root / www / HPL_pdlamch.html @ 1
Historique | Voir | Annoter | Télécharger (2,77 ko)
| 1 |
<HTML>
|
|---|---|
| 2 |
<HEAD>
|
| 3 |
<TITLE>HPL_pdlamch HPL 2.0 Library Functions September 10, 2008</TITLE> |
| 4 |
</HEAD>
|
| 5 |
|
| 6 |
<BODY BGCOLOR="WHITE" TEXT = "#000000" LINK = "#0000ff" VLINK = "#000099" |
| 7 |
ALINK = "#ffff00"> |
| 8 |
|
| 9 |
<H1>Name</H1> |
| 10 |
<B>HPL_pdlamch</B> determines machine-specific arithmetic constants. |
| 11 |
|
| 12 |
<H1>Synopsis</H1> |
| 13 |
<CODE>#include "hpl.h"</CODE><BR><BR> |
| 14 |
<CODE>double</CODE> |
| 15 |
<CODE>HPL_pdlamch(</CODE> |
| 16 |
<CODE>MPI_Comm</CODE> |
| 17 |
<CODE>COMM</CODE>, |
| 18 |
<CODE>const HPL_T_MACH</CODE> |
| 19 |
<CODE>CMACH</CODE> |
| 20 |
<CODE>);</CODE> |
| 21 |
|
| 22 |
<H1>Description</H1> |
| 23 |
<B>HPL_pdlamch</B> |
| 24 |
determines machine-specific arithmetic constants such as |
| 25 |
the relative machine precision (eps), the safe minimum(sfmin) such that |
| 26 |
1/sfmin does not overflow, the base of the machine (base), the precision |
| 27 |
(prec), the number of (base) digits in the mantissa (t), whether |
| 28 |
rounding occurs in addition (rnd = 1.0 and 0.0 otherwise), the minimum |
| 29 |
exponent before (gradual) underflow (emin), the underflow threshold |
| 30 |
(rmin)- base**(emin-1), the largest exponent before overflow (emax), the |
| 31 |
overflow threshold (rmax) - (base**emax)*(1-eps). |
| 32 |
|
| 33 |
<H1>Arguments</H1> |
| 34 |
<PRE>
|
| 35 |
COMM (global/local input) MPI_Comm |
| 36 |
The MPI communicator identifying the process collection. |
| 37 |
</PRE>
|
| 38 |
<PRE>
|
| 39 |
CMACH (global input) const HPL_T_MACH |
| 40 |
Specifies the value to be returned by HPL_pdlamch |
| 41 |
= HPL_MACH_EPS, HPL_pdlamch := eps (default) |
| 42 |
= HPL_MACH_SFMIN, HPL_pdlamch := sfmin |
| 43 |
= HPL_MACH_BASE, HPL_pdlamch := base |
| 44 |
= HPL_MACH_PREC, HPL_pdlamch := eps*base |
| 45 |
= HPL_MACH_MLEN, HPL_pdlamch := t |
| 46 |
= HPL_MACH_RND, HPL_pdlamch := rnd |
| 47 |
= HPL_MACH_EMIN, HPL_pdlamch := emin |
| 48 |
= HPL_MACH_RMIN, HPL_pdlamch := rmin |
| 49 |
= HPL_MACH_EMAX, HPL_pdlamch := emax |
| 50 |
= HPL_MACH_RMAX, HPL_pdlamch := rmax |
| 51 |
|
| 52 |
where |
| 53 |
|
| 54 |
eps = relative machine precision, |
| 55 |
sfmin = safe minimum, |
| 56 |
base = base of the machine, |
| 57 |
prec = eps*base, |
| 58 |
t = number of digits in the mantissa, |
| 59 |
rnd = 1.0 if rounding occurs in addition, |
| 60 |
emin = minimum exponent before underflow, |
| 61 |
rmin = underflow threshold, |
| 62 |
emax = largest exponent before overflow, |
| 63 |
rmax = overflow threshold. |
| 64 |
</PRE>
|
| 65 |
|
| 66 |
</BODY>
|
| 67 |
</HTML>
|