Cost of accessing an object in Scala

Question

A bit of context:

case class Finger(length: Int)
case class Arm(thumb: Finger)
case class Body(rightArm: Arm, leftArm: Arm)
case class Human(body: Body)

def processHumanFinger(human: Human) = println(human.body.rightArm.thumb.length)

def processFingerOnly(finger: Finger) = println(finger.length)

val john = Human(Body(Arm(Finger(1)), Arm(Finger(3))))
val johnFinger = john.body.rightArm.thumb.length

I was wondering what was the difference in memory and time costs between those two pieces of code :

Sample 1

(1 to 1000000).foreach(_ => processFingerOnly(johnFinger))

and

Sample 2

(1 to 1000000).foreach(_ => processHumanFinger(john))

My question is mainly about knowing if the cost of accessing the human object in depth a million times (sample 2) is more expensive (in time and/or memory) than accessing it in depth once when assigning the val johnFinger then passing only the Finger object to the function (sample 1). But I am interested in every differences that exists between these two pieces of code !

Answer 1

Using the following code, you can do a check to see how long each sample took.

Preparation

case class Finger(length: Int)
case class Arm(thumb: Finger)
case class Body(rightArm: Arm, leftArm: Arm)
case class Human(body: Body)

def processHumanFinger(human: Human) = human.body.rightArm.thumb.length
def processFingerOnly(finger: Finger) = finger.length

val john = Human(Body(Arm(Finger(1)), Arm(Finger(3))))
val johnFinger = john.body.rightArm.thumb.length

def time[R](block: => R): R = {
    val t0 = System.nanoTime()
    val result = block    // call-by-name
    val t1 = System.nanoTime()
    println("Elapsed time: " + (t1 - t0) + "ns")
    result
}

Sample 1

time {(1 to 10000000).foreach(_ => processFingerOnly(johnFinger))}

Sample 2

time {(1 to 10000000).foreach(_ => processHumanFinger(john))}

Through this simple timer, we can see that accessing an object at 1-level is around twice as fast as accessing an object at 4-level.

However, this could be a naiive approach, so I'm more than happy to hear people's opinions and feel free to edit!