The GPU version on the NVIDIA CUDA platform is fully incorporated into the main trunk. Currently a subset of QMCPACK that is commonly used for solid-state and molecular systems using bspline single-particle orbitals is supported.

Supported features

Boundary conditions:

• Periodic and open boundary conditions are supported, but not mixed boundary conditions
• Complex wave functions (e.g. fixed phase) are not yet supported

Wavefunctions:

• Single Slater determinants with 3D B-spline orbitals. Only real-valued wave functions is supported, but tiling complex orbitals to supercells is supported as long as each k-point is a multiple of half a G-vector of the supercell.
• Mixed basis representation in which orbitals are represented as:
  • 1D splines times spherical harmonics in spherical regions (muffin tins) around atoms
  • 3D B-splines in the interstitial region
• One-body and two-body Jastrows represented as 1D B-splines are supported.

Hamiltonians:

• Nonlocal pseudopotentials (semilocal form)
• Coulomb interaction
• Model Periodic Coulomb (MPC) interaction

Drivers:

• Variational Monte Carlo (VMC)
• Wave function optimization. Only Jastrow can be optimized at present
• Diffusion Monte Carlo (DMC)

Compiling the GPU code

• First, compile and install the einspline library. When building, run
  ./configure --enable-cuda ...
• To include GPU support in QMCPACK, one of these methods can be used. The nvcc compiler should be in your path.
  • Run cmake with the argument
    -DQMC_CUDA=1.
  • If using toolchains, set QMC_CUDA in the toolchain file as
    SET(QMC_CUDA 1)

Using the GPU code

• When running QMCPACK with GPUs, you should run one MPI process per GPU on each node. Add the argument --gpu to the qmcapp command line, e.g.,
  mpirun -np 16 qmcapp --gpu myfile.xml
• Inside myfile.xml, you need to add the attribute gpu="yes" in several places:
• Inside the <determinantset> element,e.g.
  
  <determinantset type="einspline" href="cBN_128_V40.TN.h5"
sort="1" tilematrix="4 0 0 0 4 0 0 0 4"
twistnum="1" gpu="yes" source="i">
• Inside each of the<qmc> elements, e.g.
  
  <qmc method="vmc" move="pbyp" gpu="yes">
• Finally, the GPU code requires many walkers to work efficiently. Presently, 128-256 walkers gets good efficiency. You may be limited by the amount of memory on the GPU card.