[CP2K-user] [CP2K:18845] Minimal build of cp2k-2023.1 fails

[Scott Milner]

**These instructions may be useful to others trying to build cp2k on older 
clusters (mine runs Centos 7, installed in 2018).**

I was never able to successfully build local.ssmp.  However, I was able to 
build local.psmp, as follows:

1) I installed a more recent gcc, version 7.5.0, by building it from source 
downloaded from GNU project.

2) I used toolchain, and requested that the compilers be installed (this 
installs gcc 12):

     install_cp2k_toolchain.sh --with-fftw --with-openblas 
--mpi-mode=openmpi --with-gcc=install

3) Then 

     make ARCH=local VERSION=psmp 

works without errors.



On Tuesday, May 16, 2023 at 7:02:49 AM UTC-4 Milner, Scott Thomas wrote:

> Dear Matthias — Thanks for your prompt reply.  The reason I am building 
> from source is that I want a version with PLUMED.  Building this minimal 
> version is a preliminary, to make sure the build works.  What I ultimately 
> want is a build that has PLUMED, and runs on a single node; my cluster does 
> not have fast interconnects, so multinode jobs are not practical there). 
>  If I understand correctly, this means the build has no need for MPI, hence 
> the flag “—mpi-mode=no”:  is that correct? 
>
> As you can see from my email, I built everything from scratch, with no 
> “=system”, precisely to avoid any incompatibilities.  It may be that 
> somehow in previous build attempts I created some “left-overs” as you say, 
> that are interfering with the current attempt.  I will start from a 
> completely fresh unpack of the cp2k tarball and try once more. 
>
> Finally, is there some way to deal with the failure of the toolchain 
> script to download necessary packages from cp2k.org (because of expired 
> certificates), other than to download each one by hand?
>
> Best regards
>
> Scott 
>
> On May 16, 2023, at 4:19 AM, Krack Matthias <matthia... at psi.ch> wrote:
>
> You don't often get email from matthia... at psi.ch. Learn why this is 
> important <https://aka.ms/LearnAboutSenderIdentification>
>
> Dear Scott
>  
> there is no need to build a serial cp2k.ssmp binary, because you can 
> download a pre-compiled binary here 
> <https://github.com/cp2k/cp2k/releases/download/v2023.1/cp2k-2023.1-Linux-gnu-x86_64.ssmp>
> .
> Note, that such a serial cp2k binary is mostly useful for testing or 
> education, but not for production runs.
>  
> I also have the gcc version 7.5.0 installed on my PC as the default system 
> compiler. With that, I can compile and run the regression tests for 
> cp2k-2023.1 without any issue using the commands you reported. So I suspect 
> that the issues you are seeing are related to the environment configuration 
> on your system. Maybe, there are left-overs from previous installation 
> attempts which interfere now with the current installation. Try to start 
> from scratch (new login session, new cp2k folder created using “git clone” 
> or “tar -xf’).
>  
> Best regards
>  
> Matthias
>  
>
> *From: *cp... at googlegroups.com <cp... at googlegroups.com> on behalf of 
> Scott Milner <st... at psu.edu>
> *Date: *Tuesday, 16 May 2023 at 04:12
> *To: *cp2k <cp... at googlegroups.com>
> *Subject: *[CP2K:18815] Minimal build of cp2k-2023.1 fails
> I am trying to build cp2k-2023.1 on x86-64 Linux architecture using 
> gcc-7.5.0 
> (which is listed as working, see https://www.cp2k.org/dev:compiler_support
> ).
>  
> I tried the absolute simplest build possible:
>      install_cp2k_toolchain.sh --with-fftw --with-openblas --mpi-mode=no
> which proceeded without errors (except that I had to download all the .b2z 
> files for the packages by hand, because the certificates have expired -- 
> grrrr).
>  
> But the compile failed:  after the obligatory
>      source 
> /home/stm9/group/SOFTWARE/cp2k-2023.1/tools/toolchain/install/setup
> and copying the architecture files to where they need to go,
> the make command
>      make -j 32 ARCH=local VERSION=ssmp
> generated errors:
>  
> /data1/group/SOFTWARE/cp2k-2023.1/lib/local/ssmp/libcp2kmpiwrap.a(message_passing.o): 
> In function `__message_passing_MOD_mp_abort':
> /data1/group/SOFTWARE/cp2k-2023.1/src/mpiwrap/message_passing.F:1156: *undefined 
> reference to `_gfortran_stop_numeric_f08'*
> collect2: error: ld returned 1 exit status
> make[3]: *** 
> [/data1/group/SOFTWARE/cp2k-2023.1/exe/local/parallel_rng_types_unittest.ssmp] 
> Error 1
> make[2]: *** [all] Error 2
> make[1]: *** [ssmp] Error 2
> make: *** [all] Error 2
>  
> This is the most frustrating build I have dealt with in a long time.  I 
> would much appreciate it if a knowledgeable CP2K-er can help build this app.
>  
> Scott Milner
> Professor of Chemical Engineering
> Penn State University
>  
>  
>  
> -- 
> You received this message because you are subscribed to the Google Groups 
> "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to cp2k+uns... at googlegroups.com.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/cp2k/c645fb64-9ccc-4c92-94e6-2ab06318d8c6n%40googlegroups.com 
> <https://groups.google.com/d/msgid/cp2k/c645fb64-9ccc-4c92-94e6-2ab06318d8c6n%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
> -- 
>
> You received this message because you are subscribed to a topic in the 
> Google Groups "cp2k" group.
> To unsubscribe from this topic, visit 
> https://groups.google.com/d/topic/cp2k/V3kMoK1bDj0/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to 
> cp2k+uns... at googlegroups.com.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/cp2k/ZRAP278MB082776353E727D6B9342A0E1F4799%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM 
> <https://groups.google.com/d/msgid/cp2k/ZRAP278MB082776353E727D6B9342A0E1F4799%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM?utm_medium=email&utm_source=footer>
> .
>
>
>

-- 
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/bd29ed4f-9b26-4a91-9858-6c00f7a545f2n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20230524/a428f0a0/attachment-0001.htm>