PGI Fortran Reference manual - The Portland Group

3F Functions and VAX Subroutinesthan or equal to 32 (less than or equal to 64 if the source or destination is INTEGER*8). It has theform:CALL MVBITS(src, a2, a3, dest, a5)Where:srcis an integer variable or array element that represents the source location.a2is an integer expression that identifies the first position in the field transferred from src.a3is an integer expression that identifies the length of the field transferred from src.destis an integer variable or array element that represents the destination location.a5is an integer expression that identifies the starting position within a4, for the bits beingtransferred.RANThe RAN subroutine returns the next number from a sequence of pseudo-random numbers ofuniform distribution over the range 0 to 1. The result is a floating point number that is uniformlydistributed in the range between 0.0 and 1.0 exclusive. It has the form:y = RAN(i)where y is set equal to the value associated by the function with the seed argument i. Theargument i must be an INTEGER*4 variable or INTEGER*4 array element.The argument i should initially be set to a large, odd integer value. The RAN function stores avalue in the argument that it later uses to calculate the next random number.There are no restrictions on the seed, although it should be initialized with different values onseparate runs in order to obtain different random numbers. The seed is updated automatically, andRAN uses the following algorithm to update the seed passed as the parameter:SEED = 6969 * SEED + 1 ! MOD2**32The value of SEED is a 32-bit number whose high-order 24 bits are converted to floating pointand returned as the result.If the command-line option to treat all REAL declarations as DOUBLE PRECISION declarationsis in effect, RAN returns a DOUBLE PRECISION value.SECNDSThe SECNDS subroutine provides system time of day, or elapsed time, as a floating point valuein seconds. It has the form:y = SECNDS(x)PGI Fortran Reference Guide 214