I’m reviewing my Leetcode Java submissions and thought, I’d see what the code would look like with t...

# getting-started

[sorianog]

I’m reviewing my Leetcode Java submissions and thought, I’d see what the code would look like with the Automagic to Kotlin conversion from Studio. One problem that I solved with updated Kotlin code seems to be less performant (by Leetcode benchmarking) 🤔
Maybe I’m not hitting their servers at the right time 😛
https://leetcode.com/problems/move-zeroes/discuss/1664500/Java-to-Kotlin-Converted-Solution%3A-Less-Performant-with-Kotlin

[Joffrey]

Not sure how leetcode performs benchmarks but this would need be repeated a number of times in order to be significant.
Also, does this count the compilation time as well?

[Joffrey]

Also, https://leetcode.com/discuss/general-discussion/644868/solution-benchmarking-seems-to-be-very-noisyinaccurate

[Joffrey]

That said, I don't see anything obvious with the converted solution, apart maybe from the use of .indices which might not benefit from the ranges optimizations (although I doubt it's not optimized).

Did you try this:

for (i in 0 until nums.length)

?

[ephemient]

is their Kotlin runner compiling to JVM or using the scripting engine? the timing makes it look like the latter

[ephemient]

(but I haven't messed around with leetcode to see if that holds up elsewhere)

[sorianog]

Appreciate the info and responses you 2.
Joffrey, I tried what you suggested and got the following:

• 01/03/2022 10:03
• Accepted
• 583 ms
• 62.7 MB
If anything, I can do my own benchmarking locally to verify some things.