Are interned strings excepted from garbage collect

问题:

I am trying to reduce time it takes to do a Gen2 collection. My app creates and holds a large number of string objects, which persist through its life.

Reducing number of scanned objects should reduce GC time. I was wondering whether intern pool is excepted from garbage collection. There isn't anything to collect there anyway. If so, I could intern all these strings and speed up GC.

回答1:

If you are trying to reduce the total amount of memory your application allocates, keep in mind that interning a string has two unwanted side effects. First, the memory allocated for interned String objects is not likely be released until the common language runtime (CLR) terminates. The reason is that the CLR's reference to the interned String object can persist after your application, or even your application domain, terminates. Second, to intern a string, you must first create the string. The memory used by the String object must still be allocated, even though the memory will eventually be garbage collected.

Ref: https://msdn.microsoft.com/en-us/library/system.string.intern(v=vs.110).aspx

Emphasis mine.

回答2:

I made a quick test and interning of strings does not seem to save them from scanning by GC. At least not in .NET 4.5 64 bit.

class Program
{
    static void Main(string[] args)
    {
        for (int i = 0; i < 20000000; ++i)
        {
            string s = i.ToString("X");
            string.Intern(s);
        }
        GC.Collect(3, GCCollectionMode.Forced, true);
        long t1 = Stopwatch.GetTimestamp();
        GC.Collect(3, GCCollectionMode.Forced, true);
        long t2 = Stopwatch.GetTimestamp();
        Console.WriteLine((double)(t2 - t1) / Stopwatch.Frequency);
    }
}

This benchmark returns 0.23s on an i5 3570k. If the strings are put into an array instead of interning, it returns 0.26s. If the strings are interned and created via (i % 10).ToString(), i.e. there's a small number of different instances, the benchmark returns microseconds.

So sadly this is not a way to bypass garbage collection. I think C# should have some way of marking strings as persistent and stop the runtime wasting time on scanning them.