home *** CD-ROM | disk | FTP | other *** search
- EPSILON(3I) Last changed: 4-13-99
-
-
- NNAAMMEE
- EEPPSSIILLOONN - Returns _r-1, where _r is the smallest number greater than 1
- in the numeric model
-
- SSYYNNOOPPSSIISS
- EEPPSSIILLOONN (([XX==]_x))
-
- IIMMPPLLEEMMEENNTTAATTIIOONN
- UNICOS, UNICOS/mk, IRIX systems
-
- CF90, MIPSpro 7 Fortran 90
-
- SSTTAANNDDAARRDDSS
- Fortran
-
- DDEESSCCRRIIPPTTIIOONN
- The EEPPSSIILLOONN intrinsic function returns a positive model number that is
- almost negligible compared to unity in the model representing numbers
- of the same type and kind type parameter as the argument. It accepts
- the following argument:
-
- _x Must be of type real. It can be scalar or array valued.
-
- EEPPSSIILLOONN is an inquiry function. The name of this intrinsic cannot be
- passed as an argument.
-
- RREETTUURRNN VVAALLUUEESS
- The result type is a scalar of the same type and kind type parameter
- as _x. The result has the value
- (1-_p)
- _b where _b and _p are as defined in the model for real numbers
- representing numbers of the same type and kind type parameter as _x.
- For information on the real number model, see the MMOODDEELLSS(3I) man page.
-
- EEXXAAMMPPLLEESS
- The following code fragment was run on several systems:
-
- REAL(KIND=4) r4
- REAL(KIND=8) r8
- REAL(KIND=16) r16
- REAL rd
- PRINT *,'EPSILON r4=',EPSILON(r4)
- PRINT *,'EPSILON r8=',EPSILON(r8)
- PRINT *,'EPSILON r16=',EPSILON(r16)
- PRINT *,'EPSILON rd=',EPSILON(rd)
- END
-
- The results of this test case on IRIX systems are as follows:
-
- EPSILON r4= 1.192092896E-7
- EPSILON r8= 2.22044604925031308E-16
- EPSILON r16= 1.23259516440783094595582588325435348E-32
- EPSILON rd= 1.192092896E-7
-
- The results of this test case on CRAY T90 systems that support IEEE
- floating-point arithmetic are as follows:
-
- EPSILON r4= 1.192092896E-7
- EPSILON r8= 2.22044604925031308E-16
- EPSILON r16= 1.92592994438723585305597794258492732E-34
- EPSILON rd= 2.22044604925031308E-16
-
- The results of this test case on UNICOS systems, except for CRAY T90
- systems that support IEEE floating-point arithmetic, are as follows:
-
- EPSILON r4= 1.192092895507812E-7
- EPSILON r8= 1.4210854715202E-14
- EPSILON r16= 5.04870979341447555463506281781E-29
- EPSILON rd= 1.4210854715202E-14
-
- The results of this test case on UNICOS/mk systems are as follows:
-
- EPSILON r4= 1.192092896E-7
- EPSILON r8= 2.22044604925031308E-16
- EPSILON r16= 2.22044604925031308E-16
- EPSILON rd= 2.22044604925031308E-16
-
- SSEEEE AALLSSOO
- MMOODDEELLSS(3I)
-
- _I_n_t_r_i_n_s_i_c _P_r_o_c_e_d_u_r_e_s _R_e_f_e_r_e_n_c_e _M_a_n_u_a_l for the printed version of this
- man page.
-