SAGESim

Scalable Agent-based GPU Enabled Simulator

Requirements

• Python 3.11+
• NVIDIA GPU with CUDA drivers or AMD GPU with ROCm 5.7.1+

Installation

Your system might require specific steps to installing mpi4py and/or cupy depending on your hardware. In that case use your systems recommended instructions to install mpi4py and cupy and execute:

pip install sagesim

Run Example

• git clone https://code.ornl.gov/sagesim/sagesim
• cd /path/to/clone_repo/examples/sir
• mpirun -n 4 python run.py --num_agents 10000 --percent_init_connections 0.1 --num_nodes 1

There are some unfortunate quirks to using CuPyx jit.rawkernel:

• nan checked by inequality to self. Unfortunate limitation of cupyx.
• Dicts and objects are unsupported.
• *args and **kwargs are unsupported.
• nested functions are unsupported.
• Be sure to use cupy data types and array routines in favor of numpy: [https://docs.cupy.dev/en/stable/reference/routines.html]
• for loops must use range iterator only. No 'for each' style loops.
• return does not seem to work well either
• break and continue are unsupported!
• Cannot reassign variables within if or for statements. Must be assigned at top level of function or new variable declared under subscope.
• -1 indexing does not necessarily work as expected, as it will access the last element of the memory block of the array instead of the logical array. Use len(my_array) - 1 instead