In Spark can we write intermediate data to executor memory , than writing to spark.loca.dir location?

Question

Hi We have spark cluster , during spark job execution , am getting sparkoutofmemory when writing intermediate data to spark.loca.dir location , but when am seeing their is