@mg6maciej Here's what I did, in Kotlin 1.1:

fun <T> List<T>.batched(batchSize: Int): List<List<T>> {
    check(batchSize > 0) { "Invalid value for parameter 'batchSize': $batchSize (should be greater than 0)" }
    val batchAmount = (size + batchSize - 1) / batchSize
    val result = List<MutableList<T>>(batchAmount) { ArrayList(batchSize) }
    var index = 0
    forEach { result[index++ / batchSize].add(it) }
    return result
}

You seem to care about optimization, so note than when called on a

LinkedList

, your implementation does a O(N)

get

call on line 13

Yes. Your and @gaetan implementations are superior to mine. If I cared about optimizations in my case, I'd choose one of them.

batch

could be added to #stdlib too.

Based on my own tests I can say that @gaetan's solution is orders of magnitude better

Yes, both in terms of speed and memory use. sublist is backed by original list.

As a personal rule, I tend to use existing functions that are generally highly optimized.

sublist

in that case.