Figure 6 shows our results for parsing Java source files.
The benchmark is based on 33 source files of lengths varying
between 0 and 800 lines of code, randomly selected from
the JSGLR and Stratego [4] project. We ran the parser three
times per test case and plotted the average run-times. The
results show linear behavior for the original JSGLR [22]
parser and for the JavaScript port, but also show a significant
performance difference for larger source files.