Presentation is loading. Please wait.

Presentation is loading. Please wait.

Computer Systems 1 Fundamentals of Computing Performance, Data & Jargon.

Similar presentations


Presentation on theme: "Computer Systems 1 Fundamentals of Computing Performance, Data & Jargon."— Presentation transcript:

1 Computer Systems 1 Fundamentals of Computing Performance, Data & Jargon

2 Computer Systems 1 (2004-2005) Performance, Data & Jargon Performance Terms Operational CPU Specific Data Representation Analogue Vs Digital ASCII Unicode Data Reduction Jargon Deciphering some computing terminology

3 Computer Systems 1 (2004-2005) Performance Terms FLOPS FLoating Point Operations Per Second How many fractional calculations can processor carry out? Processors often have a FPU (Floating Point Unit) designed to do these calculations  Improve number intensive tasks  Such as graphics rendering  Also called maths-coprocessor Another way to measure processor speed MFLOPS = MegaFLOPS (MillionFLOPS /s) GFLOPS = GigaFLOPS (BillionFLOPS /s) Thought to be inaccurate Some operations can take longer than others

4 Computer Systems 1 (2004-2005) Performance Terms ‘Benchmark’ The measuring stick against which something is measured Tests which have been devised to measure a specific type of equipment or program Used to measure and compare performance of computer software of hardware ‘Overclock(ing)’ Processes happen in sync with clock cycle Theory: Quicker Clock = Quicker Computer Forcing a CPU to run faster than it’s approved and tested speed Used to force a little more power out of the processor Risks attached Overheating of CPU Processing errors

5 Computer Systems 1 (2004-2005) CPU Types / Performance CISC Complex Instruction Set Computer Supports up to 200 instructions Instructions for complex operations Original Pentium processors & x86 RISC Reduced Instruction Set Computer Less instructions than CISC Less Instructions = Faster (?) Cheaper to produce Puts bigger burden on software (?) Motorola processors found in Apple computers Pentium 2, 3 & 4 Use CRISC Combination of CISC & RISC

6 Computer Systems 1 (2004-2005) CPU Performance ‘Pipelining’ Technique used in modern processors CPU fetches next instruction while processing current instruction  A bit like a factory assembly line Pipeline split into segments  Each segment executes a specific function in sync with other segments then data is moved along  Sync is produce by clock cycle Pipeline Segments  Fetch  Decode  Execute Can reduce bottleneck restrictions

7 Computer Systems 1 (2004-2005) CPU Performance ‘Pipelining’ Example: Four instructions to carry out  1, 2, 3 & 4 Pipeline Segments  Fetch (F), Decode (D), Execute (E) Clock Cycle F(1)F(2) D(1) F(3) D(2) E(1) F(4) D(3) E(2) D(4) E(3)E(4)

8 Computer Systems 1 (2004-2005) CPU Performance ‘Hyper-Threading’ Hyper-Threading Technology (HT Technology) is a groundbreaking technology that enables a processor to execute two threads in parallel — allowing you and your software to multi-task more effectively than ever before. (Intel)

9 Computer Systems 1 (2004-2005) CPU Performance ‘Hyper-Threading’ Based on the Simultaneous Multi-Threading (SMT) method Parallel execution of process threads Allows programs to run multiple threads in parallel on one processor A Thread can be thought of as a single task that makes up a process Involves creating two logical processors System becomes a pretend dual-processor machine  A logical processor can be multi-tasking  Provides multi-tasking through time-slicing  Is a theoretical parallel processing system  Fools the OS into thinking there are two processors

10 Computer Systems 1 (2004-2005) Problem Prevention / Solving UPS Un-interruptible Power Supply Battery backup System Can also perform power conditioning and surge protection ‘Cleans’ power supply POST Power On Self Test Checks performed by the BIOS when computer is powered up Beep codes Also have POST diagnostic cards

11 Computer Systems 1 (2004-2005) Analogue A continuously varying signal Varying in terms of frequency, amplitude, or both A signal that is constantly changing Analogue can represent many values

12 Computer Systems 1 (2004-2005) Digital A signal with discrete value changes Signal levels are either on or off Often thought of as either 1 or 0, especially in computers, but digital can represent many different values using discrete levels

13 Computer Systems 1 (2004-2005) Basic Data Representation ASCII American Standard Code for Information Interchange Represents English language characters using numbers (0 to 127) 7 bits for each character E.g- 115 is a small ‘s’ Extended ASCII (uses 8 bits = more characters) http://www.asciitable.com/

14 Computer Systems 1 (2004-2005) Basic Data Representation Unicode Another standard for representing language characters Uses 16 bits for each character > 65,000 unique characters Useful for complex languages (Greek, Japanese, etc.) special characters, symbols, etc. May ultimately replace ASCII www.unicode.org

15 Computer Systems 1 (2004-2005) Data Reduction ‘Compression’ Using an algorithm to reduce the size of a computer file or program Loss-less compression Information is compacted without destroying or removing any data from the file E.g- zipping a file Lossy compression Information is compacted but information deemed to be insignificant or redundant is thrown away E.g- MP3

16 Computer Systems 1 (2004-2005) Jargon Busting AI Artificial Intelligence Attempting to make the computer behave like a human or think for itself Used in games to simulate human and random behaviour Expert systems or knowledge bases allow decisions or recommendations to be given based on input Meta Information about information Thing of Meta as meaning “about” Meta-data = data about data Meta-language = a language about a language

17 Computer Systems 1 (2004-2005) Jargon Busting HCI Human Computer Interaction How the user and computer communicate and interact with each other Make interaction faster, easier and better Many different forms of interaction: Haptic: movement, gestures, physical interaction Auditory: using sounds to interact Pervasive Computing “Ubiquitous Computing” Integration of computing technology to everyday life Introducing technology into everyday ‘ordinary’ objects Allowing everything to be networked and live

18 Computer Systems 1 (2004-2005) CS1: Week 9 What you know now: Performance Hyper-threading Pipelining Problem solving Data Representation Analogue Vs Digital ASCII Unicode Compression Jargon AI HCI other jargon...


Download ppt "Computer Systems 1 Fundamentals of Computing Performance, Data & Jargon."

Similar presentations


Ads by Google