Linux

A compatible NVIDIA graphics driver (version 450.80.02 or newer) must be installed on Linux operating systems. CUDA-enabled GPUs with compute capabilities of 3.5 or higher are supported. A specifically tuned build of Open MPI is bundled in the installation and will be used for running nanoFluidX.

Single Node

After installation at <INSTALL_DIR>, follow these steps to verify nanoFluidX has been installed correctly.

  1. Login as a regular user. Do not use root user.
  2. Set environment source to: 
    <INSTALL_DIR>/altair/hwcfdsolvers/nanofluidx/linux64/set_nFX_environment.sh
  3. Run system tests with the following: 
    system_tests.py
  4. A successful test on a multi-GPU node will produce output similar to the following:
    Figure 1.
  5. Copy dambreak case from demo_cases to the location of choice: 
    <LOC>: cp <INSTALL_DIR>/ altair/hwcfdsolvers/nanofluidx/linux64/demo_cases/dambreak <LOC>/dambreak.
  6. Change directory to <LOC>: cd <LOC>.
  7. Run the following: 
    nanoFluidX:nfxmpirun_singlenode -np 1 $nFX_SP -i dambreak_3D_H300_dx0p005_wt.cfg -m 10
  8. The test run is successful when the message Finished (reached maximum steps) appears at the end.
    Note: Simulation time may vary.
    Figure 2.

Multi Node

After installation at <INSTALL_DIR> on each node, test the installation on each node separately. In addition, follow these steps to verify nanoFluidX has been installed correctly for multi-node operation.

  1. Login as a regular user. Do not use root user.
  2. Set environment source to: 
    <INSTALL_DIR>/altair/hwcfdsolvers/nanofluidx/linux64/ set_nFX_environment.sh multinode
  3. Copy dambreak case from demo_cases to the location of choice $LOC:
    1. For systems with shared storage: cp 
      cp <INSTALL_DIR>/altair/hwcfdsolvers/nanofluidx/ linux64/demo_cases/dambreak <LOC>/dambreak
    2. Repeat the copy operation for all nodes with local storage.
  4. Change directory to <LOC>: cd <LOC>
  5. Run the following: 
    nanoFluidX: nfxmpirun_multinode --hostfile HOSTFILE -np X $nFX_SP -i dambreak_3D_H300_dx0p005_wt.cfg -m 10
    1. Use a suitable HOSTFILE where each node is repeated once per GPU on the node.
    2. Set X equal to the number of GPUs across all nodes.
  6. The test run is successful when the message Finished (reached maximum steps) appears at the end.
    Figure 3.