PETSC and HDF5 compilation issues

[andreamarini]

I didn't manage to compile HDF5 using the Intel and PETSC using the CUDA compilers.

Details below:

>ifort --version
ifort (IFORT) 2021.11.0 20231010
Copyright (C) 1985-2023 Intel Corporation.  All rights reserved.

Intel fails while compiling SLEPC. The error is reported below as it is the same of gcc 13.2

>nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Wed_Sep_21_10:33:58_PDT_2022
Cuda compilation tools, release 11.8, V11.8.89
Build cuda_11.8.r11.8/compiler.31833905_0

CUDA compiler instead fails in compiling HDF5.

[attacc]

The log files you sent are empty

…

On Fri, 9 Feb 2024 at 11:35, Andrea Marini ***@***.***> wrote: I didn't manage to compile HDF5 using the Intel and PETSC using the CUDA compilers. Details below: >ifort --version ifort (IFORT) 2021.11.0 20231010 Copyright (C) 1985-2023 Intel Corporation. All rights reserved. Intel fails while compiling SLEPC. Config and compilation logs attached. intel_slepc_compilation_logs.tar.gz <https://github.com/yambo-code/yambo-libraries/files/14220582/intel_slepc_compilation_logs.tar.gz> >nvcc --version nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2022 NVIDIA Corporation Built on Wed_Sep_21_10:33:58_PDT_2022 Cuda compilation tools, release 11.8, V11.8.89 Build cuda_11.8.r11.8/compiler.31833905_0 CUDA compiler instead fails in compiling HDF5. Config and compilation logs attached. cuda_hdf5_compilation_logs.tar.gz <https://github.com/yambo-code/yambo-libraries/files/14220589/cuda_hdf5_compilation_logs.tar.gz> — Reply to this email directly, view it on GitHub <#1>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AED6PCIKQABV6EUURTFD3WTYSX3VLAVCNFSM6AAAAABDBHR4UOVHI2DSMVQWIX3LMV43ASLTON2WKOZSGEZDMOBWG44TMMI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

-- ========================================================== Phone: +33491172806 Mobile[FR]: +33662922897 Mobile[IT]: +393384315437 Fax: +33491418916 Address: CINaM Campus de Luminy 163 Avenue de Luminy, Case 913 13288 Marseille Cedex 9 Skype: claudioattaccalite Editor Fellow for: www.scipost.org web site: www.attaccalite.com ===========================================================

[andreamarini]

Also gcc 13.2.0 with openmpi-5.0.0 fails the slepc compilation

/data/yambo/sources/master.1/lib/slepc/slepc-3.20.1/src/nep/f90-mod/slepcnepmod.F90:61:13:

   61 |         use slepcmfn
      |             1
Fatal Error: Cannot open module file ‘slepcmfn.mod’ for reading at (1): No such file or directory
compilation terminated.
gmake[6]: *** [gmakefile:220: installed-arch-linux2-c-debug-complex/obj/src/nep/f90-mod/slepcnepmod.o] Error 1
gmake[5]: *** [makefile:79: slepc_libs] Error 2
*******************************ERROR************************************
  Error during compile, check installed-arch-linux2-c-debug-complex/lib/slepc/conf/make.log
  Send all contents of ./installed-arch-linux2-c-debug-complex/lib/slepc/conf to [email protected]
************************************************************************
gmake[4]: *** [makefile:51: all] Error 1
make[3]: *** [GNUmakefile:4: all] Error 2
*** Installing SLEPc at prefix location: /opt/libs/yambo/GF-13.2.0_OPENMPI-5.0.0_SIO/MPI-yes_PAR_IO-no_PAR_LINALG-yes/gfortran/mpifort/single  ***

[andreamarini]

The log files you sent are empty

Acc. In the Intel case it appears the same error of gcc.

For CUDA please check and if you need I will reproduce the logs.

Sorry

[sangallidavide]

The slepc compilation with CUDA is an open issue (although I get a different error)

See also here: yambo-code/yambo#63
The discussion involves the Slepc team