为什么HashSet<string>(StringComparer.InvariantCultureIgnoreCase) (Perf_HashSet_CaseInsensitive)表现这么差?Perf_HashSet的解决方案相对于20倍左右。
// perf Contains(): 1M iterations, 25 size (unsuccessful lookup)
Test Duration (ms)
Perf_HashSet 43 <-
Perf_Dictionary 49
Perf_HybridDictionary 63
Perf_ListDictionary 223
Perf_List 225
Perf_HashSet_CaseInsensitive 903 <-代码:
// <TargetFramework>net5.0</TargetFramework>
[TestFixture]
public class ContainsPerfTests
{
private const int iterations = 1_000_000;
private const int size = 25;
[Test]
[Explicit]
public void Perf_List()
{
var list = new List<string>();
for (int i = 0; i < size; i++)
{
list.Add(Guid.NewGuid().ToString().ToLowerInvariant());
}
var x = Guid.NewGuid().ToString();
var sw = Stopwatch.StartNew();
for (int i = 0; i < iterations; i++)
{
var contains = list.Contains(x.ToLowerInvariant());
}
sw.Stop();
Console.WriteLine(sw.ElapsedMilliseconds);
}
[Test]
[Explicit]
public void Perf_HashSet()
{
var hashSet = new HashSet<string>();
for (int i = 0; i < size; i++)
{
hashSet.Add(Guid.NewGuid().ToString().ToLowerInvariant());
}
var x = Guid.NewGuid().ToString();
var sw = Stopwatch.StartNew();
for (int i = 0; i < iterations; i++)
{
var contains = hashSet.Contains(x.ToLowerInvariant());
}
sw.Stop();
Console.WriteLine(sw.ElapsedMilliseconds);
}
[Test]
[Explicit]
public void Perf_HashSet_CaseInsensitive()
{
var hashSetCaseInsensitive = new HashSet<string>(StringComparer.InvariantCultureIgnoreCase);
for (int i = 0; i < size; i++)
{
hashSetCaseInsensitive.Add(Guid.NewGuid().ToString().ToLowerInvariant());
}
var x = Guid.NewGuid().ToString();
var sw = Stopwatch.StartNew();
for (int i = 0; i < iterations; i++)
{
var contains = hashSetCaseInsensitive.Contains(x);
}
sw.Stop();
Console.WriteLine(sw.ElapsedMilliseconds);
}
[Test]
[Explicit]
public void Perf_Dictionary()
{
var dictionary = new Dictionary<string, bool>();
for (int i = 0; i < size; i++)
{
dictionary.Add(Guid.NewGuid().ToString().ToLowerInvariant(), false);
}
var x = Guid.NewGuid().ToString();
var sw = Stopwatch.StartNew();
for (int i = 0; i < iterations; i++)
{
var contains = dictionary.ContainsKey(x.ToLowerInvariant());
}
sw.Stop();
Console.WriteLine(sw.ElapsedMilliseconds);
}
[Test]
[Explicit]
public void Perf_HybridDictionary()
{
var hybridDictionary = new HybridDictionary(caseInsensitive: true);
for (int i = 0; i < size; i++)
{
hybridDictionary.Add(Guid.NewGuid().ToString().ToLowerInvariant(), null);
}
var x = Guid.NewGuid().ToString();
var sw = Stopwatch.StartNew();
for (int i = 0; i < iterations; i++)
{
var contains = hybridDictionary.Contains(x);
}
sw.Stop();
Console.WriteLine(sw.ElapsedMilliseconds);
}
[Test]
[Explicit]
public void Perf_ListDictionary()
{
var listDictionary = new ListDictionary();
for (int i = 0; i < size; i++)
{
listDictionary.Add(Guid.NewGuid().ToString().ToLowerInvariant(), null);
}
var x = Guid.NewGuid().ToString();
var sw = Stopwatch.StartNew();
for (int i = 0; i < iterations; i++)
{
var contains = listDictionary.Contains(x.ToLowerInvariant());
}
sw.Stop();
Console.WriteLine(sw.ElapsedMilliseconds);
}
}发布于 2021-10-01 19:56:49
性能差异很大的主要原因是,对于不区分大小写的哈希集使用区分区域性的比较,而区分大小写的哈希集使用序号比较。
在不传递哈希集任何相等比较器的情况下,哈希集使用该类型的默认相等比较器,它调用Equals方法。注意String.Equals
执行序号比较(区分大小写和区域性不敏感)。
而StringComparer.InvariantCultureIgnoreCase
使用不变区域性的单词比较规则执行不区分大小写的字符串比较。
必须指出,“不变文化”与“文化不敏感”或“序号”并不相同:
( Console.WriteLine(StringComparer.InvariantCultureIgnoreCase.Equals( "ss",“true”);//真Console.WriteLine(StringComparer.OrdinalIgnoreCase.Equals( "ss",“true”));// false
不变文化仍然有排序规则,查找和应用这些规则需要时间。
将InvariantCultureIgnoreCase更改为OrdinalIgnoreCase将使执行时间几乎相同。这并不意味着你的基准没有其他问题,正如评论所指出的那样。
https://stackoverflow.com/questions/69410717
复制相似问题