root / www / HPL_dlamch.html
Historique | Voir | Annoter | Télécharger (3,39 ko)
| 1 | 1 | equemene | <HTML>
|
|---|---|---|---|
| 2 | 1 | equemene | <HEAD>
|
| 3 | 1 | equemene | <TITLE>HPL_dlamch HPL 2.0 Library Functions September 10, 2008</TITLE> |
| 4 | 1 | equemene | </HEAD>
|
| 5 | 1 | equemene | |
| 6 | 1 | equemene | <BODY BGCOLOR="WHITE" TEXT = "#000000" LINK = "#0000ff" VLINK = "#000099" |
| 7 | 1 | equemene | ALINK = "#ffff00"> |
| 8 | 1 | equemene | |
| 9 | 1 | equemene | <H1>Name</H1> |
| 10 | 1 | equemene | <B>HPL_dlamch</B> determines machine-specific arithmetic constants. |
| 11 | 1 | equemene | |
| 12 | 1 | equemene | <H1>Synopsis</H1> |
| 13 | 1 | equemene | <CODE>#include "hpl.h"</CODE><BR><BR> |
| 14 | 1 | equemene | <CODE>double</CODE> |
| 15 | 1 | equemene | <CODE>HPL_dlamch(</CODE> |
| 16 | 1 | equemene | <CODE>const HPL_T_MACH</CODE> |
| 17 | 1 | equemene | <CODE>CMACH</CODE> |
| 18 | 1 | equemene | <CODE>);</CODE> |
| 19 | 1 | equemene | |
| 20 | 1 | equemene | <H1>Description</H1> |
| 21 | 1 | equemene | <B>HPL_dlamch</B> |
| 22 | 1 | equemene | determines machine-specific arithmetic constants such as |
| 23 | 1 | equemene | the relative machine precision (eps), the safe minimum (sfmin) such |
| 24 | 1 | equemene | that 1 / sfmin does not overflow, the base of the machine (base), the |
| 25 | 1 | equemene | precision (prec), the number of (base) digits in the mantissa (t), |
| 26 | 1 | equemene | whether rounding occurs in addition (rnd=1.0 and 0.0 otherwise), the |
| 27 | 1 | equemene | minimum exponent before (gradual) underflow (emin), the underflow |
| 28 | 1 | equemene | threshold (rmin) base**(emin-1), the largest exponent before overflow |
| 29 | 1 | equemene | (emax), the overflow threshold (rmax) (base**emax)*(1-eps). |
| 30 | 1 | equemene | |
| 31 | 1 | equemene | <H1>Arguments</H1> |
| 32 | 1 | equemene | <PRE>
|
| 33 | 1 | equemene | CMACH (local input) const HPL_T_MACH |
| 34 | 1 | equemene | Specifies the value to be returned by HPL_dlamch |
| 35 | 1 | equemene | = HPL_MACH_EPS, HPL_dlamch := eps (default) |
| 36 | 1 | equemene | = HPL_MACH_SFMIN, HPL_dlamch := sfmin |
| 37 | 1 | equemene | = HPL_MACH_BASE, HPL_dlamch := base |
| 38 | 1 | equemene | = HPL_MACH_PREC, HPL_dlamch := eps*base |
| 39 | 1 | equemene | = HPL_MACH_MLEN, HPL_dlamch := t |
| 40 | 1 | equemene | = HPL_MACH_RND, HPL_dlamch := rnd |
| 41 | 1 | equemene | = HPL_MACH_EMIN, HPL_dlamch := emin |
| 42 | 1 | equemene | = HPL_MACH_RMIN, HPL_dlamch := rmin |
| 43 | 1 | equemene | = HPL_MACH_EMAX, HPL_dlamch := emax |
| 44 | 1 | equemene | = HPL_MACH_RMAX, HPL_dlamch := rmax |
| 45 | 1 | equemene | |
| 46 | 1 | equemene | where |
| 47 | 1 | equemene | |
| 48 | 1 | equemene | eps = relative machine precision, |
| 49 | 1 | equemene | sfmin = safe minimum, |
| 50 | 1 | equemene | base = base of the machine, |
| 51 | 1 | equemene | prec = eps*base, |
| 52 | 1 | equemene | t = number of digits in the mantissa, |
| 53 | 1 | equemene | rnd = 1.0 if rounding occurs in addition, |
| 54 | 1 | equemene | emin = minimum exponent before underflow, |
| 55 | 1 | equemene | rmin = underflow threshold, |
| 56 | 1 | equemene | emax = largest exponent before overflow, |
| 57 | 1 | equemene | rmax = overflow threshold. |
| 58 | 1 | equemene | </PRE>
|
| 59 | 1 | equemene | |
| 60 | 1 | equemene | <H1>Example</H1> |
| 61 | 1 | equemene | <CODE>#include "hpl.h"</CODE><BR><BR> |
| 62 | 1 | equemene | <PRE>
|
| 63 | 1 | equemene | int main(int argc, char *argv[]) |
| 64 | 1 | equemene | {
|
| 65 | 1 | equemene | double eps; |
| 66 | 1 | equemene | eps = HPL_dlamch( HPL_MACH_EPS ); |
| 67 | 1 | equemene | printf("eps=%18.8e\n", eps);
|
| 68 | 1 | equemene | exit(0); return(0); |
| 69 | 1 | equemene | } |
| 70 | 1 | equemene | </PRE>
|
| 71 | 1 | equemene | |
| 72 | 1 | equemene | <H1>References</H1> |
| 73 | 1 | equemene | This function has been manually translated from the Fortran 77 LAPACK |
| 74 | 1 | equemene | auxiliary function dlamch.f (version 2.0 -- 1992), that was itself |
| 75 | 1 | equemene | based on the function ENVRON by Malcolm and incorporated suggestions |
| 76 | 1 | equemene | by Gentleman and Marovich. See |
| 77 | 1 | equemene | |
| 78 | 1 | equemene | Malcolm M. A., Algorithms to reveal properties of floating-point |
| 79 | 1 | equemene | arithmetic., Comms. of the ACM, 15, 949-951 (1972). |
| 80 | 1 | equemene | |
| 81 | 1 | equemene | Gentleman W. M. and Marovich S. B., More on algorithms that reveal |
| 82 | 1 | equemene | properties of floating point arithmetic units., Comms. of the ACM, |
| 83 | 1 | equemene | 17, 276-277 (1974). |
| 84 | 1 | equemene | |
| 85 | 1 | equemene | </BODY>
|
| 86 | 1 | equemene | </HTML> |