Why does segmentation fault arise in fortran under calling subroutine nested gradually in two functions with double precision?Calling a subroutine in Fortran (Segmentation fault)Fortran: Calling a Subroutine in a FunctionSegmentation fault occurs at top of subroutine when C code calls Fortran subroutineSegmentation fault when passing a function as argument in a subroutineSegmentation fault, when passing functions with array arguments to subroutine in FortranSegmentation fault when calling C from FortranUsing CMPLX function for double or quad precision in FortranJulia calling Fortran. signal (11): Segmentation fault errorAre the specific double precision functions in Fortran useful?How to pass the kind for real precision to functions and subroutines in Fortran?
Does this Foo machine halt?
Team goes to lunch frequently, I do intermittent fasting but still want to socialize
Converting Piecewise function to C code
How can a surrogate pass on genes to a fertilized embryo?
How can I iterate this process?
Improving software when the author can see no need for improvement
Why does Intel's Haswell chip allow FP multiplication to be twice as fast as addition?
Shabbat clothing on shabbat chazon
How can I tell if a flight itinerary is fake?
Plausibility of Ice Eaters in the Arctic
What does Apple mean by "This may decrease battery life"?
Non-OR journals which regularly publish OR research
Max Order of an Isogeny Class of Rational Elliptic Curves is 8?
Why aren’t emergency services using callsigns?
Does a code snippet compile? Or does it get compiled?
Ordering a word list
Senior dev discreetly remoting in to computer and watching a coworker
Look mom! I made my own (Base 10) numeral system!
In reversi, can you overwrite two chips in one move?
Best gun to modify into a monsterhunter weapon?
Ex-contractor published company source code and secrets online
How do I explain to a team that the project they will work on for six months will certainly be cancelled?
During the Space Shuttle Columbia Disaster of 2003, Why Did The Flight Director Say, "Lock the doors."?
Author changing name
Why does segmentation fault arise in fortran under calling subroutine nested gradually in two functions with double precision?
Calling a subroutine in Fortran (Segmentation fault)Fortran: Calling a Subroutine in a FunctionSegmentation fault occurs at top of subroutine when C code calls Fortran subroutineSegmentation fault when passing a function as argument in a subroutineSegmentation fault, when passing functions with array arguments to subroutine in FortranSegmentation fault when calling C from FortranUsing CMPLX function for double or quad precision in FortranJulia calling Fortran. signal (11): Segmentation fault errorAre the specific double precision functions in Fortran useful?How to pass the kind for real precision to functions and subroutines in Fortran?
.everyoneloves__top-leaderboard:empty,.everyoneloves__mid-leaderboard:empty,.everyoneloves__bot-mid-leaderboard:empty margin-bottom:0;
I try to call subroutine SPLEV of FITPACK library through two functions ('wer' and 'qwe') nested one into another (the code is below).
The following message appears under execution of compiled program:
QWE
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
0 0x7F3EE4BF3E08
1 0x7F3EE4BF2F90
2 0x7F3EE453A4AF
3 0x4041B6 in splev_
4 0x400BD0 in value.3386 at pr.f90:?
5 0x400A6B in MAIN__ at pr.f90:?
Ошибка сегментирования (сделан дамп памяти)
If I compile my program with flags -g -fbacktrace -fsanitize=address,zero,undefined the follow output message appears:
QWE
0.37051690837706980
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
0 0x7FAB5F45CE08
1 0x7FAB5F45BF90
2 0x7FAB5EDA34AF
3 0x4075F0 in splev_ at splev.f:73 (discriminator 2)
4 0x400DDE in value.3386 at pr.f90:87
5 0x400FFA in qwe.3406 at pr.f90:43
6 0x400F88 in wer.3403 at pr.f90:48
7 0x400D08 in MAIN__ at pr.f90:38
Ошибка сегментирования (сделан дамп памяти)
If I compile my program with flags -g -fbacktrace -Wall -fcheck=all the follow output message appears:
QWE
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
0 0x7F2BE6F0FE08
1 0x7F2BE6F0EF90
2 0x7F2BE68564AF
3 0x4075F0 in splev_ at splev.f:73 (discriminator 2)
4 0x400DDE in value.3386 at pr.f90:87
5 0x400C46 in MAIN__ at pr.f90:35
Ошибка сегментирования (сделан дамп памяти)
If I compile my program with flags -g -fbacktrace -fsanitize=address the follow output message appears:
QWE
ASAN:SIGSEGV
=================================================================
==4796==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000000 (pc 0x000000408f67 bp 0x7ffe7a134440 sp 0x7ffe7a1341e0 T0)
0 0x408f66 in splev_ /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/splev.f:73
1 0x40145d in value.3386 (/home/yurchvlad/Science/Coll_Int/F90/f90DP/1/curfit+0x40145d)
2 0x4011a3 in intcoll /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/pr.f90:35
3 0x401849 in main /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/pr.f90:2
4 0x7fcad9b3282f in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x2082f)
5 0x400d38 in _start (/home/yurchvlad/Science/Coll_Int/F90/f90DP/1/curfit+0x400d38)
AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: SEGV /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/splev.f:73 splev_
==4796==ABORTING
Firstly I will show the code and then I will give some information about subroutines CURFIT and SPLEV of library FITPACK which are playing there a principal role.
Here is my code. This is just a test program, i.e. it is not confusion, that I interpolate there array of values of analytical function.
PROGRAM IntColl
USE Constants
IMPLICIT NONE
INTEGER :: i, nen ! i = counter
! nen, nmn, ne is sirvice variables, which
! appear on exit of CURFIT and needed on entry
! of SPLEV and SPLINT
REAL(DP) :: foo
REAL(DP) :: MOM1 ! dimensionless neutrino momentum
REAL(DP) :: dmg ( 1 : 2 * NG) ! dimensionless momentum grid
REAL(DP) :: endf( 1 : 2 * NG) ! electron neutrino distribution function
! muon neutrino distribution function
! electron and positron distribution function
REAL(DP) :: ten ( 1 : 2 * NG + k + 1) ! service arrays:
! ten is array arising on exit of working of CURFIT
! and contain knots of the spline (for endf, mndf and edf correspondingly).
REAL(DP) :: cen ( 1 : 2 * NG + k + 1) ! needed on entry of SPLEV and SPLINT
! cen appear on exit of CURFIT, contain coefficients of spline
! (for endf, mndf and edf correspondingly) and needed on entry of SPLEV and SPLINT.
REAL(DP) :: w ( 1 : 2 * NG + k + 1) ! w is array of weights for points on entry of CURFIT.
DO i = 1, 2 * NG
dmg(i) = i / 10.D+00 ! filling arrays to give their
endf(i) = eq_nu_di_fu(dmg(i)) ! on entry into subroutine
w(i) = 1.d+00 ! CURFIT
END DO
MOM1 = .53D+00
PRINT *, 'QWE'
CALL spline(dmg, endf, nen, ten, cen)
foo = value(MOM1, ten, nen, cen)
PRINT *, foo
PRINT *, wer(MOM1)
CONTAINS
REAL(DP) FUNCTION qwe(q) ! qwe and wer is "wrappers" for using
REAL(DP) :: q ! of subroutines spline > curfit
qwe = value(q, ten, nen, cen) ! in main program
END FUNCTION qwe
REAL(DP) FUNCTION wer(q)
REAL(DP) :: q
wer = qwe(q)
END FUNCTION wer
SUBROUTINE spline(x, y, n, t, c) ! spline is "hand-made wrapper" for
IMPLICIT NONE ! more convenient using of subroutine
! CURFIT in main program
INTEGER :: m, nest, n, lwrk, ier
INTEGER, PARAMETER :: iopt = 0
INTEGER :: iwrk( 1 : 10 * NG )
REAL(DP) :: xb, xe, fp
REAL(DP) :: wrk( 1 : 2 * NG * (k + 1) + (2 * NG + k + 1) * (7 + 3 * k) )
REAL(DP) :: x( 1 : 2 * NG), y(1: 2 * NG )
REAL(DP) :: t( 1 : 2 * NG + k + 1 )
REAL(DP) :: c( 1 : 2 * NG + k + 1 )
xb = 0.d+00
xe = x(2 * NG)
m = 2 * NG
nest = m + k + 1
lwrk = 2 * NG * (k + 1) + nest * (7 + 3 * k)
CALL curfit(iopt, m, x, y, w, xb, xe, k, s, nest, n, t, c, fp, wrk, lwrk, iwrk, ier)
END SUBROUTINE spline
REAL(DP) FUNCTION value(q, t, n, c) ! value is "hand-made wrapper" for
IMPLICIT NONE ! more convenient using of subroutine
! SPLEV in main program
INTEGER :: n, ier ! SPLEV should work only after
INTEGER, PARAMETER :: m = 1 ! CURFIT edned its working
REAL(DP) :: q
REAL(DP) :: t( 1 : 2 * NG + k + 1 )
REAL(DP) :: c( 1 : 2 * NG + k + 1 )
REAL(DP) :: ddmg(1), sddmg(1)
ddmg(1) = q
CALL splev(t, n, c, k, ddmg, sddmg, m, ier)
value = sddmg(1)
END FUNCTION value
REAL(DP) FUNCTION eq_nu_di_fu(y) ! eq_nu_di_fy givev values for array
IMPLICIT NONE ! to interpolate
REAL(DP) :: y
eq_nu_di_fu = 1 / (EXP(y) + 1)
END FUNCTION eq_nu_di_fu
END PROGRAM IntColl
The module Constants is there:
MODULE CONSTANTS
INTEGER, PARAMETER :: DP = SELECTED_REAL_KIND(15, 307)
INTEGER, PARAMETER :: NG = 200 ! NUMBER OF KNOTS OF GRID
INTEGER , PARAMETER :: K = 3 ! THE ORDER OF SPLINE
REAL(DP), PARAMETER :: S = 0.D+00 ! CUBIC SPLINE SMOOTHING FACTOR
END MODULE
Now, subroutines CURFIT and SPLEV appearing in above code with all their dependensies are in follow sources:
https://github.com/jbaayen/fitpackpp/tree/master/fitpack
where these subroutines are in double precision
and
http://www.netlib.org/dierckx/
where these subroutines are in single precision.
It is very important to mention that with single precision above scheme works!
Of course, if I use subroutines of single precision I modify all the types of all variables in corrisponding way.
What else have I observed:
straightforward using of FUNCTION value works.
If the line
PRINT *, 'QWE'
of the main program is commented, the value 'foo' also is not printed.
fortran gfortran
asked Mar 26 at 17:55
Vlad YurchenkoVlad Yurchenko
11 bronze badge
|
show 7 more comments
I try to call subroutine SPLEV of FITPACK library through two functions ('wer' and 'qwe') nested one into another (the code is below).
The following message appears under execution of compiled program:
QWE
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
0 0x7F3EE4BF3E08
1 0x7F3EE4BF2F90
2 0x7F3EE453A4AF
3 0x4041B6 in splev_
4 0x400BD0 in value.3386 at pr.f90:?
5 0x400A6B in MAIN__ at pr.f90:?
Ошибка сегментирования (сделан дамп памяти)
If I compile my program with flags -g -fbacktrace -fsanitize=address,zero,undefined the follow output message appears:
QWE
0.37051690837706980
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
0 0x7FAB5F45CE08
1 0x7FAB5F45BF90
2 0x7FAB5EDA34AF
3 0x4075F0 in splev_ at splev.f:73 (discriminator 2)
4 0x400DDE in value.3386 at pr.f90:87
5 0x400FFA in qwe.3406 at pr.f90:43
6 0x400F88 in wer.3403 at pr.f90:48
7 0x400D08 in MAIN__ at pr.f90:38
Ошибка сегментирования (сделан дамп памяти)
If I compile my program with flags -g -fbacktrace -Wall -fcheck=all the follow output message appears:
QWE
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
0 0x7F2BE6F0FE08
1 0x7F2BE6F0EF90
2 0x7F2BE68564AF
3 0x4075F0 in splev_ at splev.f:73 (discriminator 2)
4 0x400DDE in value.3386 at pr.f90:87
5 0x400C46 in MAIN__ at pr.f90:35
Ошибка сегментирования (сделан дамп памяти)
If I compile my program with flags -g -fbacktrace -fsanitize=address the follow output message appears:
QWE
ASAN:SIGSEGV
=================================================================
==4796==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000000 (pc 0x000000408f67 bp 0x7ffe7a134440 sp 0x7ffe7a1341e0 T0)
0 0x408f66 in splev_ /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/splev.f:73
1 0x40145d in value.3386 (/home/yurchvlad/Science/Coll_Int/F90/f90DP/1/curfit+0x40145d)
2 0x4011a3 in intcoll /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/pr.f90:35
3 0x401849 in main /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/pr.f90:2
4 0x7fcad9b3282f in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x2082f)
5 0x400d38 in _start (/home/yurchvlad/Science/Coll_Int/F90/f90DP/1/curfit+0x400d38)
AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: SEGV /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/splev.f:73 splev_
==4796==ABORTING
Firstly I will show the code and then I will give some information about subroutines CURFIT and SPLEV of library FITPACK which are playing there a principal role.
Here is my code. This is just a test program, i.e. it is not confusion, that I interpolate there array of values of analytical function.
PROGRAM IntColl
USE Constants
IMPLICIT NONE
INTEGER :: i, nen ! i = counter
! nen, nmn, ne is sirvice variables, which
! appear on exit of CURFIT and needed on entry
! of SPLEV and SPLINT
REAL(DP) :: foo
REAL(DP) :: MOM1 ! dimensionless neutrino momentum
REAL(DP) :: dmg ( 1 : 2 * NG) ! dimensionless momentum grid
REAL(DP) :: endf( 1 : 2 * NG) ! electron neutrino distribution function
! muon neutrino distribution function
! electron and positron distribution function
REAL(DP) :: ten ( 1 : 2 * NG + k + 1) ! service arrays:
! ten is array arising on exit of working of CURFIT
! and contain knots of the spline (for endf, mndf and edf correspondingly).
REAL(DP) :: cen ( 1 : 2 * NG + k + 1) ! needed on entry of SPLEV and SPLINT
! cen appear on exit of CURFIT, contain coefficients of spline
! (for endf, mndf and edf correspondingly) and needed on entry of SPLEV and SPLINT.
REAL(DP) :: w ( 1 : 2 * NG + k + 1) ! w is array of weights for points on entry of CURFIT.
DO i = 1, 2 * NG
dmg(i) = i / 10.D+00 ! filling arrays to give their
endf(i) = eq_nu_di_fu(dmg(i)) ! on entry into subroutine
w(i) = 1.d+00 ! CURFIT
END DO
MOM1 = .53D+00
PRINT *, 'QWE'
CALL spline(dmg, endf, nen, ten, cen)
foo = value(MOM1, ten, nen, cen)
PRINT *, foo
PRINT *, wer(MOM1)
CONTAINS
REAL(DP) FUNCTION qwe(q) ! qwe and wer is "wrappers" for using
REAL(DP) :: q ! of subroutines spline > curfit
qwe = value(q, ten, nen, cen) ! in main program
END FUNCTION qwe
REAL(DP) FUNCTION wer(q)
REAL(DP) :: q
wer = qwe(q)
END FUNCTION wer
SUBROUTINE spline(x, y, n, t, c) ! spline is "hand-made wrapper" for
IMPLICIT NONE ! more convenient using of subroutine
! CURFIT in main program
INTEGER :: m, nest, n, lwrk, ier
INTEGER, PARAMETER :: iopt = 0
INTEGER :: iwrk( 1 : 10 * NG )
REAL(DP) :: xb, xe, fp
REAL(DP) :: wrk( 1 : 2 * NG * (k + 1) + (2 * NG + k + 1) * (7 + 3 * k) )
REAL(DP) :: x( 1 : 2 * NG), y(1: 2 * NG )
REAL(DP) :: t( 1 : 2 * NG + k + 1 )
REAL(DP) :: c( 1 : 2 * NG + k + 1 )
xb = 0.d+00
xe = x(2 * NG)
m = 2 * NG
nest = m + k + 1
lwrk = 2 * NG * (k + 1) + nest * (7 + 3 * k)
CALL curfit(iopt, m, x, y, w, xb, xe, k, s, nest, n, t, c, fp, wrk, lwrk, iwrk, ier)
END SUBROUTINE spline
REAL(DP) FUNCTION value(q, t, n, c) ! value is "hand-made wrapper" for
IMPLICIT NONE ! more convenient using of subroutine
! SPLEV in main program
INTEGER :: n, ier ! SPLEV should work only after
INTEGER, PARAMETER :: m = 1 ! CURFIT edned its working
REAL(DP) :: q
REAL(DP) :: t( 1 : 2 * NG + k + 1 )
REAL(DP) :: c( 1 : 2 * NG + k + 1 )
REAL(DP) :: ddmg(1), sddmg(1)
ddmg(1) = q
CALL splev(t, n, c, k, ddmg, sddmg, m, ier)
value = sddmg(1)
END FUNCTION value
REAL(DP) FUNCTION eq_nu_di_fu(y) ! eq_nu_di_fy givev values for array
IMPLICIT NONE ! to interpolate
REAL(DP) :: y
eq_nu_di_fu = 1 / (EXP(y) + 1)
END FUNCTION eq_nu_di_fu
END PROGRAM IntColl
The module Constants is there:
MODULE CONSTANTS
INTEGER, PARAMETER :: DP = SELECTED_REAL_KIND(15, 307)
INTEGER, PARAMETER :: NG = 200 ! NUMBER OF KNOTS OF GRID
INTEGER , PARAMETER :: K = 3 ! THE ORDER OF SPLINE
REAL(DP), PARAMETER :: S = 0.D+00 ! CUBIC SPLINE SMOOTHING FACTOR
END MODULE
Now, subroutines CURFIT and SPLEV appearing in above code with all their dependensies are in follow sources:
https://github.com/jbaayen/fitpackpp/tree/master/fitpack
where these subroutines are in double precision
and
http://www.netlib.org/dierckx/
where these subroutines are in single precision.
It is very important to mention that with single precision above scheme works!
Of course, if I use subroutines of single precision I modify all the types of all variables in corrisponding way.
What else have I observed:
straightforward using of FUNCTION value works.
If the line
PRINT *, 'QWE'
of the main program is commented, the value 'foo' also is not printed.
fortran gfortran
asked Mar 26 at 17:55
Vlad YurchenkoVlad Yurchenko
11 bronze badge
Use compiler checks to identify bugs. You can find many yourself. Try:gfortran -g -fbacktrace -Wall -fcheck=alland alsogfortran -g -fbacktrace -fsanitize=address,zero,undefined. Compile your code with these options, re-run it and report the results.
– Vladimir F
Mar 26 at 20:23
@Vladimir F-g -fbacktrace -Wall -fcheck=allreturnsfpcurf.f:186:35: if(fpold-fp.gt.acc) npl1 = rn*fpms/(fpold-fp) 1 Warning: Possible change of value in conversion from REAL(8) to INTEGER(4) at (1) [-Wconversion]and-g -fbacktrace -fsanitize=address,zero,undefinedreturnsgfortran: error: unrecognized argument to -fsanitize= option: ‘zero’
– Vlad Yurchenko
Mar 27 at 7:19
What happens if you run the program compiled with those flags?
– Ian Bush
Mar 27 at 7:21
@Ian Bush if I run the program compiled with those flags the same result is returned
– Vlad Yurchenko
Mar 27 at 7:25
I seriously doubt the output was exactly the same, there should at least be line numbers in the backtrace. Try removing the "zero" from the sanitize flag. If needed, remove also the "undefined" and re-rurn.
– Vladimir F
Mar 27 at 7:28
|
show 7 more comments
I try to call subroutine SPLEV of FITPACK library through two functions ('wer' and 'qwe') nested one into another (the code is below).
The following message appears under execution of compiled program:
QWE
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
0 0x7F3EE4BF3E08
1 0x7F3EE4BF2F90
2 0x7F3EE453A4AF
3 0x4041B6 in splev_
4 0x400BD0 in value.3386 at pr.f90:?
5 0x400A6B in MAIN__ at pr.f90:?
Ошибка сегментирования (сделан дамп памяти)
If I compile my program with flags -g -fbacktrace -fsanitize=address,zero,undefined the follow output message appears:
QWE
0.37051690837706980
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
0 0x7FAB5F45CE08
1 0x7FAB5F45BF90
2 0x7FAB5EDA34AF
3 0x4075F0 in splev_ at splev.f:73 (discriminator 2)
4 0x400DDE in value.3386 at pr.f90:87
5 0x400FFA in qwe.3406 at pr.f90:43
6 0x400F88 in wer.3403 at pr.f90:48
7 0x400D08 in MAIN__ at pr.f90:38
Ошибка сегментирования (сделан дамп памяти)
If I compile my program with flags -g -fbacktrace -Wall -fcheck=all the follow output message appears:
QWE
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
0 0x7F2BE6F0FE08
1 0x7F2BE6F0EF90
2 0x7F2BE68564AF
3 0x4075F0 in splev_ at splev.f:73 (discriminator 2)
4 0x400DDE in value.3386 at pr.f90:87
5 0x400C46 in MAIN__ at pr.f90:35
Ошибка сегментирования (сделан дамп памяти)
If I compile my program with flags -g -fbacktrace -fsanitize=address the follow output message appears:
QWE
ASAN:SIGSEGV
=================================================================
==4796==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000000 (pc 0x000000408f67 bp 0x7ffe7a134440 sp 0x7ffe7a1341e0 T0)
0 0x408f66 in splev_ /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/splev.f:73
1 0x40145d in value.3386 (/home/yurchvlad/Science/Coll_Int/F90/f90DP/1/curfit+0x40145d)
2 0x4011a3 in intcoll /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/pr.f90:35
3 0x401849 in main /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/pr.f90:2
4 0x7fcad9b3282f in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x2082f)
5 0x400d38 in _start (/home/yurchvlad/Science/Coll_Int/F90/f90DP/1/curfit+0x400d38)
AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: SEGV /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/splev.f:73 splev_
==4796==ABORTING
Firstly I will show the code and then I will give some information about subroutines CURFIT and SPLEV of library FITPACK which are playing there a principal role.
Here is my code. This is just a test program, i.e. it is not confusion, that I interpolate there array of values of analytical function.
PROGRAM IntColl
USE Constants
IMPLICIT NONE
INTEGER :: i, nen ! i = counter
! nen, nmn, ne is sirvice variables, which
! appear on exit of CURFIT and needed on entry
! of SPLEV and SPLINT
REAL(DP) :: foo
REAL(DP) :: MOM1 ! dimensionless neutrino momentum
REAL(DP) :: dmg ( 1 : 2 * NG) ! dimensionless momentum grid
REAL(DP) :: endf( 1 : 2 * NG) ! electron neutrino distribution function
! muon neutrino distribution function
! electron and positron distribution function
REAL(DP) :: ten ( 1 : 2 * NG + k + 1) ! service arrays:
! ten is array arising on exit of working of CURFIT
! and contain knots of the spline (for endf, mndf and edf correspondingly).
REAL(DP) :: cen ( 1 : 2 * NG + k + 1) ! needed on entry of SPLEV and SPLINT
! cen appear on exit of CURFIT, contain coefficients of spline
! (for endf, mndf and edf correspondingly) and needed on entry of SPLEV and SPLINT.
REAL(DP) :: w ( 1 : 2 * NG + k + 1) ! w is array of weights for points on entry of CURFIT.
DO i = 1, 2 * NG
dmg(i) = i / 10.D+00 ! filling arrays to give their
endf(i) = eq_nu_di_fu(dmg(i)) ! on entry into subroutine
w(i) = 1.d+00 ! CURFIT
END DO
MOM1 = .53D+00
PRINT *, 'QWE'
CALL spline(dmg, endf, nen, ten, cen)
foo = value(MOM1, ten, nen, cen)
PRINT *, foo
PRINT *, wer(MOM1)
CONTAINS
REAL(DP) FUNCTION qwe(q) ! qwe and wer is "wrappers" for using
REAL(DP) :: q ! of subroutines spline > curfit
qwe = value(q, ten, nen, cen) ! in main program
END FUNCTION qwe
REAL(DP) FUNCTION wer(q)
REAL(DP) :: q
wer = qwe(q)
END FUNCTION wer
SUBROUTINE spline(x, y, n, t, c) ! spline is "hand-made wrapper" for
IMPLICIT NONE ! more convenient using of subroutine
! CURFIT in main program
INTEGER :: m, nest, n, lwrk, ier
INTEGER, PARAMETER :: iopt = 0
INTEGER :: iwrk( 1 : 10 * NG )
REAL(DP) :: xb, xe, fp
REAL(DP) :: wrk( 1 : 2 * NG * (k + 1) + (2 * NG + k + 1) * (7 + 3 * k) )
REAL(DP) :: x( 1 : 2 * NG), y(1: 2 * NG )
REAL(DP) :: t( 1 : 2 * NG + k + 1 )
REAL(DP) :: c( 1 : 2 * NG + k + 1 )
xb = 0.d+00
xe = x(2 * NG)
m = 2 * NG
nest = m + k + 1
lwrk = 2 * NG * (k + 1) + nest * (7 + 3 * k)
CALL curfit(iopt, m, x, y, w, xb, xe, k, s, nest, n, t, c, fp, wrk, lwrk, iwrk, ier)
END SUBROUTINE spline
REAL(DP) FUNCTION value(q, t, n, c) ! value is "hand-made wrapper" for
IMPLICIT NONE ! more convenient using of subroutine
! SPLEV in main program
INTEGER :: n, ier ! SPLEV should work only after
INTEGER, PARAMETER :: m = 1 ! CURFIT edned its working
REAL(DP) :: q
REAL(DP) :: t( 1 : 2 * NG + k + 1 )
REAL(DP) :: c( 1 : 2 * NG + k + 1 )
REAL(DP) :: ddmg(1), sddmg(1)
ddmg(1) = q
CALL splev(t, n, c, k, ddmg, sddmg, m, ier)
value = sddmg(1)
END FUNCTION value
REAL(DP) FUNCTION eq_nu_di_fu(y) ! eq_nu_di_fy givev values for array
IMPLICIT NONE ! to interpolate
REAL(DP) :: y
eq_nu_di_fu = 1 / (EXP(y) + 1)
END FUNCTION eq_nu_di_fu
END PROGRAM IntColl
The module Constants is there:
MODULE CONSTANTS
INTEGER, PARAMETER :: DP = SELECTED_REAL_KIND(15, 307)
INTEGER, PARAMETER :: NG = 200 ! NUMBER OF KNOTS OF GRID
INTEGER , PARAMETER :: K = 3 ! THE ORDER OF SPLINE
REAL(DP), PARAMETER :: S = 0.D+00 ! CUBIC SPLINE SMOOTHING FACTOR
END MODULE
Now, subroutines CURFIT and SPLEV appearing in above code with all their dependensies are in follow sources:
https://github.com/jbaayen/fitpackpp/tree/master/fitpack
where these subroutines are in double precision
and
http://www.netlib.org/dierckx/
where these subroutines are in single precision.
It is very important to mention that with single precision above scheme works!
Of course, if I use subroutines of single precision I modify all the types of all variables in corrisponding way.
What else have I observed:
straightforward using of FUNCTION value works.
If the line
PRINT *, 'QWE'
of the main program is commented, the value 'foo' also is not printed.
fortran gfortran
asked Mar 26 at 17:55
Vlad YurchenkoVlad Yurchenko
11 bronze badge
I try to call subroutine SPLEV of FITPACK library through two functions ('wer' and 'qwe') nested one into another (the code is below).
The following message appears under execution of compiled program:
QWE
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
0 0x7F3EE4BF3E08
1 0x7F3EE4BF2F90
2 0x7F3EE453A4AF
3 0x4041B6 in splev_
4 0x400BD0 in value.3386 at pr.f90:?
5 0x400A6B in MAIN__ at pr.f90:?
Ошибка сегментирования (сделан дамп памяти)
If I compile my program with flags -g -fbacktrace -fsanitize=address,zero,undefined the follow output message appears:
QWE
0.37051690837706980
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
0 0x7FAB5F45CE08
1 0x7FAB5F45BF90
2 0x7FAB5EDA34AF
3 0x4075F0 in splev_ at splev.f:73 (discriminator 2)
4 0x400DDE in value.3386 at pr.f90:87
5 0x400FFA in qwe.3406 at pr.f90:43
6 0x400F88 in wer.3403 at pr.f90:48
7 0x400D08 in MAIN__ at pr.f90:38
Ошибка сегментирования (сделан дамп памяти)
If I compile my program with flags -g -fbacktrace -Wall -fcheck=all the follow output message appears:
QWE
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
0 0x7F2BE6F0FE08
1 0x7F2BE6F0EF90
2 0x7F2BE68564AF
3 0x4075F0 in splev_ at splev.f:73 (discriminator 2)
4 0x400DDE in value.3386 at pr.f90:87
5 0x400C46 in MAIN__ at pr.f90:35
Ошибка сегментирования (сделан дамп памяти)
If I compile my program with flags -g -fbacktrace -fsanitize=address the follow output message appears:
QWE
ASAN:SIGSEGV
=================================================================
==4796==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000000 (pc 0x000000408f67 bp 0x7ffe7a134440 sp 0x7ffe7a1341e0 T0)
0 0x408f66 in splev_ /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/splev.f:73
1 0x40145d in value.3386 (/home/yurchvlad/Science/Coll_Int/F90/f90DP/1/curfit+0x40145d)
2 0x4011a3 in intcoll /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/pr.f90:35
3 0x401849 in main /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/pr.f90:2
4 0x7fcad9b3282f in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x2082f)
5 0x400d38 in _start (/home/yurchvlad/Science/Coll_Int/F90/f90DP/1/curfit+0x400d38)
AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: SEGV /home/yurchvlad/Science/Coll_Int/F90/f90DP/1/splev.f:73 splev_
==4796==ABORTING
Firstly I will show the code and then I will give some information about subroutines CURFIT and SPLEV of library FITPACK which are playing there a principal role.
Here is my code. This is just a test program, i.e. it is not confusion, that I interpolate there array of values of analytical function.
PROGRAM IntColl
USE Constants
IMPLICIT NONE
INTEGER :: i, nen ! i = counter
! nen, nmn, ne is sirvice variables, which
! appear on exit of CURFIT and needed on entry
! of SPLEV and SPLINT
REAL(DP) :: foo
REAL(DP) :: MOM1 ! dimensionless neutrino momentum
REAL(DP) :: dmg ( 1 : 2 * NG) ! dimensionless momentum grid
REAL(DP) :: endf( 1 : 2 * NG) ! electron neutrino distribution function
! muon neutrino distribution function
! electron and positron distribution function
REAL(DP) :: ten ( 1 : 2 * NG + k + 1) ! service arrays:
! ten is array arising on exit of working of CURFIT
! and contain knots of the spline (for endf, mndf and edf correspondingly).
REAL(DP) :: cen ( 1 : 2 * NG + k + 1) ! needed on entry of SPLEV and SPLINT
! cen appear on exit of CURFIT, contain coefficients of spline
! (for endf, mndf and edf correspondingly) and needed on entry of SPLEV and SPLINT.
REAL(DP) :: w ( 1 : 2 * NG + k + 1) ! w is array of weights for points on entry of CURFIT.
DO i = 1, 2 * NG
dmg(i) = i / 10.D+00 ! filling arrays to give their
endf(i) = eq_nu_di_fu(dmg(i)) ! on entry into subroutine
w(i) = 1.d+00 ! CURFIT
END DO
MOM1 = .53D+00
PRINT *, 'QWE'
CALL spline(dmg, endf, nen, ten, cen)
foo = value(MOM1, ten, nen, cen)
PRINT *, foo
PRINT *, wer(MOM1)
CONTAINS
REAL(DP) FUNCTION qwe(q) ! qwe and wer is "wrappers" for using
REAL(DP) :: q ! of subroutines spline > curfit
qwe = value(q, ten, nen, cen) ! in main program
END FUNCTION qwe
REAL(DP) FUNCTION wer(q)
REAL(DP) :: q
wer = qwe(q)
END FUNCTION wer
SUBROUTINE spline(x, y, n, t, c) ! spline is "hand-made wrapper" for
IMPLICIT NONE ! more convenient using of subroutine
! CURFIT in main program
INTEGER :: m, nest, n, lwrk, ier
INTEGER, PARAMETER :: iopt = 0
INTEGER :: iwrk( 1 : 10 * NG )
REAL(DP) :: xb, xe, fp
REAL(DP) :: wrk( 1 : 2 * NG * (k + 1) + (2 * NG + k + 1) * (7 + 3 * k) )
REAL(DP) :: x( 1 : 2 * NG), y(1: 2 * NG )
REAL(DP) :: t( 1 : 2 * NG + k + 1 )
REAL(DP) :: c( 1 : 2 * NG + k + 1 )
xb = 0.d+00
xe = x(2 * NG)
m = 2 * NG
nest = m + k + 1
lwrk = 2 * NG * (k + 1) + nest * (7 + 3 * k)
CALL curfit(iopt, m, x, y, w, xb, xe, k, s, nest, n, t, c, fp, wrk, lwrk, iwrk, ier)
END SUBROUTINE spline
REAL(DP) FUNCTION value(q, t, n, c) ! value is "hand-made wrapper" for
IMPLICIT NONE ! more convenient using of subroutine
! SPLEV in main program
INTEGER :: n, ier ! SPLEV should work only after
INTEGER, PARAMETER :: m = 1 ! CURFIT edned its working
REAL(DP) :: q
REAL(DP) :: t( 1 : 2 * NG + k + 1 )
REAL(DP) :: c( 1 : 2 * NG + k + 1 )
REAL(DP) :: ddmg(1), sddmg(1)
ddmg(1) = q
CALL splev(t, n, c, k, ddmg, sddmg, m, ier)
value = sddmg(1)
END FUNCTION value
REAL(DP) FUNCTION eq_nu_di_fu(y) ! eq_nu_di_fy givev values for array
IMPLICIT NONE ! to interpolate
REAL(DP) :: y
eq_nu_di_fu = 1 / (EXP(y) + 1)
END FUNCTION eq_nu_di_fu
END PROGRAM IntColl
The module Constants is there:
MODULE CONSTANTS
INTEGER, PARAMETER :: DP = SELECTED_REAL_KIND(15, 307)
INTEGER, PARAMETER :: NG = 200 ! NUMBER OF KNOTS OF GRID
INTEGER , PARAMETER :: K = 3 ! THE ORDER OF SPLINE
REAL(DP), PARAMETER :: S = 0.D+00 ! CUBIC SPLINE SMOOTHING FACTOR
END MODULE
Now, subroutines CURFIT and SPLEV appearing in above code with all their dependensies are in follow sources:
https://github.com/jbaayen/fitpackpp/tree/master/fitpack
where these subroutines are in double precision
and
http://www.netlib.org/dierckx/
where these subroutines are in single precision.
It is very important to mention that with single precision above scheme works!
Of course, if I use subroutines of single precision I modify all the types of all variables in corrisponding way.
What else have I observed:
straightforward using of FUNCTION value works.
If the line
PRINT *, 'QWE'
of the main program is commented, the value 'foo' also is not printed.
fortran gfortran
fortran gfortran
asked Mar 26 at 17:55
Vlad YurchenkoVlad Yurchenko
11 bronze badge
asked Mar 26 at 17:55
Vlad YurchenkoVlad Yurchenko
11 bronze badge
edited Mar 27 at 12:05
Vlad Yurchenko
asked Mar 26 at 17:55
Vlad YurchenkoVlad Yurchenko
11 bronze badge
asked Mar 26 at 17:55
Vlad YurchenkoVlad Yurchenko
11 bronze badge
asked Mar 26 at 17:55
Vlad YurchenkoVlad Yurchenko
11 bronze badge
11 bronze badge
Use compiler checks to identify bugs. You can find many yourself. Try:gfortran -g -fbacktrace -Wall -fcheck=alland alsogfortran -g -fbacktrace -fsanitize=address,zero,undefined. Compile your code with these options, re-run it and report the results.
– Vladimir F
Mar 26 at 20:23
@Vladimir F-g -fbacktrace -Wall -fcheck=allreturnsfpcurf.f:186:35: if(fpold-fp.gt.acc) npl1 = rn*fpms/(fpold-fp) 1 Warning: Possible change of value in conversion from REAL(8) to INTEGER(4) at (1) [-Wconversion]and-g -fbacktrace -fsanitize=address,zero,undefinedreturnsgfortran: error: unrecognized argument to -fsanitize= option: ‘zero’
– Vlad Yurchenko
Mar 27 at 7:19
What happens if you run the program compiled with those flags?
– Ian Bush
Mar 27 at 7:21
@Ian Bush if I run the program compiled with those flags the same result is returned
– Vlad Yurchenko
Mar 27 at 7:25
I seriously doubt the output was exactly the same, there should at least be line numbers in the backtrace. Try removing the "zero" from the sanitize flag. If needed, remove also the "undefined" and re-rurn.
– Vladimir F
Mar 27 at 7:28
|
show 7 more comments
Use compiler checks to identify bugs. You can find many yourself. Try:gfortran -g -fbacktrace -Wall -fcheck=alland alsogfortran -g -fbacktrace -fsanitize=address,zero,undefined. Compile your code with these options, re-run it and report the results.
– Vladimir F
Mar 26 at 20:23
@Vladimir F-g -fbacktrace -Wall -fcheck=allreturnsfpcurf.f:186:35: if(fpold-fp.gt.acc) npl1 = rn*fpms/(fpold-fp) 1 Warning: Possible change of value in conversion from REAL(8) to INTEGER(4) at (1) [-Wconversion]and-g -fbacktrace -fsanitize=address,zero,undefinedreturnsgfortran: error: unrecognized argument to -fsanitize= option: ‘zero’
– Vlad Yurchenko
Mar 27 at 7:19
What happens if you run the program compiled with those flags?
– Ian Bush
Mar 27 at 7:21
@Ian Bush if I run the program compiled with those flags the same result is returned
– Vlad Yurchenko
Mar 27 at 7:25
I seriously doubt the output was exactly the same, there should at least be line numbers in the backtrace. Try removing the "zero" from the sanitize flag. If needed, remove also the "undefined" and re-rurn.
– Vladimir F
Mar 27 at 7:28
Use compiler checks to identify bugs. You can find many yourself. Try:
gfortran -g -fbacktrace -Wall -fcheck=all and also gfortran -g -fbacktrace -fsanitize=address,zero,undefined. Compile your code with these options, re-run it and report the results.– Vladimir F
Mar 26 at 20:23
Use compiler checks to identify bugs. You can find many yourself. Try:
gfortran -g -fbacktrace -Wall -fcheck=all and also gfortran -g -fbacktrace -fsanitize=address,zero,undefined. Compile your code with these options, re-run it and report the results.– Vladimir F
Mar 26 at 20:23
@Vladimir F
-g -fbacktrace -Wall -fcheck=all returns fpcurf.f:186:35: if(fpold-fp.gt.acc) npl1 = rn*fpms/(fpold-fp) 1 Warning: Possible change of value in conversion from REAL(8) to INTEGER(4) at (1) [-Wconversion] and -g -fbacktrace -fsanitize=address,zero,undefined returns gfortran: error: unrecognized argument to -fsanitize= option: ‘zero’– Vlad Yurchenko
Mar 27 at 7:19
@Vladimir F
-g -fbacktrace -Wall -fcheck=all returns fpcurf.f:186:35: if(fpold-fp.gt.acc) npl1 = rn*fpms/(fpold-fp) 1 Warning: Possible change of value in conversion from REAL(8) to INTEGER(4) at (1) [-Wconversion] and -g -fbacktrace -fsanitize=address,zero,undefined returns gfortran: error: unrecognized argument to -fsanitize= option: ‘zero’– Vlad Yurchenko
Mar 27 at 7:19
What happens if you run the program compiled with those flags?
– Ian Bush
Mar 27 at 7:21
What happens if you run the program compiled with those flags?
– Ian Bush
Mar 27 at 7:21
@Ian Bush if I run the program compiled with those flags the same result is returned
– Vlad Yurchenko
Mar 27 at 7:25
@Ian Bush if I run the program compiled with those flags the same result is returned
– Vlad Yurchenko
Mar 27 at 7:25
I seriously doubt the output was exactly the same, there should at least be line numbers in the backtrace. Try removing the "zero" from the sanitize flag. If needed, remove also the "undefined" and re-rurn.
– Vladimir F
Mar 27 at 7:28
I seriously doubt the output was exactly the same, there should at least be line numbers in the backtrace. Try removing the "zero" from the sanitize flag. If needed, remove also the "undefined" and re-rurn.
– Vladimir F
Mar 27 at 7:28
|
show 7 more comments
0
active
oldest
votes
Your Answer
StackExchange.ifUsing("editor", function ()
StackExchange.using("externalEditor", function ()
StackExchange.using("snippets", function ()
StackExchange.snippets.init();
);
);
, "code-snippets");
StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "1"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);
else
createEditor();
);
function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader:
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
,
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);
);
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f55363485%2fwhy-does-segmentation-fault-arise-in-fortran-under-calling-subroutine-nested-gra%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
0
active
oldest
votes
0
active
oldest
votes
active
oldest
votes
active
oldest
votes
Is this question similar to what you get asked at work? Learn more about asking and sharing private information with your coworkers using Stack Overflow for Teams.
Is this question similar to what you get asked at work? Learn more about asking and sharing private information with your coworkers using Stack Overflow for Teams.
Thanks for contributing an answer to Stack Overflow!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f55363485%2fwhy-does-segmentation-fault-arise-in-fortran-under-calling-subroutine-nested-gra%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Use compiler checks to identify bugs. You can find many yourself. Try:
gfortran -g -fbacktrace -Wall -fcheck=alland alsogfortran -g -fbacktrace -fsanitize=address,zero,undefined. Compile your code with these options, re-run it and report the results.– Vladimir F
Mar 26 at 20:23
@Vladimir F
-g -fbacktrace -Wall -fcheck=allreturnsfpcurf.f:186:35: if(fpold-fp.gt.acc) npl1 = rn*fpms/(fpold-fp) 1 Warning: Possible change of value in conversion from REAL(8) to INTEGER(4) at (1) [-Wconversion]and-g -fbacktrace -fsanitize=address,zero,undefinedreturnsgfortran: error: unrecognized argument to -fsanitize= option: ‘zero’– Vlad Yurchenko
Mar 27 at 7:19
What happens if you run the program compiled with those flags?
– Ian Bush
Mar 27 at 7:21
@Ian Bush if I run the program compiled with those flags the same result is returned
– Vlad Yurchenko
Mar 27 at 7:25
I seriously doubt the output was exactly the same, there should at least be line numbers in the backtrace. Try removing the "zero" from the sanitize flag. If needed, remove also the "undefined" and re-rurn.
– Vladimir F
Mar 27 at 7:28