| 摘要 |
A plurality of user interactions in a baseline computing system are intercepted and recorded, in an artifact file, with a benchmarking tool. The recorded user interactions from the artifact file are replayed and injected to a software stack of a system under test (whose performance is to be compared to the baseline system), using the benchmarking tool. During the replay, screen updates in the system under test are sequentially compared with expected screen images from the artifact file, until a match is observed in a given case. A response time associated with each of the matches is recorded. |
