Dynamic distribution of simulation load

A method comprising:

• assigning ownership of objects of a simulation scenario to a first compute node;
• identifying that a processing load of the first compute node has reached a threshold;
• determining how many clusters of the objects are owned by the first compute node;
• re-clustering the objects, wherein a second quantity of clusters is determined that exceeds a first quantity of one or more clusters owned by the first compute node;
• forming the second quantity of clusters, wherein the forming the second quantity of clusters comprises using the second quantity as a selected quantity of cluster centroids for the second quantity of clusters;
• and
• reassigning the ownership of the objects to a plurality of compute nodes including the first compute node and a second compute node, wherein the reassigning causes at least one cluster included in the second quantity of clusters to be owned by the second compute node.

A non-transitory computer-readable medium having stored thereon a set of instructions, which when performed by one or more processors, causes the one or more processors to perform operations comprising:

• assigning ownership of objects of a simulation scenario to a first compute node;
• identifying that a processing load of the first compute node has reached a threshold;
• determining how many clusters of the objects are owned by the first compute node;
• re-clustering the objects, wherein a second quantity of clusters is determined that exceeds a first quantity of one or more clusters owned by the first compute node;
• forming the second quantity of clusters, wherein the forming the second quantity of clusters comprises using the second quantity as a selected quantity of cluster centroids for the second quantity of clusters;
• and
• reassigning the ownership of the objects to a plurality of compute nodes including the first compute node and a second compute node, wherein the reassigning causes at least one cluster included in the second quantity of clusters to be owned by the second compute node.