Appearance
ForgeCAT UTC Distributed Simulation
This benchmark measures distributed Monte Carlo execution for the ForgeCAT UTC cyclone simulation adapter.
Workload
Monte Carlo simulation of cyclone intensity distributions.
Model:
UTC-1.0
Simulation size:
100,000,000 iterations
Network Topology
Agents used: 20
Shard size: 5,000,000 iterations per agent
Measured Result
Iterations: 100,000,000
Agents: 20
Wall time:
0.305 seconds
Derived Metrics
Aggregate throughput:
≈ 327 million iterations/sec
Per-agent throughput:
≈ 16.3 million iterations/sec
Scaling Projection
Assuming comparable per-agent performance:
| Agents | Expected Throughput | Estimated Time for 1B |
|---|---|---|
| 20 | ~327M/s | ~3.06 s |
| 60 | ~980M/s | ~1.02 s |
| 120 | ~1.96B/s | ~0.51 s |
Interpretation
Measured results demonstrate that lightweight stochastic workloads can achieve hundreds of millions of iterations per second on modest Forge networks.
Under similar conditions, billion-iteration simulations approach sub-second execution with moderate network width.
Boundary
Results apply specifically to the ForgeCAT UTC workload and should not be generalized to all simulation models.