| 1 | <html> |
|---|
| 2 | <head> |
|---|
| 3 | <title> |
|---|
| 4 | A Tour of NTL: Examples: Floating Point Classes </title> |
|---|
| 5 | </head> |
|---|
| 6 | |
|---|
| 7 | <body bgcolor="#fff9e6"> |
|---|
| 8 | <center> |
|---|
| 9 | <a href="tour-ex5.html"><img src="arrow1.gif" alt="[Previous]" align=bottom></a> |
|---|
| 10 | <a href="tour-examples.html"><img src="arrow2.gif" alt="[Up]" align=bottom></a> |
|---|
| 11 | <img src="arrow3.gif" alt="[Next]" align=bottom> |
|---|
| 12 | </center> |
|---|
| 13 | |
|---|
| 14 | <h1> |
|---|
| 15 | <p align=center> |
|---|
| 16 | A Tour of NTL: Examples: Floating Point Classes |
|---|
| 17 | </p> |
|---|
| 18 | </h1> |
|---|
| 19 | |
|---|
| 20 | <p> <hr> <p> |
|---|
| 21 | |
|---|
| 22 | NTL also supports arbitrary precision floating point with |
|---|
| 23 | the class <tt>RR</tt>. |
|---|
| 24 | Additionally, it supports two specialized classes: <tt>quad_float</tt>, |
|---|
| 25 | which gives a form of quadruple precision, but without an extended |
|---|
| 26 | exponent range, |
|---|
| 27 | and <tt>xdouble</tt>, |
|---|
| 28 | which gives double precision, but with an extended exponent range. |
|---|
| 29 | The advantage of the latter two classes is efficiency. |
|---|
| 30 | |
|---|
| 31 | <p> |
|---|
| 32 | |
|---|
| 33 | Here again is a program that reads a list of numbers from the input, |
|---|
| 34 | and outputs the sum of their squares, using the class <tt>RR</tt>. |
|---|
| 35 | <p> |
|---|
| 36 | |
|---|
| 37 | <pre> |
|---|
| 38 | #include <NTL/RR.h> |
|---|
| 39 | |
|---|
| 40 | int main() |
|---|
| 41 | { |
|---|
| 42 | RR acc, val; |
|---|
| 43 | |
|---|
| 44 | acc = 0; |
|---|
| 45 | while (SkipWhiteSpace(cin)) { |
|---|
| 46 | cin >> val; |
|---|
| 47 | acc += val*val; |
|---|
| 48 | } |
|---|
| 49 | |
|---|
| 50 | cout << acc << "\n"; |
|---|
| 51 | } |
|---|
| 52 | </pre> |
|---|
| 53 | |
|---|
| 54 | <p> |
|---|
| 55 | |
|---|
| 56 | The precision used for the computation can be set by executing |
|---|
| 57 | <pre> |
|---|
| 58 | RR::SetPrecision(p); |
|---|
| 59 | </pre> |
|---|
| 60 | which sets the effective precision to <tt>p</tt> bits. |
|---|
| 61 | By default, <tt>p=150</tt>. |
|---|
| 62 | All of the basic arithmetic operations compute their results |
|---|
| 63 | by rounding to the nearest <tt>p</tt>-bit floating point number. |
|---|
| 64 | The semantics of this are exactly the same as in the IEEE floating |
|---|
| 65 | point standard (except that there are no special values, like |
|---|
| 66 | "infinity" and "not a number"). |
|---|
| 67 | |
|---|
| 68 | <p> |
|---|
| 69 | |
|---|
| 70 | The number of <i>decimal</i> digits of precision that are used when |
|---|
| 71 | printing an <tt>RR</tt> can be set be executing |
|---|
| 72 | <pre> |
|---|
| 73 | RR::SetOutputPrecision(d); |
|---|
| 74 | </pre> |
|---|
| 75 | which sets the output precision to <tt>d</tt>. |
|---|
| 76 | By default, <tt>d=10</tt>. |
|---|
| 77 | |
|---|
| 78 | <p> |
|---|
| 79 | See <a href="RR.txt"><tt>RR.txt</tt></a> for details. |
|---|
| 80 | |
|---|
| 81 | <p> |
|---|
| 82 | |
|---|
| 83 | By replacing the occurences of <tt>RR</tt> by either <tt>quad_float</tt> |
|---|
| 84 | or <tt>xdouble</tt>, one gets an equivalent program using one of the |
|---|
| 85 | other floating point classes. |
|---|
| 86 | The output precision for these two classes can be controlled just |
|---|
| 87 | as with <tt>RR</tt>. |
|---|
| 88 | See <a href="quad_float.txt"><tt>quad_float.txt</tt></a> and |
|---|
| 89 | <a href="xdouble.txt"><tt>xdouble.txt</tt></a> |
|---|
| 90 | for details. |
|---|
| 91 | |
|---|
| 92 | <p> |
|---|
| 93 | |
|---|
| 94 | |
|---|
| 95 | <center> |
|---|
| 96 | <a href="tour-ex5.html"><img src="arrow1.gif" alt="[Previous]" align=bottom></a> |
|---|
| 97 | <a href="tour-examples.html"><img src="arrow2.gif" alt="[Up]" align=bottom></a> |
|---|
| 98 | <img src="arrow3.gif" alt="[Next]" align=bottom> |
|---|
| 99 | </center> |
|---|
| 100 | |
|---|
| 101 | </body> |
|---|
| 102 | </html> |
|---|
