Measured parallel processing in Java

I don't know much about Java because it can process in parallel, so I ran Parallel Stream and Completable Future, which I think I often use.

Measurement method

Processing to fetch data from DB on AWS individually with findById from 30 IDs.

Implementation environment

local

MacBook Pro (13-inch, 2018, Four Thunderbolt 3 Ports)
2.3 GHz Intel Core i5
16 GB 2133 MHz LPDDR3
Java1.8
Springboot2

AWS RDS

db.t2.micro --Tokyo Region

Implementation

series

`Serial`


        List<Long> list = createIdList(); //30 cases
        List<Currency> currencies = new ArrayList<>();
        list.forEach(id -> currencyService.findById(id)
                .ifPresent(currency -> currencies.add(currency)));

`result`


3644ms

ParallelStream

`ParallelStream`


        List<Long> list = createIdList(); //30 cases
        List<Currency> currencies = new ArrayList<>();
        list.parallelStream()
                .forEach(id -> currencyService.findById(id)
                                .ifPresent(currency -> currencies.add(currency)));

`result`


607ms

CompletableFuture

`CompletableFuture`


        List<Long> list = createIdList(); //30 cases
        List<Currency> currencies = new ArrayList<>();
        ExecutorService ex = Executors.newWorkStealingPool();
        List<CompletableFuture<Void>> futures = new ArrayList<>();
        for (long id : list) {
            futures.add(CompletableFuture.runAsync(() -> currencyService
                    .findById(id).ifPresent(c -> currencies.add(c)), ex));
        }
        CompletableFuture.allOf((futures.toArray(new CompletableFuture[futures.size()]))).join();

`result`


604ms

By the way, when I tried to make it single thread with ʻExecutors.newSingleThreadExecutor ();`, it was 4123ms.

Postscript

In this experiment, we are going to get data to RDS, but it seems that we need to be careful when connecting externally with Parallel Stream. http://taichiw.hatenablog.com/entry/2018/09/30/224907

Also, I added it to the List in the loop, but it seems that it is not good because it is a so-called side effect.