Commit f61865f4638158170082dca249903e6f846da6be
1 parent
25c42432dc
Exists in
master
and in
3 other branches
git-svn-id: https://lxsd.femto-st.fr/svn/fvn@14 b657c933-2333-4658-acf2-d3c7c2708721
Showing 1 changed file with 195 additions and 1 deletions Side-by-side Diff
fvnlib.f90
| ... | ... | @@ -22,7 +22,8 @@ |
| 22 | 22 | ! if you find it usefull it would be kind to give credits ;-) Nevertheless, you |
| 23 | 23 | ! may give credits to quadpack authors. |
| 24 | 24 | ! |
| 25 | -! Version 1.1 | |
| 25 | +! svn version | |
| 26 | +! June 2007 : added some complex trigonometric functions | |
| 26 | 27 | ! |
| 27 | 28 | ! TO DO LIST : |
| 28 | 29 | ! + Order eigenvalues and vectors in decreasing eigenvalue's modulus order -> atm |
| ... | ... | @@ -36,6 +37,11 @@ |
| 36 | 37 | !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! |
| 37 | 38 | |
| 38 | 39 | implicit none |
| 40 | +! We define pi and i for the module | |
| 41 | +real(kind=8),parameter :: fvn_pi = 3.141592653589793_8 | |
| 42 | +complex(kind=8),parameter :: fvn_i = (0._8,1._8) | |
| 43 | + | |
| 44 | + | |
| 39 | 45 | ! All quadpack routines are private to the module |
| 40 | 46 | private :: d1mach,dqag,dqag_2d_inner,dqag_2d_outer,dqage,dqage_2d_inner, & |
| 41 | 47 | dqage_2d_outer,dqk15,dqk15_2d_inner,dqk15_2d_outer,dqk21,dqk21_2d_inner,dqk21_2d_outer, & |
| ... | ... | @@ -1772,6 +1778,194 @@ |
| 1772 | 1778 | include "fvn_quadpack/dqk21_2d_inner.f" |
| 1773 | 1779 | include "fvn_quadpack/dqk15_2d_outer.f" |
| 1774 | 1780 | |
| 1781 | +!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! | |
| 1782 | +! | |
| 1783 | +! Trigonometric functions | |
| 1784 | +! | |
| 1785 | +! fvn_z_acos, fvn_z_asin : complex arc cosine and sine | |
| 1786 | +! fvn_d_acosh : arc cosinus hyperbolic | |
| 1787 | +! fvn_d_asinh : arc sinus hyperbolic | |
| 1788 | +! | |
| 1789 | +!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! | |
| 1790 | +function fvn_z_acos(z) | |
| 1791 | + ! double complex arccos function adapted from | |
| 1792 | + ! the c gsl library | |
| 1793 | + ! http://www.gnu.org/software/gsl/ | |
| 1794 | + implicit none | |
| 1795 | + complex(kind=8) :: fvn_z_acos | |
| 1796 | + complex(kind=8) :: z | |
| 1797 | + real(kind=8) :: rz,iz,x,y,a,b,y2,r,s,d,apx,am1 | |
| 1798 | + real(kind=8),parameter :: a_crossover=1.5_8,b_crossover = 0.6417_8 | |
| 1799 | + complex(kind=8),parameter :: i=(0._8,1._8) | |
| 1800 | + real(kind=8) :: r_res,i_res | |
| 1801 | + | |
| 1802 | + rz=dble(z) | |
| 1803 | + iz=aimag(z) | |
| 1804 | + if ( iz == 0._8 ) then | |
| 1805 | + fvn_z_acos=fvn_z_acos_real(rz) | |
| 1806 | + return | |
| 1807 | + end if | |
| 1808 | + | |
| 1809 | + x=dabs(rz) | |
| 1810 | + y=dabs(iz) | |
| 1811 | + r=fvn_d_hypot(x+1.,y) | |
| 1812 | + s=fvn_d_hypot(x-1.,y) | |
| 1813 | + a=0.5*(r + s) | |
| 1814 | + b=x/a | |
| 1815 | + y2=y*y | |
| 1816 | + | |
| 1817 | + if (b <= b_crossover) then | |
| 1818 | + r_res=dacos(b) | |
| 1819 | + else | |
| 1820 | + if (x <= 1.) then | |
| 1821 | + d=0.5*(a+x)*(y2/(r+x+1)+(s + (1 - x))) | |
| 1822 | + r_res=datan(dsqrt(d)/x) | |
| 1823 | + else | |
| 1824 | + apx=a+x | |
| 1825 | + d=0.5*(apx/(r+x+1)+apx/(s + (x - 1))) | |
| 1826 | + r_res=datan((y*dsqrt(d))/x); | |
| 1827 | + end if | |
| 1828 | + end if | |
| 1829 | + | |
| 1830 | + if (a <= a_crossover) then | |
| 1831 | + if (x < 1.) then | |
| 1832 | + am1=0.5*(y2 / (r + (x + 1)) + y2 / (s + (1 - x))) | |
| 1833 | + else | |
| 1834 | + am1=0.5*(y2 / (r + (x + 1)) + (s + (x - 1))) | |
| 1835 | + end if | |
| 1836 | + i_res = dlog(1.+(am1 + sqrt (am1 * (a + 1)))); | |
| 1837 | + else | |
| 1838 | + i_res = dlog (a + dsqrt (a*a - 1.)); | |
| 1839 | + end if | |
| 1840 | + if (rz <0.) then | |
| 1841 | + r_res=fvn_pi-r_res | |
| 1842 | + end if | |
| 1843 | + i_res=-sign(1._8,iz)*i_res | |
| 1844 | + fvn_z_acos=dcmplx(r_res)+fvn_i*dcmplx(i_res) | |
| 1845 | + | |
| 1846 | +end function fvn_z_acos | |
| 1847 | + | |
| 1848 | +function fvn_z_acos_real(r) | |
| 1849 | + ! return the double complex arc cosinus for a | |
| 1850 | + ! double precision argument | |
| 1851 | + implicit none | |
| 1852 | + real(kind=8) :: r | |
| 1853 | + complex(kind=8) :: fvn_z_acos_real | |
| 1854 | + | |
| 1855 | + if (dabs(r)<=1._8) then | |
| 1856 | + fvn_z_acos_real=dcmplx(dacos(r)) | |
| 1857 | + return | |
| 1858 | + end if | |
| 1859 | + if (r < 0._8) then | |
| 1860 | + fvn_z_acos_real=dcmplx(fvn_pi)-fvn_i*dcmplx(fvn_d_acosh(-r)) | |
| 1861 | + else | |
| 1862 | + fvn_z_acos_real=fvn_i*dcmplx(fvn_d_acosh(r)) | |
| 1863 | + end if | |
| 1864 | +end function | |
| 1865 | + | |
| 1866 | + | |
| 1867 | +function fvn_z_asin(z) | |
| 1868 | + ! double complex arcsin function derived from | |
| 1869 | + ! the c gsl library | |
| 1870 | + ! http://www.gnu.org/software/gsl/ | |
| 1871 | + implicit none | |
| 1872 | + complex(kind=8) :: fvn_z_asin | |
| 1873 | + complex(kind=8) :: z | |
| 1874 | + real(kind=8) :: rz,iz,x,y,a,b,y2,r,s,d,apx,am1 | |
| 1875 | + real(kind=8),parameter :: a_crossover=1.5_8,b_crossover = 0.6417_8 | |
| 1876 | + real(kind=8) :: r_res,i_res | |
| 1877 | + | |
| 1878 | + rz=dble(z) | |
| 1879 | + iz=aimag(z) | |
| 1880 | + if ( iz == 0._8 ) then | |
| 1881 | + ! z is real | |
| 1882 | + fvn_z_asin=fvn_z_asin_real(rz) | |
| 1883 | + return | |
| 1884 | + end if | |
| 1885 | + | |
| 1886 | + x=dabs(rz) | |
| 1887 | + y=dabs(iz) | |
| 1888 | + r=fvn_d_hypot(x+1.,y) | |
| 1889 | + s=fvn_d_hypot(x-1.,y) | |
| 1890 | + a=0.5*(r + s) | |
| 1891 | + b=x/a | |
| 1892 | + y2=y*y | |
| 1893 | + | |
| 1894 | + if (b <= b_crossover) then | |
| 1895 | + r_res=dasin(b) | |
| 1896 | + else | |
| 1897 | + if (x <= 1.) then | |
| 1898 | + d=0.5*(a+x)*(y2/(r+x+1)+(s + (1 - x))) | |
| 1899 | + r_res=datan(x/dsqrt(d)) | |
| 1900 | + else | |
| 1901 | + apx=a+x | |
| 1902 | + d=0.5*(apx/(r+x+1)+apx/(s + (x - 1))) | |
| 1903 | + r_res=datan(x/(y*dsqrt(d))); | |
| 1904 | + end if | |
| 1905 | + end if | |
| 1906 | + | |
| 1907 | + if (a <= a_crossover) then | |
| 1908 | + if (x < 1.) then | |
| 1909 | + am1=0.5*(y2 / (r + (x + 1)) + y2 / (s + (1 - x))) | |
| 1910 | + else | |
| 1911 | + am1=0.5*(y2 / (r + (x + 1)) + (s + (x - 1))) | |
| 1912 | + end if | |
| 1913 | + i_res = dlog(1.+(am1 + sqrt (am1 * (a + 1)))); | |
| 1914 | + else | |
| 1915 | + i_res = dlog (a + dsqrt (a*a - 1.)); | |
| 1916 | + end if | |
| 1917 | + r_res=sign(1._8,rz)*r_res | |
| 1918 | + i_res=sign(1._8,iz)*i_res | |
| 1919 | + fvn_z_asin=dcmplx(r_res)+fvn_i*dcmplx(i_res) | |
| 1920 | + | |
| 1921 | +end function fvn_z_asin | |
| 1922 | + | |
| 1923 | +function fvn_z_asin_real(r) | |
| 1924 | + ! return the double complex arc sinus for a | |
| 1925 | + ! double precision argument | |
| 1926 | + implicit none | |
| 1927 | + real(kind=8) :: r | |
| 1928 | + complex(kind=8) :: fvn_z_asin_real | |
| 1929 | + | |
| 1930 | + if (dabs(r)<=1._8) then | |
| 1931 | + fvn_z_asin_real=dcmplx(dasin(r)) | |
| 1932 | + return | |
| 1933 | + end if | |
| 1934 | + if (r < 0._8) then | |
| 1935 | + fvn_z_asin_real=dcmplx(-fvn_pi/2._8)+fvn_i*dcmplx(fvn_d_acosh(-r)) | |
| 1936 | + else | |
| 1937 | + fvn_z_asin_real=dcmplx(fvn_pi/2._8)-fvn_i*dcmplx(fvn_d_acosh(r)) | |
| 1938 | + end if | |
| 1939 | +end function fvn_z_asin_real | |
| 1940 | + | |
| 1941 | +function fvn_d_acosh(r) | |
| 1942 | + ! return the arc hyperbolic cosine | |
| 1943 | + implicit none | |
| 1944 | + real(kind=8) :: r | |
| 1945 | + real(kind=8) :: fvn_d_acosh | |
| 1946 | + if (r >=1) then | |
| 1947 | + fvn_d_acosh=dlog(r+dsqrt(r*r-1)) | |
| 1948 | + else | |
| 1949 | + !! TODO : Better error handling!!!!!! | |
| 1950 | + stop "Argument to fvn_d_acosh lesser than 1" | |
| 1951 | + end if | |
| 1952 | +end function fvn_d_acosh | |
| 1953 | + | |
| 1954 | +function fvn_d_asinh(r) | |
| 1955 | + ! return the arc hyperbolic sine | |
| 1956 | + implicit none | |
| 1957 | + real(kind=8) :: r | |
| 1958 | + real(kind=8) :: fvn_d_asinh | |
| 1959 | + fvn_d_asinh=dlog(r+dsqrt(r*r+1)) | |
| 1960 | +end function fvn_d_asinh | |
| 1961 | + | |
| 1962 | +function fvn_d_hypot(a,b) | |
| 1963 | + implicit none | |
| 1964 | + ! return the euclidian norm of vector(a,b) | |
| 1965 | + real(kind=8) :: a,b | |
| 1966 | + real(kind=8) :: fvn_d_hypot | |
| 1967 | + fvn_d_hypot=dsqrt(a*a+b*b) | |
| 1968 | +end function | |
| 1775 | 1969 | |
| 1776 | 1970 | |
| 1777 | 1971 |