Applying efficient memory access and parallelization techniques still is not sufficient to efficiently handle the huge processing loads caused by large datasets. We presented algorithmic optimizations to reduce this workload. We introduced three techniques which each can achieve a significant reduction of rendering times. Our goal was to minimize the additional memory requirements of newly introduced data structures.