faster zipf (#280)

* zipf

* rem props

* config

Author: bitfaster

BitFaster.Caching.HitRateAnalysis/BitFaster.Caching.HitRateAnalysis.csproj

@@ -1,4 +1,4 @@
-<Project Sdk="Microsoft.NET.Sdk">
+<Project Sdk="Microsoft.NET.Sdk">
 
   <PropertyGroup>
     <OutputType>Exe</OutputType>
@@ -13,10 +13,14 @@
     <NoWarn>1701;1702,CS8002</NoWarn>
   </PropertyGroup>
 
+  <ItemGroup>
+    <Compile Include="..\BitFaster.Caching.ThroughputAnalysis\FastZipf.cs" Link="FastZipf.cs" />
+  </ItemGroup>
+
   <ItemGroup>
     <PackageReference Include="ConsoleTables" Version="2.4.2" />
     <PackageReference Include="CsvHelper" Version="28.0.1" />
-    <PackageReference Include="EasyConsole" Version="1.1.0" >
+    <PackageReference Include="EasyConsole" Version="1.1.0">
         <NoWarn>NU1701</NoWarn>
     </PackageReference>
     <PackageReference Include="MathNet.Numerics" Version="5.0.0" />

BitFaster.Caching.HitRateAnalysis/Zipfian/Runner.cs

@@ -5,6 +5,7 @@
 using System.Text;
 using System.Threading.Tasks;
 using BitFaster.Caching.Lru;
+using BitFaster.Caching.ThroughputAnalysis;
 using MathNet.Numerics;
 using MathNet.Numerics.Distributions;
 
@@ -58,8 +59,7 @@ public static void Run()
             {
                 Console.WriteLine($"Generating Zipfian distribution with {sampleCount} samples, s = {sValues[index]}, N = {n}");
                 var sw = Stopwatch.StartNew();
-                zipdfDistribution[index] = new int[sampleCount];
-                Zipf.Samples(zipdfDistribution[index], sValues[index], n);
+                zipdfDistribution[index] = FastZipf.Generate(new Random(666), sampleCount, sValues[index], n);
                 Console.WriteLine($"Took {sw.Elapsed} for s = {sValues[index]}.");
             });
 

BitFaster.Caching.ThroughputAnalysis/FastZipf.cs

@@ -0,0 +1,54 @@
+using System;
+using System.Linq;
+using System.Threading.Tasks;
+using MathNet.Numerics;
+
+namespace BitFaster.Caching.ThroughputAnalysis
+{
+    // produces the same output as MathNet.Numerics Zipf.Samples(random, samples[], s, n)
+    // but about 20x faster.
+    public class FastZipf
+    {
+        static Random srandom = new Random(666);
+
+        public static int[] Generate(Random random, int sampleCount, double s, int n)
+        {
+            double[] num = new double[sampleCount];
+            int[] samples = new int[sampleCount];
+
+            for (int i = 0; i < sampleCount; i++)
+            {
+                while (num[i] == 0.0)
+                {
+                    num[i] = random.NextDouble();
+                }
+            }
+
+            double num2 = 1.0 / SpecialFunctions.GeneralHarmonic(n, s);
+
+            Parallel.ForEach(Enumerable.Range(0, samples.Length), (x, j) =>
+            {
+                double num3 = 0.0;
+                int i;
+
+                for (i = 1; i <= n; i++)
+                {
+                    num3 += num2 / Math.Pow(i, s);
+                    if (num3 >= num[x])
+                    {
+                        break;
+                    }
+                }
+
+                samples[x] = i;
+            });
+
+            return samples;
+        }
+
+        public static int[] Generate(int sampleCount, double s, int n)
+        {
+            return Generate(srandom, sampleCount, s, n);
+        }
+    }
+}

BitFaster.Caching.ThroughputAnalysis/Program.cs

@@ -1,5 +1,9 @@
 using System;
+using System.ComponentModel.DataAnnotations;
+using System.Diagnostics;
 using BitFaster.Caching.ThroughputAnalysis;
+using Iced.Intel;
+using MathNet.Numerics.Distributions;
 
 Host.PrintInfo();
 

BitFaster.Caching.ThroughputAnalysis/ThroughputBenchConfig.cs

@@ -25,10 +25,7 @@ public ZipfConfig(int iterations, int sampleCount, double s, int n)
         {
             this.iterations = iterations;
 
-            Random random = new Random(666);
-
-            samples = new int[sampleCount];
-            Zipf.Samples(random, samples, s, n);
+            samples = FastZipf.Generate(sampleCount, s, n);
         }
 
         public int Iterations => iterations;