Benchmarking GPT-4 on Algorithmic Problems: A Systematic Evaluation of Prompting Strategies