Abstract | ||
---|---|---|
Large clusters and supercomputers are simulated to aid in design. Many devices, such as hard drives, are slow to simulate. Our approach is to use a genetic algorithm to fit parameters for an analytical model of a device. Fitting focuses on aggregate accuracy rather than request-level accuracy since individual request times are irrelevant in large simulations. The model is fitted to traces from a physical device or a known device-accurate model. This is done once, offline, before running the simulation. Execution of the model is fast, since it only requires a modest amount of floating point math and no event queueing. Only a few floating point numbers are needed for state. Compared to an event-driven model, this trades a little accuracy for a large gain in performance. |
Year | DOI | Venue |
---|---|---|
2011 | 10.1145/2148600.2148617 | SC Companion |
Keywords | Field | DocType |
fast performance model,event-driven model,analytical model,large cluster,floating point number,large simulation,device-accurate model,large gain,aggregate accuracy,request-level accuracy,floating point math | Cluster (physics),Floating point,Computer science,Parallel computing,Queueing theory,Genetic algorithm,Distributed computing | Conference |
Citations | PageRank | References |
0 | 0.34 | 1 |
Authors | ||
12 |
Name | Order | Citations | PageRank |
---|---|---|---|
Adam Crume | 1 | 190 | 8.09 |
Carlos Maltzahn | 2 | 1201 | 87.49 |
Jason Cope | 3 | 255 | 11.16 |
Sam Lang | 4 | 0 | 0.34 |
Rob Ross | 5 | 18 | 4.81 |
Phil Carns | 6 | 0 | 0.34 |
Chris Carothers | 7 | 0 | 0.34 |
Ning Liu | 8 | 201 | 8.13 |
Curtis Janssen | 9 | 0 | 0.34 |
John Bent | 10 | 95 | 9.02 |
Stephan Eidenbenz | 11 | 1302 | 103.03 |
Meghan McClelland | 12 | 0 | 0.34 |