mirror of https://gitlab.com/QEF/q-e.git
Merge branch 'smallfixes' into 'develop'
CUDA MPI problem See merge request QEF/q-e!1798
This commit is contained in:
commit
e9faeebabf
|
@ -645,11 +645,10 @@ available:\\
|
|||
& CUDA Toolkit installed on the workstation \\
|
||||
& or available on the compute nodes of the HPC facility.\\
|
||||
\texttt{--with-cuda-mpi=value} & enable usage of a CUDA-aware MPI library (default: no).\\
|
||||
\end{tabular}{ll}
|
||||
\end{tabular}
|
||||
|
||||
To modify or extend \configure, see the Wiki pages on GitLab:
|
||||
\texttt{https://gitlab.com/QEF/q-e/-/wikis}.
|
||||
For advanced users only!
|
||||
To modify or extend \configure\ (advanced users only!), see the Wiki pages
|
||||
on GitLab: \texttt{https://gitlab.com/QEF/q-e/-/wikis}.
|
||||
|
||||
\subsubsection{Manual configuration}
|
||||
\label{SubSec:manconf}
|
||||
|
@ -665,7 +664,8 @@ if you want to link internal FFTW). For a correct choice of preprocessing
|
|||
flags, refer to the documentation in \texttt{include/defs.h.README}.
|
||||
|
||||
Even if \configure\ works, you may need to tweak the \texttt{make.inc}
|
||||
file. It is very simple, but please note that if you change any settings
|
||||
file. It is very simple, but please note that a) you must know what you are
|
||||
doing, and b) if you change any settings
|
||||
(e.g. preprocessing, compilation flags)
|
||||
after a previous, successful or failed, compilation, you must run
|
||||
\texttt{make clean} before recompiling, unless you know exactly which
|
||||
|
@ -676,7 +676,7 @@ unless you use option \texttt{--save}.
|
|||
\subsection{Libraries}
|
||||
\label{Sec:Libraries}
|
||||
|
||||
\qe\ contains a copy of some needed external libraries:
|
||||
\qe\ downloads a copy of the following external libraries if needed:
|
||||
\begin{itemize}
|
||||
\item FoX for reading and writing xml files;
|
||||
\item BLAS (\texttt{http://www.netlib.org/blas/}) and LAPACK
|
||||
|
|
12
README.md
12
README.md
|
@ -7,8 +7,10 @@ Optimization)
|
|||
[![License: GPL v2](https://img.shields.io/badge/License-GPL%20v2-blue.svg)](https://www.gnu.org/licenses/old-licenses/gpl-2.0.en.html)
|
||||
|
||||
## USAGE
|
||||
Quick installation instructions for the impatient. Go to the directory
|
||||
where this file is. Using "make"
|
||||
Quick installation instructions for CPU-based machines. For GPU execution, see
|
||||
file [README_GPU.md](README_GPU.md). Go to the directory where this file is.
|
||||
|
||||
Using "make"
|
||||
(`[]` means "optional"):
|
||||
```
|
||||
./configure [options]
|
||||
|
@ -35,7 +37,8 @@ is prepended onto all install directories.
|
|||
|
||||
For more information, see the general documentation in directory Doc/,
|
||||
package-specific documentation in \*/Doc/, and the web site
|
||||
http://www.quantum-espresso.org/. Documentation for developers
|
||||
http://www.quantum-espresso.org/. Technical documentation for users and
|
||||
developers
|
||||
can be found on [Wiki page on gitlab](https://gitlab.com/QEF/q-e/-/wikis/home).
|
||||
|
||||
## PACKAGES
|
||||
|
@ -62,9 +65,6 @@ The following libraries have been isolated and partially encapsulated in view of
|
|||
- LRlib: performs a variety of tasks connected with (time-dependent) DFPT, to be used also in connection with Many-Body Perturbation Theory.
|
||||
- upflib: pseudopotential-related code.
|
||||
|
||||
## GPU-enabled version
|
||||
Since Feb.2021 this repository also works for GPU's (currently only NVIDIA). See file [README_GPU.md](README_GPU.md).
|
||||
|
||||
## Contributing
|
||||
Quantum ESPRESSO is an open project: contributions are welcome.
|
||||
Read the [Contribution Guidelines](CONTRIBUTING.md) to see how you
|
||||
|
|
|
@ -4839,7 +4839,7 @@ END SUBROUTINE mp_type_free
|
|||
#if defined(__GPU_MPI)
|
||||
msglen = size(msg_d)
|
||||
ierr = cudaDeviceSynchronize() ! This syncs __GPU_MPI
|
||||
CALL reduce_base_real_to_gpu( 2 * msglen, msg_d, res_h, gid, -1 )
|
||||
CALL reduce_base_real_to_gpu( 2 * msglen, msg_d, res_d, gid, -1 )
|
||||
RETURN ! Sync not needed after MPI call
|
||||
#else
|
||||
ALLOCATE( msg_h, source=msg_d ) ! This syncs __MPI case
|
||||
|
|
Loading…
Reference in New Issue