[Users] Compilation error with ifort

Roland Haas rhaas at illinois.edu
Tue Sep 7 18:40:19 CDT 2021


Hello Lorenzo,

the heavy Fortran code in PITTNull has been giving ifort trouble every
once in a while. Basically it tries to optimize very hard and sometimes
just fails and gives up (or encounters various errors). Sometimes it
may also use too much memory. It is not clear from the error message
part you included what it would be.

The option list looks ok to me.

Please take a look at the "general guidelines for questionss" in
http://einsteintoolkit.org/support.html

If I were in your shoes I would try these two things:

* compile with -O1 (if that works one can restrict it to just that one
  file see eg what is done in
  repos/mclachlan/ML_ADMConstraints/src/make.code.deps but using -O1
  instead of -fp-model precise)

* compile with -j1 (which might help memory consumption if it is a
  per-user memory limit and not per process)

The full make.log would be helpful.

Yours,
Roland

> Hi,
> I am trying to compile the Einstein Toolkit on MarconiA3 and I am getting
> the following error:
> 
> ifort: error #10106: Fatal error in
> /marconi/prod/opt/compilers/intel/pe-xe-2018/binary/compilers_and_libraries_2018.5.274/linux/bin/intel64/fortcom,
> terminated by kill signal
> compilation aborted for
> /marconi/home/userexternal/lennoggi/ET_2020_11/Cactus/configs/Spritz_UpwindCT/build/NullConstr/NullConstr_R00.f90
> (code 1)
> 
> I have tried to compile both at O3 and O2, but that makes no difference. I
> am attaching the optionlist I am using.
> 
> Do you have any clue about what may be causing the error?
> 
> Thank you,
> Lorenzo



-- 
My email is as private as my paper mail. I therefore support encrypting
and signing email messages. Get my PGP key from http://keys.gnupg.net.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
Url : http://lists.einsteintoolkit.org/pipermail/users/attachments/20210907/1ecad7c6/attachment.bin 


More information about the Users mailing list