide .slidesh share.n are.net/s t/sam1 am178 7896/i 96/intr ntro-ch ch-05a 05a-peter-nor norton