which recording mode(HTML or URL) is faster execution?...

HP LoadRunner Software Forum

This particular user exhibits a lack of performance testing knowledge where they are looking for users to execute as fast as possible. This implies no think time and no pacing set for the user. This has the effect of collapsing the client-server model, which uses as one of it's foundation sets the gap in requests from one user to support the requests and responses from the next. By closing these gaps to zero, you wind up with a model which is in no way representative of end user behavior and also cannot be used as a predictor of expected behavior in production.

Such tests also typically produce engineering ghosts: Performance defects which appear in test but which would never appear in production due to the behavior of users in test vs users in prod.

This user appears, by virtue of the question, not to understand the value of pacing users to match a production level of end user behavior. This question is also commonly tied to the use of the community or evaluation licenses (limited number of users) and a desire to run users as fast as possible to emulate the transactional load of a larger population of users. This assumption fails because the session load, the resources set aside for each session, are not comparable to production either.

which recording mode(HTML or URL) is faster execution?...

Be wary of the tester who says to you, "well one of the test users represents n users in production. Politely show that tester the door.