I have been through my code and re-written the base routine as free-form Fortran; and it works exactly as the older version…including an error!? Note: The following floating-point exceptions are signalling: IEEE_OVERFLOW_FLAG IEEE_DENORMAL I believe I have checked each variable and set them as dou…

If it does not show the line where the error occurs, you can try to compile your program in debug mode (-g option). You may obtain more info when executing.

The error is happening in runtime, so not sure debugging in compiling would help.(?) There are 30000 steps and the program ends by producing a lot of data - all correct and complete. At present, no other file is used, so I could output data at given points, but I’m unsure what I am looking to find…

You are right, -g is not sufficient but you can use ieee_set_halting_mode: program main use, intrinsic :: iso_fortran_env, only: dp=>real64 use, intrinsic :: ieee_exceptions implicit none real(dp) :: x call ieee_set_halting_mode(ieee_overflow, .true.) x = 1d308 x = x*x print *, x…

Yes, sorry. I refer to it as an error simply because it is not a clean finish; something is not quite right somewhere and I always like to know where. So, not an error but the note tells me that something is in there that is incorrect. It might not be an error now but it might prove to be later on!…

[image] garynewport: EDIT: Looks like this command doesn’t work under x86… I am working on an Intel Core i3, therefore a x86_64 architecture. Which compiler do you use?

And you were totally correct! :slight_smile: However, I’m not getting output beyond the original note. I believe I have set it all correctly and run it accordingly, but all I got was this… gfortran -g dummymain.f90 && ./a.out > dummymain.lst Note: The following floating-point exceptions are sign…

Not answering your question, but I must say that .eqv. is rarely needed in Fortran and that if (massonly .eqv. .false.) then write(6,201) i, time, dtime, CorM, CorR, CorL, totL, fluxM, smass else write(6,202) i, time, dtime, CorM, fluxM, smass…

I was looking for something to replace the .eqv. with! And that is such an obvious answer, but not one I spotted; thank you. There’s no shorthand for .and., etc is there? Couldn’t find it when I searched. EDIT: Is there a NOT equivalence? I have one select statement where the only option is if m…

Write if (.not. massonly) then call stelcor(CorM, CorR, CorL, fluxM, time, iprint) end if or preferably IMO using the one-line if if (.not. massonly) call stelcor(CorM, CorR, CorL, fluxM, time, iprint)

I did wonder if that was it. Sorry, I don’t like the one-line if (unless the action is particularly simple and small). You did state IMO, so we’re happy to disagree on this minor difference. :slight_smile:

How do I chase down this error?

Help

garynewport October 19, 2021, 10:23am 20

That is excellent; thank you!

Topic		Replies	Views
Note: The following floating-point exceptions are signalling: IEEE_INVALID_FLAG IEEE_DIVIDE_BY_ZERO	16	7878	April 4, 2023
Undetected format error	34	627	June 10, 2025
Resolving floating point exceptions signalled after program termination Help	9	2063	April 25, 2023
Exceptions proposal Language enhancement	52	3030	March 7, 2023
ErrorFx: Fortran exception library Announcements	30	1534	October 19, 2021

How do I chase down this error?

Related topics