Align benchmark scaling and minimum samples

# Motivation

We want to have an aligned strategy across all of our benchmarks in how we use the scaling factor and samples.

# Modification

This PR makes sure that any inner loop is scaled by the scaling factor. Moreover, we want to run more than 1 iteration. To do this we set an almost infinite maximum duration and a maximum iteration count.

# Result

More consistent benchmarks

Author: FranzBusch

Benchmarks/Benchmarks/NIOCoreBenchmarks/Benchmarks.swift

@@ -23,7 +23,12 @@ let benchmarks = {
 
     Benchmark(
         "NIOAsyncChannel.init",
-        configuration: Benchmark.Configuration(metrics: defaultMetrics)
+        configuration: .init(
+            metrics: defaultMetrics,
+            scalingFactor: .kilo,
+            maxDuration: .seconds(10_000_000),
+            maxIterations: 10
+        )
     ) { benchmark in
         // Elide the cost of the 'EmbeddedChannel'. It's only used for its pipeline.
         var channels: [EmbeddedChannel] = []

Benchmarks/Benchmarks/NIOPosixBenchmarks/Benchmarks.swift

@@ -26,8 +26,9 @@ let benchmarks = {
         "TCPEcho",
         configuration: .init(
             metrics: defaultMetrics,
-            timeUnits: .milliseconds,
-            scalingFactor: .mega
+            scalingFactor: .mega,
+            maxDuration: .seconds(10_000_000),
+            maxIterations: 5
         )
     ) { benchmark in
         try runTCPEcho(
@@ -43,8 +44,9 @@ let benchmarks = {
         "TCPEchoAsyncChannel",
         configuration: .init(
             metrics: defaultMetrics,
-            timeUnits: .milliseconds,
             scalingFactor: .mega,
+            maxDuration: .seconds(10_000_000),
+            maxIterations: 5,
             // We are expecting a bit of allocation variance due to an allocation
             // in the Concurrency runtime which happens when resuming a continuation.
             thresholds: [.mallocCountTotal: .init(absolute: [.p90: 2000])],

Benchmarks/Thresholds/5.10/NIOCoreBenchmarks.NIOAsyncChannel.init.p90.json

@@ -1,3 +1,3 @@
 {
-  "mallocCountTotal" : 8
-}
+  "mallocCountTotal" : 8000
+}

Benchmarks/Thresholds/5.10/NIOPosixBenchmarks.TCPEcho.p90.json

@@ -1,3 +1,3 @@
 {
-  "mallocCountTotal" : 108
+  "mallocCountTotal" : 555
 }

Benchmarks/Thresholds/5.10/NIOPosixBenchmarks.TCPEchoAsyncChannel.p90.json

@@ -1,3 +1,3 @@
 {
-  "mallocCountTotal" : 165000
-}
+  "mallocCountTotal" : 164375
+}

Benchmarks/Thresholds/5.8/NIOCoreBenchmarks.NIOAsyncChannel.init.p90.json

@@ -1,3 +1,3 @@
 {
-  "mallocCountTotal" : 8
-}
+  "mallocCountTotal" : 8000
+}

Benchmarks/Thresholds/5.8/NIOPosixBenchmarks.TCPEcho.p90.json

@@ -1,3 +1,3 @@
 {
-  "mallocCountTotal" : 110
+  "mallocCountTotal" : 556
 }

Benchmarks/Thresholds/5.9/NIOCoreBenchmarks.NIOAsyncChannel.init.p90.json

@@ -1,3 +1,3 @@
 {
-  "mallocCountTotal" : 8
-}
+  "mallocCountTotal" : 8000
+}

Benchmarks/Thresholds/5.9/NIOPosixBenchmarks.TCPEcho.p90.json

@@ -1,3 +1,3 @@
 {
-  "mallocCountTotal" : 110
+  "mallocCountTotal" : 556
 }

Benchmarks/Thresholds/5.9/NIOPosixBenchmarks.TCPEchoAsyncChannel.p90.json

@@ -1,3 +1,3 @@
 {
-  "mallocCountTotal" : 164000
-}
+  "mallocCountTotal" : 164381
+}