1 / 28

Numeric Representation in a Computer

Numeric Representation in a Computer. Learning Objectives Understand how numbers are stored in a computer and how the computer operates on them. Topics Numbers on a computer Precision and accuracy Numeric operators Precedence Exercises Summary. Numbers: precision and accuracy.

essien
Télécharger la présentation

Numeric Representation in a Computer

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Numeric Representation in a Computer Learning Objectives Understand how numbers are stored in a computer and how the computer operates on them. • Topics • Numbers on a computer • Precision and accuracy • Numeric operators • Precedence • Exercises • Summary AE6382 Design Computing

  2. Numbers: precision and accuracy • Low precision:  = 3.14 • High precision:  = 3.140101011 • Low accuracy:  = 3.10212 • High accuracy:  = 3.14159 • High accuracy & precision:  = 3.141592653 Good Accuracy Good Precision Good Precision Poor Accuracy Good Accuracy Poor Precision Poor Accuracy Poor Precision AE6382 Design Computing

  3. Computer Memory • Numbers and the results of numeric computations (along with other data such as text, graphics, documents, etc) must be stored somewhere in a computer. • That “somewhere” is “memory”. • Memory comes in a variety of types and speeds: • Cache – in the CPU itself (fastest) • RAM - external to the CPU (fast) • Disk - physical media, external to the CPU, r/w • CDROM - physical media, (slow) • Tape - physical media, (slowest) • Memory is measured in “bytes” (and kilobytes, megabytes, gigabytes, and terabytes.) AE6382 Design Computing

  4. Computer Memory is Varied… AE6382 Design Computing

  5. Memory in MATLAB AE6382 Design Computing

  6. Inside the Bytes • In the previous slide, we see: • What is the size of the variable “i” • What does “class” represent? • How many bytes are used to store the value? Name Size Bytes Class k 1x1 8 double array s 1x12 24 char array x 1x200 1600 double array Grand total is 213 elements using 1632 bytes AE6382 Design Computing

  7. Numbers and their Bases • Numbers we use are DECIMAL (or base 10): • Digits: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 • 123 = 1*102 + 2*101 + 3*100 • But we can always use other bases: • Octal (base 8): • Digits: 0, 1, 2, 3, 4, 5, 6, 7 • 123 = 1*82 + 2*81 + 3*80 • 1238 = 64+16+3 = 8310 • Binary (base 2): • Digits: 0, 1 • 1011 = 1*23 + 0*22 + 1*21 + 1*20 • 10112 = 8+0+2+1 = 1110 • 1238 = 001 010 011 = 10100112 • Hexidecimal (base 16): • Digits: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, F • 123 = 1*162 + 2*161 + 3*160 • 12316 = 256 + 32 + 3 = 29110 • 12316 = 0001 0010 0011 = 1001000112 AE6382 Design Computing

  8. Inside the Bytes • A byte is the smallest memory allocation available. • A byte contains 8 bits so that: • Smallest: 0 0 0 0 0 0 0 0 = 010 • Largest: 1 1 1 1 1 1 1 1 = 1*27+1*26+1*25+1*24+1*23+1*22+1*21+1*20 = 25510 (or 28-1) • Result: a single byte can be used to store an integer number ranging from 0 to 255 (256 different numbers) • If negative numbers are included, one bit must be dedicated to the sign, leaving only 7 bits for the number • Smallest: 0 • Largest: +12710 or -12810 AE6382 Design Computing

  9. Inside the Bytes • 2^8-1 (255) is not a very big number, so computers generally use multiple bytes to represent numbers: • Here is some vocabulary: • byte = 8 bits • single (precision) = 4 bytes • double (precision) = 8 bytes • quad (precision) = 16 bytes • char = 2 bytes (used to be 1 byte) • Note: • A word is the basic size of the CPU registers and for Pentium chips it is 4 bytes or 32 bits; it is 8 bytes for the new Itanium and some unix chipsets; it was 2 bytes for early Intel chips; some game consoles use 16 byte words. AE6382 Design Computing

  10. Inside the Bytes: Exercise • Let’s investigate how MATLAB stores numbers: • Try the following MATLAB commands: • format short • 2^8 • 2^8-1 %(is the answer correct?) • 2^64 • 2^64-1 % (is the answer correct?) • Question: what is the largest value of the exponent so that the answer above is correct? AE6382 Design Computing

  11. Inside the Bytes • How are fractional (floating point) numbers stored in a computer? • The IEEE 754 double format consists of three fields: • a 52-bit fraction, f • an 11-bit biased exponent, e • and a 1-bit sign, s • These fields are stored contiguously in 8 bytes (or 2 successively addressed 4-byte words): AE6382 Design Computing

  12. Inside the Bytes • Because only a finite number of bits can be used for each number, not all possible numbers can be represented: • Negative numbers less than -(2-2-52) x 21023(negative overflow) • Negative numbers greater than -2-1022(negative underflow) • Positive numbers less than 2-1022(positive underflow) • Positive numbers greater than (2-2-52) x 21023(positive overflow) • Zero (actually is a special combination of bits) AE6382 Design Computing

  13. Inside the Bytes • Others sources of error in computation: • Errors in the input data - measurement errors, errors introduced by the conversion of decimal data to binary, roundoff errors. • Roundoff errors during computation (as discussed) • Truncation errors - using approximate calculation is inevitable when computing quantities involving limits and other infinite processes on a computer • Never try to compare two floating point numbers for equality because all 16 digits would have to match perfectly… AE6382 Design Computing

  14. Describing error • Precision • The smallest difference that can be represented on the computer (help eps) • Accuracy • How close your answer is to the “actual” or “real” answer. • Recognize: • MATLAB (and other programs that use IEEE doubles) give you 15-16 “good” digits • MATLAB COMMANDS • realmin, realmax, eps (try with help) AE6382 Design Computing

  15. Inside the bytes • Back to MATLAB - what does all this mean? • We must pay attention to the math! (on any computer!) • Given a finite (limited) number of bits, almost all computations will result in numbers that can’t be represented, • Remember that the result of a floating-point computation must be ROUNDED to fit back into it’s finite representation • Always check your results - design programs to allow for independent verification of computations! AE6382 Design Computing

  16. SUMMARY • Describe memory. List different kinds of memory. • What is IEEE 754? Describe how MATLAB represents numbers. • Draw a number line and identify ranges where computers will generate errors. • Describe three potential sources of errors in computation. • Describe precision. Describe accuracy. • Describe how we can protect ourselves from computation error. AE6382 Design Computing

  17. Numeric computation in MATLAB AE6382 Design Computing

  18. Simple Math and Evaluation • Expressions are evaluated using standard algebraic hierarchy, with parenthesis overriding the normal convention. (help ops) PRECEDENCE - the order expressions are evaluated! AE6382 Design Computing

  19. Practice: Simple Calculations • What are the results of the following MATLAB expressions? • 5 ^ 2 + 3 / 2 ^ 3 – 25 • (5-3)^2 • (5-3)*(2-1)\8 • 2*pi • sin(pi/2) • Try entering the values above in MATLAB • Try entering help ops NOTEUsing spaces before and after operators is a matter of style and readability AE6382 Design Computing

  20. Variables and Names • A variable is a placeholder in memory • Variables containvalues • Variable names: • Are case sensitive: Cost, cost, COST are different • May contain up to 31 characters (more are ignored) • Must start with a letter, • May contain numbers and letters • May NOT contain punctuation except “_” • How do I view the contents of a variable? • Just type the variable name without a following “;” AE6382 Design Computing

  21. Practice: simple computations • Try the following MATLAB code: • Example 1 X = 5 ^ 2; Y = 2 * 2; Z = X * Y; Z • Example 2 Price = 19.95; Tax = 0.07; Units = 3; Cost = (Units * Price) * (1.0 + Tax ) • What about the “;”? What is Y? What is Price? AE6382 Design Computing

  22. Where are Variables Stored? • Variables are stored in a “workspace” • Workspace commands: • who – display variables • whos – display variables and sizes • clear – removes variables from the work space • help clear – display help on command “clear” • clc - clears the command window (screen) • home - move cursor to top-left of command window • Workspace is shown in the upper left MATLAB pane • Double-click on any variable to view it’s contents! AE6382 Design Computing

  23. How are numbers displayed? • Display is different from storage. • MATLAB computes with 15-16 significant digits (IEEE double), but often shows less! • The format command controls how values are displayed: • format short; format long • format short e; format long e; • format short g; format long g • format rat; format short • format compact; format loose • Try using help format AE6382 Design Computing

  24. MATLAB Built-in Functions • MATLAB offers a wealth of built-in math functions that can be quite helpful for many computational problems • Elementary MATLAB functions (help elfun) • Trigonometric functions • Exponential functions • Complex functions • Rounding and remainder functions • Specialized MATLAB functions (help specfun) • Specialized math functions • Number theoretic functions • Coordinate transformations • WE can write of own functions, also! AE6382 Design Computing

  25. Practice: Matlab Expressions • Write MATLAB expressions for the following: AE6382 Design Computing

  26. Summary • Topics • Memory • Numbers on a computer • Precision and accuracy • Numeric operators • Precedence • Writing complicated expressions. AE6382 Design Computing

  27. Lecture references online • Computer memory • http://www-und.ida.liu.se/~annsa582/tutorial/ • http://www.crisinc.addr.com/memory.html • http://www.howstuffworks.com/computer-memory.htm • Numeric representation • http://www.scri.fsu.edu/~jac/MAD3401/Backgrnd/ieee.html • http://www.math.byu.edu/~schow/work/floating_point_system.htm • http://www.cs.utah.edu/~zachary/isp/applets/FP/FP.html • http://www.research.microsoft.com/~hollasch/cgindex/coding/ieeefloat.html • http://docs.sun.com/htmlcoll/coll.648.2/iso-8859-1/NUMCOMPGD/ncg_goldberg.html • http://www.cs.unc.edu/~dm/UNC/COMP205/LECTURES/ERROR/lec23/node4.html • Standards • http://standards.ieee.org/ AE6382 Design Computing

  28. format rat format short format long clear who whos realmin realmax eps help arith help slash help ops help format MATLAB commands used AE6382 Design Computing

More Related