C#•2mo ago

✅ I don't undertand the error: capacity was less than the current size

Hey I have two versions of code and second one doesn't work first version:

int[] array = Enumerable.Range(0, 1000000000).ToArray();
var lst = new List<double>();

var stopWatch = new Stopwatch();
stopWatch.Start();

for (int i = 0; i < array.Length; i++)
{
    var result = Math.Pow(i, 100000000000);
    lst.Add(result);
}

stopWatch.Stop();
Console.WriteLine(stopWatch.Elapsed);

int[] array = Enumerable.Range(0, 1000000000).ToArray();
var lst = new List<double>();

var stopWatch = new Stopwatch();
stopWatch.Start();

for (int i = 0; i < array.Length; i++)
{
    var result = Math.Pow(i, 100000000000);
    lst.Add(result);
}

stopWatch.Stop();
Console.WriteLine(stopWatch.Elapsed);

second version:

int[] array = Enumerable.Range(0, 1000000000).ToArray();
var lst = new List<double>();

var stopWatch = new Stopwatch();
stopWatch.Start();

Parallel.For(0, array.Length, i =>
{
    var result = Math.Pow(i, 100000000000);
    lst.Add(result);
});

stopWatch.Stop();
Console.WriteLine(stopWatch.Elapsed);

int[] array = Enumerable.Range(0, 1000000000).ToArray();
var lst = new List<double>();

var stopWatch = new Stopwatch();
stopWatch.Start();

Parallel.For(0, array.Length, i =>
{
    var result = Math.Pow(i, 100000000000);
    lst.Add(result);
});

stopWatch.Stop();
Console.WriteLine(stopWatch.Elapsed);

second version always throws the exception:

Unhandled exception. System.AggregateException: One or more errors occurred. (capacity was less than the current size. (Parameter 'value')) (capacity was less than the current size. (Parameter 'value')) (capacity was less than the current size. (Parameter 'value')) (capacity was less than the current size.

11 Replies

❤RieBi&❤•2mo ago

You can't just use regular list in multiple threads and appends elements to it on different threads

SWEETPONYOP•2mo ago

simple for takes 25 seconds in this case, that is why I decided to use Parallel.For but maybe I was wrong and it shouldn't be faster

❤RieBi&❤•2mo ago

Your code is not thread safe It's very dangerous Look into that

SWEETPONYOP•2mo ago

hm ur right, I will change List to smth concurrent like ConcurrentBag and try to see results forgot about thread safety simple for was x2 faster than Parallel.For idk why

❤RieBi&❤•2mo ago

With concurrent bag and same code?

SWEETPONYOP•2mo ago

yes

❤RieBi&❤•2mo ago

Nothing unexpected tbf

blueberriesiftheywerecats•2mo ago

I think it would work if you provide needed capacity when creating list

Anton•2mo ago

parallel has an overload with indices iirc, use that and an array

❤RieBi&❤•2mo ago

Yeah so when multiple threads call the Add() at the same time, it would be essentially the same as doing it without parallel, but with parallel you also get overhead of syncing threads. So I decided to do it myself:

using System.Diagnostics;


var max = 100_000_000;
int[] array = Enumerable.Range(0, max).ToArray();
var lst = new List<double>();

var arr = new double[max];

var stopWatch = new Stopwatch();
stopWatch.Start();

for (int i = 0; i < array.Length; i++)
{
    var result = Math.Pow(i, 100000000000);
    arr[i] = result;
}

stopWatch.Stop();
Console.WriteLine(stopWatch.Elapsed);

var arr2 = new double[max];
stopWatch.Restart();

Parallel.For(0, array.Length, i =>
{
    var result = Math.Pow(i, 100000000000);
    arr2[i] = result;
});

stopWatch.Stop();
Console.WriteLine(stopWatch.Elapsed);

using System.Diagnostics;


var max = 100_000_000;
int[] array = Enumerable.Range(0, max).ToArray();
var lst = new List<double>();

var arr = new double[max];

var stopWatch = new Stopwatch();
stopWatch.Start();

for (int i = 0; i < array.Length; i++)
{
    var result = Math.Pow(i, 100000000000);
    arr[i] = result;
}

stopWatch.Stop();
Console.WriteLine(stopWatch.Elapsed);

var arr2 = new double[max];
stopWatch.Restart();

Parallel.For(0, array.Length, i =>
{
    var result = Math.Pow(i, 100000000000);
    arr2[i] = result;
});

stopWatch.Stop();
Console.WriteLine(stopWatch.Elapsed);

The second one runs almost 2x as fast as the former one. But I've got a shitty machine, so it might be even faster on your one. The key here is that every thread operates on their own indices, and yep it's faster

SWEETPONYOP•2mo ago

oh, thanks for this!

Gaming

Programming

✅ I don't undertand the error: capacity was less than the current size