[QE-users] [QE-GPU] Running GPU calculations optimally

[Anson Thomas]

Dear all,
I am trying to use Quantum ESPRESSO v 7.1 with GPU acceleration.
However, since I am new to this field, I need to learn about the flags and
variables that could be exported to achieve optimum performance.

Here is a bunch of variables that were suggested to me:

export GPU_FORCE_64BIT_PTR=0
export GPU_MAX_HEAP_SIZE=100
export GPU_USE_SYNC_OBJECTS=1
export GPU_MAX_ALLOC_PERCENT=100
export GPU_SINGLE_ALLOC_PERCENT=100

export OMP_NUM_THREADS=1
export MPI_PER_GPU=20
export ENABLE_MPS=false
export LS_HYPERTHREAD=true

I could not find the documentation about these variables, nor do I
understand how these work.
The GPUs are not being utilized fully. I say this since I tried to run some
basic profiling, all of which indicate the same:

[image: image.png]


I am using a SLURM  script to run my jobs. Every GPU node on the cluster
has 2 Tesla V100-SXM2-16GB and 40CPU cores.
The job is launched using "mpirun -np $SLURM_NTASKS pw.x -ndiag 1 -in
input.in > output.out"

Could anyone shed some light on the following:
1. What flags and variables could be set, and how do I decide which are
relevant for my calculations?
2. Where can I read more about these variables and flags to test which ones
are relevant for the calculations?
3. Can the mpirun command be modified for better performance?

Any insights would be very helpful.



Sincerely,
*Anson Thomas*
Research Scholar,
Department of Chemistry,
Indian Institute of Technology, Roorkee
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20231114/ace61862/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 54296 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20231114/ace61862/attachment.png>