github.com/gopherd/gonum@v0.0.4/mathext/internal/amos/amoslib/d1mach.f (about) 1 *DECK D1MACH 2 DOUBLE PRECISION FUNCTION D1MACH(I) 3 C***BEGIN PROLOGUE D1MACH 4 C***DATE WRITTEN 750101 (YYMMDD) 5 C***REVISION DATE 890213 (YYMMDD) 6 C***CATEGORY NO. R1 7 C***KEYWORDS LIBRARY=SLATEC,TYPE=DOUBLE PRECISION(R1MACH-S D1MACH-D), 8 C MACHINE CONSTANTS 9 C***AUTHOR FOX, P. A., (BELL LABS) 10 C HALL, A. D., (BELL LABS) 11 C SCHRYER, N. L., (BELL LABS) 12 C***PURPOSE Returns double precision machine dependent constants 13 C***DESCRIPTION 14 C 15 C D1MACH can be used to obtain machine-dependent parameters 16 C for the local machine environment. It is a function 17 C subprogram with one (input) argument, and can be called 18 C as follows, for example 19 C 20 C D = D1MACH(I) 21 C 22 C where I=1,...,5. The (output) value of D above is 23 C determined by the (input) value of I. The results for 24 C various values of I are discussed below. 25 C 26 C D1MACH( 1) = B**(EMIN-1), the smallest positive magnitude. 27 C D1MACH( 2) = B**EMAX*(1 - B**(-T)), the largest magnitude. 28 C D1MACH( 3) = B**(-T), the smallest relative spacing. 29 C D1MACH( 4) = B**(1-T), the largest relative spacing. 30 C D1MACH( 5) = LOG10(B) 31 C 32 C Assume double precision numbers are represented in the T-digit, 33 C base-B form 34 C 35 C sign (B**E)*( (X(1)/B) + ... + (X(T)/B**T) ) 36 C 37 C where 0 .LE. X(I) .LT. B for I=1,...,T, 0 .LT. X(1), and 38 C EMIN .LE. E .LE. EMAX. 39 C 40 C The values of B, T, EMIN and EMAX are provided in I1MACH as 41 C follows: 42 C I1MACH(10) = B, the base. 43 C I1MACH(14) = T, the number of base-B digits. 44 C I1MACH(15) = EMIN, the smallest exponent E. 45 C I1MACH(16) = EMAX, the largest exponent E. 46 C 47 C To alter this function for a particular environment, 48 C the desired set of DATA statements should be activated by 49 C removing the C from column 1. Also, the values of 50 C D1MACH(1) - D1MACH(4) should be checked for consistency 51 C with the local operating system. 52 C 53 C***REFERENCES FOX P.A., HALL A.D., SCHRYER N.L.,*FRAMEWORK FOR A 54 C PORTABLE LIBRARY*, ACM TRANSACTIONS ON MATHEMATICAL 55 C SOFTWARE, VOL. 4, NO. 2, JUNE 1978, PP. 177-188. 56 C***ROUTINES CALLED XERROR 57 C***END PROLOGUE D1MACH 58 C 59 INTEGER SMALL(4) 60 INTEGER LARGE(4) 61 INTEGER RIGHT(4) 62 INTEGER DIVER(4) 63 INTEGER LOG10(4) 64 C 65 DOUBLE PRECISION DMACH(5) 66 SAVE DMACH 67 C 68 C EQUIVALENCE (DMACH(1),SMALL(1)) 69 C EQUIVALENCE (DMACH(2),LARGE(1)) 70 C EQUIVALENCE (DMACH(3),RIGHT(1)) 71 C EQUIVALENCE (DMACH(4),DIVER(1)) 72 C EQUIVALENCE (DMACH(5),LOG10(1)) 73 C 74 C MACHINE CONSTANTS FOR THE IBM PC 75 C ASSUMES THAT ALL ARITHMETIC IS DONE IN DOUBLE PRECISION 76 C ON 8088, I.E., NOT IN 80 BIT FORM FOR THE 8087. 77 C 78 DATA DMACH(1) / 2.23D-308 / 79 C DATA SMALL(1),SMALL(2) / 2002288515, 1050897 / 80 DATA DMACH(2) / 1.79D-308 / 81 C DATA LARGE(1),LARGE(2) / 1487780761, 2146426097 / 82 DATA DMACH(3) / 1.11D-16 / 83 C DATA RIGHT(1),RIGHT(2) / -1209488034, 1017118298 / 84 DATA DMACH(4) / 2.22D-16 / 85 C DATA DIVER(1),DIVER(2) / -1209488034, 1018166874 / 86 DATA DMACH(5) / 0.3010299956639812 / 87 C DATA LOG10(1),LOG10(2) / 1352628735, 1070810131 / 88 C 89 C 90 C***FIRST EXECUTABLE STATEMENT D1MACH 91 IF (I .LT. 1 .OR. I .GT. 5) 92 1 CALL XERROR ('D1MACH -- I OUT OF BOUNDS', 25, 1, 2) 93 C 94 D1MACH = DMACH(I) 95 RETURN 96 C 97 END