The "__GPU_MPI" flag is not strictly necessary to compile and run the MPI
parallel versions with GPU.
It only enables more efficient communications when the MPI library has been
compiled with the necessary support for GPU-aware communications. The only
essential flag for MPI is the -D__MPI.
The failure's origin might be that your MPI library lacks such support. But you
had better ask your system administrator about this.
You can remove the GPU_MPI flag and see if the code works.
Hope it helps
Pietro
________________________________
Da: users <users-boun...@lists.quantum-espresso.org> per conto di Siwakorn
Sukharom <siwakorn....@ncr.nstda.or.th>
Inviato: giovedì 25 maggio 2023 06:34
A: Paolo Giannozzi <paolo.gianno...@uniud.it>; Quantum ESPRESSO users Forum
<users@lists.quantum-espresso.org>
Oggetto: Re: [QE-users] R: [QE-GPU] compilation issue :
NVFORTRAN-F-1225-Unmatched directive
Now, I can compiled GPU version for QE 7.2. Thank you for the suggestion. It
can run fine on 1 gpu. However, I still having a problem with running multiple
gpus. I have added the --with-cuda-mpi=yes during the configuration, and I can
see that there is -D__GPU_MPI tag on DFLAGS in make.inc
When compiled and run on multiple gpus, I get the error message below:
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Error in routine fft_buffers (4):
Allocation failed
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
stopping ...
MPICH ERROR [Rank 0] [job id 77558.0] [Thu May 25 11:11:15 2023]
[x1000c2s0b0n1] - Abort(4) (rank 0 in comm 0): application called
MPI_Abort(MPI_COMM_WORLD, 4) - process 0
aborting job:
application called MPI_Abort(MPI_COMM_WORLD, 4) - process 0
What is happening?
________________________________
From: Paolo Giannozzi <paolo.gianno...@uniud.it>
Sent: Wednesday, May 24, 2023 1:05:38 PM
To: Siwakorn Sukharom; Quantum ESPRESSO users Forum
Subject: Re: [QE-users] R: [QE-GPU] compilation issue :
NVFORTRAN-F-1225-Unmatched directive
On 24/05/2023 05:47, Siwakorn Sukharom wrote:
>
> I have added -D__OPENACC to DFLAGS for QE 7.2 and 7.1
"-D_OPENACC", with one underscore
Paolo
. The compilation
> was failed for QE 7.2, but success for QE 7.1. The
> https://linkprotect.cudasvc.com/url?a=https%3a%2f%2fmake.inc&c=E,1,RCjygUVrHz_y_vPiylqafXnFbjj3qIZgqer4PcMPTMAL1nWOb0F2WH7i2o4bEwXkmPY9pBKdip8024dPJMlrqTn00xrWzoEusar_ykVw&typo=1
> is almost
> the same.
>
> ------------------------------------------------------------------------
> *From:* Paolo Giannozzi <paolo.gianno...@uniud.it>
> *Sent:* Tuesday, May 23, 2023 5:29:09 PM
> *To:* Quantum ESPRESSO users Forum; Siwakorn Sukharom
> *Subject:* Re: [QE-users] R: [QE-GPU] compilation issue :
> NVFORTRAN-F-1225-Unmatched directive
> On 5/23/23 12:13, Siwakorn Sukharom wrote:
>
>> I can successfully compile on QE version 7.1, but not 7.2. The error
>> is still the same. Will focus on 7.1 instead.
>
> Not a good idea. You just need to add _OPENACC to DFLAGS (something that
> the Cray compiler should take care of, but it doesn't).
>
> Paolo
> --
> Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
> Univ. Udine, via delle Scienze 208, 33100 Udine Italy, +39-0432-558216
>
> ------------------------------------------------------------------------
>
> Disclaimer:
>
> This e-mail and any files transmitted with it may contain confidential
> and proprietary information of the National Science and Technology
> Development Agency (NSTDA), Thailand. They are intended solely for the
> use of the addressed individuals or entities. If you are not the
> intended recipient, you are required to immediately delete this e-mail
> and its contents from your system. Any disclosure, distribution, or
> action based upon the contents of this e-mail is strictly prohibited.
> Any views or opinions presented in this e-mail are solely those of the
> sender and do not necessarily represent those of NSTDA. NSTDA does not
> accept any responsibility for the content of this message or the
> consequences of any actions taken on the basis of the information
> provided. NSTDA accepts no liability for any damage caused by any virus
> or malware which may be inserted in this e-mail during transmission.
--
Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
Univ. Udine, via delle Scienze 206, 33100 Udine Italy, +39-0432-558216
________________________________
Disclaimer:
This e-mail and any files transmitted with it may contain confidential and
proprietary information of the National Science and Technology Development
Agency (NSTDA), Thailand. They are intended solely for the use of the addressed
individuals or entities. If you are not the intended recipient, you are
required to immediately delete this e-mail and its contents from your system.
Any disclosure, distribution, or action based upon the contents of this e-mail
is strictly prohibited. Any views or opinions presented in this e-mail are
solely those of the sender and do not necessarily represent those of NSTDA.
NSTDA does not accept any responsibility for the content of this message or the
consequences of any actions taken on the basis of the information provided.
NSTDA accepts no liability for any damage caused by any virus or malware which
may be inserted in this e-mail during transmission.
_______________________________________________
The Quantum ESPRESSO community stands by the Ukrainian
people and expresses its concerns about the devastating
effects that the Russian military offensive has on their
country and on the free and peaceful scientific, cultural,
and economic cooperation amongst peoples
_______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing list users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users
