如何提高精度？

zyl · 发表于 2017-10-17 20:55:35

程序定义的变量是单精度，但是想提高计算精度。我把变量都定义成为REAL(KIND=8)之后，程序报实参与虚参不匹配，所以困惑了。给位大大，帮帮忙改下呗，谢谢谢谢！

[Fortran] 纯文本查看 复制代码

001PROGRAM MAIN 
002    
003 IMPLICIT NONE
004 INTEGER I 
005REAL T  
006REAL  CCU,CNBSN 
007REAL RES        
008REAL*8 ACU,ASC  
009REAL B,RRR  
010REAL CAV 
011REAL  t1,t2
012REAL*16  Q,F 
013REAL y   
014REAL Iop  
015 F=0.000
016  
017 t1=0.0
018 Iop=8930.0
019 B=12.84
020 RRR=150
021 ACU=3.80233E-5
022 ASC=3.80233E-5
023 t2=2.0
024 y=5.0
025 Q=Iop**2*y*(EXP(-2.0*y*t1/2.0)-EXP(-2.0*y*t2/2.0)) &
026              /(2.0*Acu**2.0)
027 WRITE(*,*) "Q=",Q
028 DO I=1,11 
029     T=89.0+(i-1)*0.1
030     CALL CNB3SN(T,B,CNBSN)
031     CALL CCCU(T,CCU)
032     CALL DIANZULV(T,B,RRR,RES)
033     CAV=(CCU*ACU+CNBSN*ASC)/(ACU+ASC)
034      F=CAV*0.001/RES+F
035END DO
036STOP
037END
038     
039 
040SUBROUTINE CNB3SN(T,B,CNBSN)
041IMPLICIT NONE
042REAL   b4,b3,b2,b1,b0
043REAL  T,B
044REAL  CNBSN
045   IF(T>0.0  .AND.  T<10.4 ) THEN
046           b4=0.0
047           b3=38.8-1.8*B+0.0634*B**2
048           b2=-110*exp(-0.434*B)
049           b1=207.0-3.83*B+2.86*B**2
050           b0=0
051          
052   ELSEIF(T>10.4  .AND.  T<26.113 ) THEN   
053           b4=0.0
054           b3=7.42
055           b2=0.0
056           b1=1522.0
057           b0=0.0
058  ELSEIF(T>=26.113 .AND. T<169.416) THEN
059           b4=0.0
060           b3=0.0
061           b2=-61.635
062           b1=19902.0
063           b0=-305807.0
064  ELSEIF(T>=169.416 .AND. T<300.0) THEN 
065           b4=0.0
066           b3=0.0
067           b2=-7.4636
068           b1=4411.0
069           b0=763801.0
070  ELSE
071             b4=0.0
072             b3=0.0
073             b2=0.0
074             b1=0.0
075             b0=1415377.0
076  ENDIF
077         CNBSN=b4*(T**4)+b3*(T**3)+b2*(T**2)+b1*T+b0
078  RETURN
079  END  
080SUBROUTINE CCCU(T,CCU)
081IMPLICIT NONE
082REAL a0,a1,a2,a3,a4,a5,a6,a7
083REAL dens,aa1,aa2,aa3,aa4,aa
084REAL T
085REAL CCU
086a0=-1.91844
087a1=-0.15973
088a2=8.61013
089a3=-18.996
090a4=21.9661
091a5=-12.7328
092a6=3.54322
093a7=-0.3797
094aa1=a0*((log10(T))**0.0)+a1*((log10(T))**1.0)
095aa2=a2*((log10(T))**2.0)+a3*((log10(T))**3.0)
096aa3=a4*((log10(T))**4.0)+a5*((log10(T))**5.0)  
097aa4=a6*((log10(T))**6.0)+a7*((log10(T))**7.0)
098aa=aa1+aa2+aa3+aa4
099dens=8960.0 
100  CCU=dens*(10.0**aa)*1.0
101 WRITE(*,*) "CCU=",CCU
102 RETURN
103 END
104     
105    
106SUBROUTINE DIANZULV(T,B,RRR,RES)   
107REAL RES1,RES2,RES3,RES
108    RES1=1.7E-8/RRR
109    RES2=1.0E-8/(2.33*10**9/T**5+9.57*10**5/T**3+163.0/T)
110    RES3=(0.37E-10+0.0005E-10*RRR)*B
111    RES=RES1+RES2+RES3
112    RETURN
113    END

pasuka · 发表于 2017-10-17 21:35:18

不想修改程序的话，可以通过添加编译器的编译指令实现
譬如gfortran可以参考 https://gcc.gnu.org/onlinedocs/g ... ran-Dialect-Options
-fdefault-real-8
Set the default real type to an 8 byte wide type. This option also affects the kind of non-double real constants like 1.0, and does promote the default width of DOUBLE PRECISION to 16 bytes if possible, unless -fdefault-double-8 is given, too. Unlike -freal-4-real-8, it does not promote variables with explicit kind declaration.
或者
-freal-4-real-8
-freal-4-real-10
-freal-4-real-16
-freal-8-real-4
-freal-8-real-10
-freal-8-real-16
Promote all REAL(KIND=M) entities to REAL(KIND=N) entities. If REAL(KIND=N) is unavailable, then an error will be issued. All other real kind types are unaffected by this option. These options should be used with care and may not be suitable for your codes. Areas of possible concern include calls to external procedures, alignment in EQUIVALENCE and/or COMMON, generic interfaces, BOZ literal constant conversion, and I/O. Inspection of the intermediate representation of the translated Fortran code, produced by -fdump-tree-original, is suggested.

zyl · 发表于 2017-10-18 10:56:06

已解决，谢谢！

[数值问题] 如何提高精度？

水王勋章

元老勋章

热心勋章