#189 fails to build with latest Intel meson (Ubuntu-22.04,ifort,icc)

[alan-arnold]

Hello again Christoph. I've been using your crest-master code with the updated QCG settings for ensemble generation #183 . Thanks for making these changes in response to #181. This compiles, builds & runs OK with the 2022.3.1 versions of Intel oneAPI Basekit and HPC toolkit on Ubuntu 22.04 with ifort and icc versions 2021.7.1.

I'm of course interested to take a look at your #189 fix that introduced the -normdock flag for a more extensive QCG grow search but, unfortunately, overnight Ubuntu updated to the latest 2023.1.0 oneAPI suite, which includes ifort and icc versions 2021.9.0, and crest-master now fails to build. I haven't changed my meson setup in any way.

As a temporary workaround, I'll try to revert to the older ifort and icc versions 2021.7.1, but in the meantime I thought you might like to know that there seems to be an issue with confscript3.f90:

~$ cd /usr/bin/crest-master
crest-master$ export FC=ifort CC=icc
crest-master$ meson setup build --prefix=$PWD/_dist -Dfortran_link_args="-lifcoremt -static -qopenmp" -Ddefault_library=static -Dla_backend=mkl --wipe
The Meson build system
Version: 0.64.1
Source dir: /usr/bin/crest-master
Build dir: /usr/bin/crest-master/build
Build type: native build
Project name: crest
Project version: 2.12.0
Fortran compiler for the host machine: ifort (intel 2021.9.0 "ifort (IFORT) 2021.9.0 20230302")
Fortran linker for the host machine: ifort ld.bfd 2.38
C compiler for the host machine: icc (intel 2021.9.0 "icc (ICC) 2021.9.0 20230302")
C linker for the host machine: icc ld.bfd 2.38
Host machine cpu family: x86_64
Host machine cpu: x86_64
Run-time dependency OpenMP found: YES 5.0-revision1
Library mkl_intel_lp64 found: YES
Library mkl_intel_thread found: YES
Library mkl_core found: YES
Program asciidoctor found: YES (/usr/bin/asciidoctor)
Configuring crest.1 with command
Build targets in project: 1

crest 2.12.0

  User defined options
    default_library  : static
    prefix           : /usr/bin/crest-master/_dist
    fortran_link_args: -lifcoremt -static -qopenmp
    la_backend       : mkl

Found ninja-1.11.1.git.kitware.jobserver-1 at /home/parallels/.local/bin/ninja
crest-master$ meson install -C build
ninja: Entering directory `/usr/bin/crest-master/build'
[44/86] Compiling Fortran object crest.p/src_confscript3.f90.o
FAILED: crest.p/src_confscript3.f90.o 
ifort -Icrest.p -I. -I.. -D_FILE_OFFSET_BITS=64 -warn general -warn truncated_source -g -traceback -O2 -g -traceback -DWITH_MKL -qopenmp -module crest.p -gen-dep=crest.p/src_confscript3.f90.o -gen-depformat=make -o crest.p/src_confscript3.f90.o -c ../src/confscript3.f90
../src/confscript3.f90: catastrophic error: **Internal compiler error: segmentation violation signal raised** Please report this error along with the circumstances in which it occurred in a Software Problem Report.  Note: File and line given may not be explicit cause of this error.
compilation aborted for ../src/confscript3.f90 (code 1)
[69/86] Compiling Fortran object crest.p/src_confparse.f90.o
ninja: build stopped: subcommand failed.
Could not rebuild /usr/bin/crest-master/build
crest-master$

[alan-arnold]

I reverted to ifort/icc 2021.6.0 from the 2022.1.0 oneAPI suite and this now compiles OK.

BTW it's a bit of a pain to go back to old saved output files and tell which exact version of crest was used for that run. For example, the header from this latest codebase (3-Apr) which I compiled today (4-Apr) looks like this

   ==============================================
   |                                            |
   |                 C R E S T                  |
   |                                            |
   |  Conformer-Rotamer Ensemble Sampling Tool  |
   |          based on the GFN methods          |
   |             P.Pracht, S.Grimme             |
   |          Universitaet Bonn, MCTC           |
   ==============================================
   Version 2.12, Mo 30. Jan 18:49:37 GMT 2023

Can you include a date stamp here which shows when the version was actually compiled? Like xtb does ;)

[github-actions[bot]]

This issue had no activity for 6 months. It will be closed in 1 week unless there is some new activity.