[1] | 1 | *DECK I1MACH |
---|
| 2 | INTEGER FUNCTION I1MACH (I) |
---|
| 3 | C***BEGIN PROLOGUE I1MACH |
---|
| 4 | C***PURPOSE Return integer machine dependent constants. |
---|
| 5 | C***LIBRARY SLATEC |
---|
| 6 | C***CATEGORY R1 |
---|
| 7 | C***TYPE INTEGER (I1MACH-I) |
---|
| 8 | C***KEYWORDS 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***DESCRIPTION |
---|
| 13 | C |
---|
| 14 | C I1MACH can be used to obtain machine-dependent parameters for the |
---|
| 15 | C local machine environment. It is a function subprogram with one |
---|
| 16 | C (input) argument and can be referenced as follows: |
---|
| 17 | C |
---|
| 18 | C K = I1MACH(I) |
---|
| 19 | C |
---|
| 20 | C where I=1,...,16. The (output) value of K above is determined by |
---|
| 21 | C the (input) value of I. The results for various values of I are |
---|
| 22 | C discussed below. |
---|
| 23 | C |
---|
| 24 | C I/O unit numbers: |
---|
| 25 | C I1MACH( 1) = the standard input unit. |
---|
| 26 | C I1MACH( 2) = the standard output unit. |
---|
| 27 | C I1MACH( 3) = the standard punch unit. |
---|
| 28 | C I1MACH( 4) = the standard error message unit. |
---|
| 29 | C |
---|
| 30 | C Words: |
---|
| 31 | C I1MACH( 5) = the number of bits per integer storage unit. |
---|
| 32 | C I1MACH( 6) = the number of characters per integer storage unit. |
---|
| 33 | C |
---|
| 34 | C Integers: |
---|
| 35 | C assume integers are represented in the S-digit, base-A form |
---|
| 36 | C |
---|
| 37 | C sign ( X(S-1)*A**(S-1) + ... + X(1)*A + X(0) ) |
---|
| 38 | C |
---|
| 39 | C where 0 .LE. X(I) .LT. A for I=0,...,S-1. |
---|
| 40 | C I1MACH( 7) = A, the base. |
---|
| 41 | C I1MACH( 8) = S, the number of base-A digits. |
---|
| 42 | C I1MACH( 9) = A**S - 1, the largest magnitude. |
---|
| 43 | C |
---|
| 44 | C Floating-Point Numbers: |
---|
| 45 | C Assume floating-point numbers are represented in the T-digit, |
---|
| 46 | C base-B form |
---|
| 47 | C sign (B**E)*( (X(1)/B) + ... + (X(T)/B**T) ) |
---|
| 48 | C |
---|
| 49 | C where 0 .LE. X(I) .LT. B for I=1,...,T, |
---|
| 50 | C 0 .LT. X(1), and EMIN .LE. E .LE. EMAX. |
---|
| 51 | C I1MACH(10) = B, the base. |
---|
| 52 | C |
---|
| 53 | C Single-Precision: |
---|
| 54 | C I1MACH(11) = T, the number of base-B digits. |
---|
| 55 | C I1MACH(12) = EMIN, the smallest exponent E. |
---|
| 56 | C I1MACH(13) = EMAX, the largest exponent E. |
---|
| 57 | C |
---|
| 58 | C Double-Precision: |
---|
| 59 | C I1MACH(14) = T, the number of base-B digits. |
---|
| 60 | C I1MACH(15) = EMIN, the smallest exponent E. |
---|
| 61 | C I1MACH(16) = EMAX, the largest exponent E. |
---|
| 62 | C |
---|
| 63 | C To alter this function for a particular environment, the desired |
---|
| 64 | C set of DATA statements should be activated by removing the C from |
---|
| 65 | C column 1. Also, the values of I1MACH(1) - I1MACH(4) should be |
---|
| 66 | C checked for consistency with the local operating system. |
---|
| 67 | C |
---|
| 68 | C***REFERENCES P. A. Fox, A. D. Hall and N. L. Schryer, Framework for |
---|
| 69 | C a portable library, ACM Transactions on Mathematical |
---|
| 70 | C Software 4, 2 (June 1978), pp. 177-188. |
---|
| 71 | C***ROUTINES CALLED (NONE) |
---|
| 72 | C***REVISION HISTORY (YYMMDD) |
---|
| 73 | C 750101 DATE WRITTEN |
---|
| 74 | C 891012 Added VAX G-floating constants. (WRB) |
---|
| 75 | C 891012 REVISION DATE from Version 3.2 |
---|
| 76 | C 891214 Prologue converted to Version 4.0 format. (BAB) |
---|
| 77 | C 900618 Added DEC RISC constants. (WRB) |
---|
| 78 | C 900723 Added IBM RS 6000 constants. (WRB) |
---|
| 79 | C 901009 Correct I1MACH(7) for IBM Mainframes. Should be 2 not 16. |
---|
| 80 | C (RWC) |
---|
| 81 | C 910710 Added HP 730 constants. (SMR) |
---|
| 82 | C 911114 Added Convex IEEE constants. (WRB) |
---|
| 83 | C 920121 Added SUN -r8 compiler option constants. (WRB) |
---|
| 84 | C 920229 Added Touchstone Delta i860 constants. (WRB) |
---|
| 85 | C 920501 Reformatted the REFERENCES section. (WRB) |
---|
| 86 | C 920625 Added Convex -p8 and -pd8 compiler option constants. |
---|
| 87 | C (BKS, WRB) |
---|
| 88 | C 930201 Added DEC Alpha and SGI constants. (RWC and WRB) |
---|
| 89 | C 930618 Corrected I1MACH(5) for Convex -p8 and -pd8 compiler |
---|
| 90 | C options. (DWL, RWC and WRB). |
---|
| 91 | C 100623 Use Fortran 95 intrinsic functions (Lionel GUEZ) |
---|
| 92 | C***END PROLOGUE I1MACH |
---|
| 93 | C |
---|
| 94 | INTEGER IMACH(16),OUTPUT |
---|
| 95 | SAVE IMACH |
---|
| 96 | EQUIVALENCE (IMACH(4),OUTPUT) |
---|
| 97 | C***FIRST EXECUTABLE STATEMENT I1MACH |
---|
| 98 | IMACH( 1) = 5 |
---|
| 99 | IMACH( 2) = 6 |
---|
| 100 | IMACH( 3) = 6 |
---|
| 101 | IMACH( 4) = 6 |
---|
| 102 | IMACH( 5) = bit_size(0) |
---|
| 103 | IMACH( 6) = IMACH( 5) / 8 |
---|
| 104 | IMACH( 7) = radix(0) |
---|
| 105 | IMACH( 8) = digits(0) |
---|
| 106 | IMACH( 9) = huge(0) |
---|
| 107 | IMACH(10) = radix(0.) |
---|
| 108 | IMACH(11) = digits(0.) |
---|
| 109 | IMACH(12) = minexponent(0.) |
---|
| 110 | IMACH(13) = maxexponent(0.) |
---|
| 111 | IMACH(14) = digits(0d0) |
---|
| 112 | IMACH(15) = minexponent(0d0) |
---|
| 113 | IMACH(16) = maxexponent(0d0) |
---|
| 114 | IF (I .LT. 1 .OR. I .GT. 16) GO TO 10 |
---|
| 115 | C |
---|
| 116 | I1MACH = IMACH(I) |
---|
| 117 | RETURN |
---|
| 118 | C |
---|
| 119 | 10 CONTINUE |
---|
| 120 | WRITE (UNIT = OUTPUT, FMT = 9000) |
---|
| 121 | 9000 FORMAT ('1ERROR 1 IN I1MACH - I OUT OF BOUNDS') |
---|
| 122 | C |
---|
| 123 | C CALL FDUMP |
---|
| 124 | C |
---|
| 125 | STOP |
---|
| 126 | END |
---|