The efficiency and modes of usage of computerized tools (DGE and some other types of tools) were studied in relation to: students' abilities; the routineness of a problem; and the type of course (all three parameters defined in the paper). It was found that there was no significant difference in modes of computer usage for routine problems; a significant difference appeared only in the group where the students had been taught to use these tools 'in real time', i.e. when it was found appropriate by the lecturer during a geometry lesson. Besides that it was found that the percentage of high-level students who improved their results with non-routine problems due to use of the computer is significantly higher than that of lower-level students.