Sorted by memory efficiency.
Memory efficiency is the usable JVM heap after Databricks overhead as a share of total instance RAM.
Tip: Click on an instance card to add it to comparison.
------------------------------------------------- The following formulae are used for calculations: ------------------------------------------------- jvm_heap_memory_mb = (instance.ram * 1024 * 0.97 - 4800) * 0.8; reserved_memory_mb = 300 mb spark_memory_fraction = 0.6 user_memory_fraction = 1-spark_memory_fraction = 0.4 spark_execution_memory_fraction = 0.5 spark_storage_memory_fraction = 1 - spark_execution_memory_fraction = 0.5