Memory usage of any spark task

Question

I developed the algorithm in Spark, and I need to report the memory size of my algorithm throughout its execution for different sizes of input data. In particular, I need maximum memory load and maximum use of the I / O of my Spark job during its execution for each node or for each artist. My questions

• Spark UI provides RDD size on the storage tab. Will adding all RDD sizes be sufficient for memory consumption or will I have to look for other things.
• If I need to compare mine with other algorithms, how to check memory consumption. Because in this case I will not have access to the codes.
• How to calculate peak IO usage. Will shuffle read / write in this case.