Abstract
This study explores a teaching methodology combining theoretical and practical components with modern parallel programming technologies. The training process utilized parallel programming languages and tools such as OpenMP, MPI, and CUDA, as well as the SimGrid simulator, which provided students with the opportunity to master real-world high-performance computing technologies. The research evaluates the method’s impact on students’ knowledge, skills, and perceptions of the learning process using tests, questionnaires, and statistical analysis. The study involved 118 students from four Kazakhstani universities. Participants averaged 20 years of age. The main problem of the study is that traditional approaches to teaching numerical methods do not meet the needs of the development of information and communication technologies in Kazakhstan, where specialists with parallel and high-performance computing skills are needed. Results showed a significant improvement in student performance, with theoretical exam scores increasing by 38.1% and practical scores by 45.9%. Statistical analysis confirmed these gains, validating the method’s effectiveness. Surveys further indicated enhanced student perceptions of program usefulness, satisfaction, and material comprehension. The methodology integrates modern technologies and tools, significantly improving theoretical and practical knowledge in parallel computing.