subroutine aptvuna (ax, ay, az, np, tol, vlen, nerr) ccbeg. cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc c c SUBROUTINE APTVUNA c c call aptvuna (ax, ay, az, np, tol, vlen, nerr) c c Version: aptvuna Updated 1990 November 26 10:00. c aptvuna Originated 1989 November 2 14:10. c c Author: Arthur L. Edwards, LLNL, L-298, Telephone (925) 422-4123. c c c Purpose: To find the np unit vectors a = (ax, ay, az) parallel to the c np initial vectors a = (ax, ay, az). Any components of the c initial vector "a" no greater than tol, or no greater than tol c times the initial length of "a", will be truncated to zero. c If all are zero, or are truncated to zero, vlen will be zero. c Flag nerr indicates any input error. c c With no truncation, c (ax, ay, az) = (ax, ay, az) / sqrt (ax**2 + ay**2 + az**2). c c History: 1990 March 14. Modified to always return a unit vector. c 1990 March 21. Deleted change of 1990 March 14. c c Input: ax, ay, az, np, tol. c c Output: ax, ay, az, vlen, nerr. c c Glossary: c c ax,ay,az Input The x, y, z components of input vector "a". Size np. c Will be truncated to zero if initially no greater c than tol, or no greater than tol times the initial c length of "a". c c ax,ay,az Output The x, y, z components of unit vector "a". Size np. c c nerr Output Indicates an input error, it not 0. c 1 if np is not positive. c c np Input Size of arrays ax, ay, az, vlen. c c tol Input Numerical tolerance limit. c On Cray computers, recommend 1.e-5 to 1.e-11. c c vlen Output Magnitude of the input vector "a", after any c truncation of components has been done, but before c division by vlen to form a unit vector. Size np. c cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc ccend. c.... Dimensioned arguments. c---- Component x of vector "a". dimension ax (1) c---- Component y of vector "a". dimension ay (1) c---- Component z of vector "a". dimension az (1) c---- Magnitude of input vector "a". dimension vlen (1) c.... Local variables. c---- Square of estimated error in "a". common /laptvuna/ aerr2 c---- A very small number. common /laptvuna/ fuz c---- Index, 1 to np. common /laptvuna/ n cbugc***DEBUG begins. cbug 9901 format (/ 'aptvuna finding unit vectors with tol=',1pe13.5) cbug 9902 format (i3,' ax,ay,az=',1p3e22.14) cbug write ( 3, 9901) tol cbug write ( 3, 9902) (n, ax(n), ay(n), az(n), n = 1, np) cbugc***DEBUG ends. c.... Initialize. c---- A very small number. fuz = 1.e-99 nerr = 0 c.... Test for input errors. if (np .le. 0) then nerr = 1 go to 210 endif c.... Test for the truncation option. c---- Truncate small components to zero. if (tol .gt. 0.0) then c---- Loop over vectors. do 110 n = 1, np aerr2 = tol**2 * amax1 (1.0, ax(n)**2 + ay(n)**2 + az(n)**2) if (ax(n)**2 .lt. aerr2) then ax(n) = 0.0 endif if (ay(n)**2 .lt. aerr2) then ay(n) = 0.0 endif if (az(n)**2 .lt. aerr2) then az(n) = 0.0 endif c---- End of loop over vectors. 110 continue c---- Tested tol. endif c.... Find the unit vectors. c---- Loop over vectors. do 120 n = 1, np vlen(n) = sqrt (ax(n)**2 + ay(n)**2 + az(n)**2) ax(n) = ax(n) / (vlen(n) + fuz) ay(n) = ay(n) / (vlen(n) + fuz) az(n) = az(n) / (vlen(n) + fuz) c---- End of loop over vectors. 120 continue cbugc***DEBUG begins. cbug 9903 format (/ 'aptvuna results:' / cbug & (i3,' vlen= ',1pe22.14 / cbug & ' ux,uy,uz=',1p3e22.14)) cbug write ( 3, 9903) (n, vlen(n), ax(n), ay(n), az(n), n = 1, np) cbugc***DEBUG ends. 210 return c.... End of subroutine aptvuna. (+1 line.) end UCRL-WEB-209832