It depends on which kind of algorithms you used (data analysis, image processing, video processing, etc). Anyway you can analyse the accuracy of your model, response time and memory space (volume). For framework you can use Spark (graphx, MLLIB, streaming, etc.)
The Performance evaluation of big data frameworks for large-scale data analytics. the behavior of frameworks has been characterized by parameters of the workloads such as block size, input data size, interconnect network or thread configuration. The analysis of the results has shown that replacing Hadoop with Spark or Flink ..