Performance of the ATLAS Silicon Pattern Recognition Algorithm in Data and Simulation at $\sqrt{s} = 7$ TeV

We study the performance of the silicon pattern recognition algorithms on Monte Carlo simulation and data at $\sqrt{s} = 7$ TeV at different stages of track reconstruction. The seed finding in the silicon detector, the subsequent track candidate building and ambiguity solving procedures are investigated and comparisons are made between data and simulation. A reasonably good description of the data by the MC simulation was found for track seed and track candidate properties. A first comparison of seed to track and generated particle information is also presented.