Loops performance in Groovy

IntroductionIn the 2018 Advent of Code challenged I solved all the puzzles in Groovy. It is pretty obvious, that choosing good data structure is the most important to obtain performant solution. However, the way we iterate over those structures is also…

Introduction

In the 2018 Advent of Code challenged I solved all the puzzles in Groovy. It is pretty obvious, that choosing good data structure is the most important to obtain performant solution. However, the way we iterate over those structures is also very significant, at least when using Groovy.

Measuring performance

I want to measure how long it takes to sum some numbers. For testing performance of loops I prepared a small function that simply sums some numbers:

void printAddingTime(String message, long to, Closure<Long> adder) {
    LocalTime start = LocalTime.now()
    long sum = adder(to)
    println("$message: $sum calculated in ${Duration.between(start, LocalTime.now()).toMillis()} ms")
}

Pseudo code for summing functions is below:

for i = 1 to n
  for j = 1 to n
    sum += i * j
  end
end

Loops types

Let’s implement the summing function in various ways.

collect and sum

First loop type is to use built-in (by Groovy) function collect and sum on collections (Range it this example):

(1..n).collect { long i ->
  (1..n).collect { long j ->
    i * j
  }.sum()
}.sum()

each

Next, let’s write the same function using each built-in function on collections (Range it this example) and then add results to accumulator variable:

long sum = 0
(1..n).each { long i ->
    (1..n).each { long j ->
        sum += i * j
    }
}
return sum

times

Now instead of using each we could use the function times built-in on Number by Groovy:

long sum = 0
n.times { long i ->
  n.times { long j ->
    sum += (i + 1)*(j+1)
  }
}
return sum

We have to add 1 to i and j because times generates numbers from 0 to n exclusive.

LongStream with sum

Java 8 came with a new feature – streams. One example of streams is LongStream. Fortunately, it has sum built-in function, which we can use:

LongStream.range(0, n).map { i ->
    LongStream.range(0, n).map { j ->
        (i + 1) * (j + 1)
    }.sum()
}.sum()

LongStream generates numbers in the same way as times function, so we also have to add 1 to i and j here.

LongStream with manual sum

Instead of sum function on LongStream, we can add all numbers manually:

long sum = 0
LongStream.range(0, n).forEach { i ->
    LongStream.range(0, n).forEach { j ->
        sum += (i + 1) * (j + 1)
    }
}
return sum

while

Of course since Groovy inherits from Java a big part of its syntax, we can use the while loop:

long sum = 0
long i = 1
while(i <= n){
    long j = 1
    while(j <= n){
        sum+= i*j
        ++j
    }
    ++i
}
return sum

for

As we can use while, we can also use for loop in Groovy:

long sum = 0
for (long i = 1; i <= n; ++i) {
    for (long j = 1; j <= n; ++j) {
        sum += i * j
    }
}
return sum

 

Results

My tests I run on Java 1.8 and Groovy 2.5.5. Script loops.groovy was fired using bash script:

#!/bin/sh
for x in 10 100 1000 10000 100000; do
  echo $x
  groovy loops.groovy $x
  echo
done

Values are in milliseconds

Loop  n 10 100 1000 10000 100000
collect + sum 7 22 216 16244 1546822
each 12 17 118 7332 706781
times 2 10 109 8264 708684
LongStream + sum 7 17 127 7679 763341
LongStream + manual sum 18 35 149 6857 680804
while 8 20 103 3166 301967
for 7 10 25 359 27966

As you can spot, for small amount of iterations using built-in Groovy functions is good enough, but for much bigger amount of iterations we should use while or for loops like in plain, old Java.

Show me the code

Code for those examples are available here. You can run those examples on your machine and check performance on your own.

You May Also Like

Recently at storm-users

I've been reading through storm-users Google Group recently. This resolution was heavily inspired by Adam Kawa's post "Football zero, Apache Pig hero". Since I've encountered a lot of insightful and very interesting information I've decided to describe some of those in this post.

  • nimbus will work in HA mode - There's a pull request open for it already... but some recent work (distributing topology files via Bittorrent) will greatly simplify the implementation. Once the Bittorrent work is done we'll look at reworking the HA pull request. (storm’s pull request)

  • pig on storm - Pig on Trident would be a cool and welcome project. Join and groupBy have very clear semantics there, as those concepts exist directly in Trident. The extensions needed to Pig are the concept of incremental, persistent state across batches (mirroring those concepts in Trident). You can read a complete proposal.

  • implementing topologies in pure python with petrel looks like this:

class Bolt(storm.BasicBolt):
    def initialize(self, conf, context):
       ''' This method executed only once '''
        storm.log('initializing bolt')

    def process(self, tup):
       ''' This method executed every time a new tuple arrived '''       
       msg = tup.values[0]
       storm.log('Got tuple %s' %msg)

if __name__ == "__main__":
    Bolt().run()
  • Fliptop is happy with storm - see their presentation here

  • topology metrics in 0.9.0: The new metrics feature allows you to collect arbitrarily custom metrics over fixed windows. Those metrics are exported to a metrics stream that you can consume by implementing IMetricsConsumer and configure with Config.java#L473. Use TopologyContext#registerMetric to register new metrics.

  • storm vs flume - some users' point of view: I use Storm and Flume and find that they are better at different things - it really depends on your use case as to which one is better suited. First and foremost, they were originally designed to do different things: Flume is a reliable service for collecting, aggregating, and moving large amounts of data from source to destination (e.g. log data from many web servers to HDFS). Storm is more for real-time computation (e.g. streaming analytics) where you analyse data in flight and don't necessarily land it anywhere. Having said that, Storm is also fault-tolerant and can write to external data stores (e.g. HBase) and you can do real-time computation in Flume (using interceptors)

That's all for this day - however, I'll keep on reading through storm-users, so watch this space for more info on storm development.

I've been reading through storm-users Google Group recently. This resolution was heavily inspired by Adam Kawa's post "Football zero, Apache Pig hero". Since I've encountered a lot of insightful and very interesting information I've decided to describe some of those in this post.

  • nimbus will work in HA mode - There's a pull request open for it already... but some recent work (distributing topology files via Bittorrent) will greatly simplify the implementation. Once the Bittorrent work is done we'll look at reworking the HA pull request. (storm’s pull request)

  • pig on storm - Pig on Trident would be a cool and welcome project. Join and groupBy have very clear semantics there, as those concepts exist directly in Trident. The extensions needed to Pig are the concept of incremental, persistent state across batches (mirroring those concepts in Trident). You can read a complete proposal.

  • implementing topologies in pure python with petrel looks like this:

class Bolt(storm.BasicBolt):
    def initialize(self, conf, context):
       ''' This method executed only once '''
        storm.log('initializing bolt')

    def process(self, tup):
       ''' This method executed every time a new tuple arrived '''       
       msg = tup.values[0]
       storm.log('Got tuple %s' %msg)

if __name__ == "__main__":
    Bolt().run()
  • Fliptop is happy with storm - see their presentation here

  • topology metrics in 0.9.0: The new metrics feature allows you to collect arbitrarily custom metrics over fixed windows. Those metrics are exported to a metrics stream that you can consume by implementing IMetricsConsumer and configure with Config.java#L473. Use TopologyContext#registerMetric to register new metrics.

  • storm vs flume - some users' point of view: I use Storm and Flume and find that they are better at different things - it really depends on your use case as to which one is better suited. First and foremost, they were originally designed to do different things: Flume is a reliable service for collecting, aggregating, and moving large amounts of data from source to destination (e.g. log data from many web servers to HDFS). Storm is more for real-time computation (e.g. streaming analytics) where you analyse data in flight and don't necessarily land it anywhere. Having said that, Storm is also fault-tolerant and can write to external data stores (e.g. HBase) and you can do real-time computation in Flume (using interceptors)

That's all for this day - however, I'll keep on reading through storm-users, so watch this space for more info on storm development.