Dear users,
I would like to use the GPU package. I followed the steps suggested in the README file.I obtained the nvc_get_devices file, and it looks like this:
Found 1 platform(s).
Using platform: NVIDIA Corporation NVIDIA CUDA Driver
CUDA Driver Version: 9.10
Device 0: “GeForce GTX 1050 Ti”
Type of device: GPU
Compute capability: 6.1
Double precision support: Yes
Total amount of global memory: 3.94537 GB
Number of compute units/multiprocessors: 6
Number of cores: 1152
Total amount of constant memory: 65536 bytes
Total amount of local/shared memory per block: 49152 bytes
Total number of registers available per block: 65536
Warp size: 32
Maximum number of threads per block: 1024
Maximum group size (# of threads per block) 1024 x 1024 x 64
Maximum item sizes (# threads for each dim) 2147483647 x 65535 x 65535
Maximum memory pitch: 2147483647 bytes
Texture alignment: 512 bytes
Clock rate: 1.62 GHz
Run time limit on kernels: Yes
Integrated: No
Support host page-locked memory mapping: Yes
Compute mode: Default
Concurrent kernel execution: Yes
Device has ECC support enabled: No
After that, I build the lammps app with the others packages I need.
I made a test to see if the app works propertly. The input is:
units lj
atom_style atomic
lattice sc 0.1111 origin 0.5 0.5 0.5
region box block 0 30 0 30 0 30 units box
create_box 1 box
create_atoms 1 box
mass 1 1.0
velocity all create 2.0 87287 dist gaussian
pair_style lj/cut 2.5
pair_coeff 1 1 1.0 1.0 2.5
neighbor 0.2 bin
neigh_modify every 10 delay 0 check no
fix 1 all nve
thermo 50
run 2500
And it only works propertly if I dont use the gpu package. When I modified the input file to run with gpu package ( pair_style lj/cut/gpu 2.5) it looks like:
LAMMPS (12 Dec 2018)
Lattice spacing in x,y,z = 2.08015 2.08015 2.08015
Created orthogonal box = (0 0 0) to (30 30 30)
1 by 2 by 2 MPI processor grid
Created 2744 atoms
Time spent = 0.000245094 secs