Skip to main content

Table 8 Performance comparison of Sequential, Lock-Based, TVar and TArray Sugarscape implementations with varying agent numbers and 50 × 50 environment on 4 cores (except Sequential)

From: A tale of lock-free agents: towards Software Transactional Memory in parallel Agent-Based Simulation

AgentsSequentialLock-BasedTVarTArray
50070.1 (0.41)67.9 (0.13)69.1 (0.34)25.7 (0.42)
1000145.0 (0.11)130.0 (0.28)136.0 (0.16)38.8 (1.43)
1500220.0 (0.14)183.0 (0.83)192.0 (0.73)40.1 (0.25)
2000213.0 (0.69)181.0 (0.84)214.0 (0.53)49.9 (0.82)
2500193.0 (0.16)272.0 (0.81)147.0 (0.32)55.2 (1.04)
  1. Best performance indicated in italic
  2. Timings in seconds (lower is better), standard deviation in parentheses