docs: improve code readability showing perf gains

Two improvements here:
1. It's better to show the slowest first.
2. It's better to not 'open Nessos.Streams', and use the
namespace explicitly.
It took me a minute to wrap my head around this code,
so I think with these 2 improvements, other will understand
it a bit better the first time they read it.

Author: knocte

docs/content/index.fsx

@@ -38,19 +38,10 @@ For simple pipelines we have observed performance improvements of a factor of fo
 Important performance tip: Make sure that FSI is running with 64-bit option set to true and fsproj option prefer 32-bit is unchecked.
 *)
 
-open Nessos.Streams
-
 let data = [|1..10000000|] |> Array.map int64
 
 // Sequential
 
-// Real: 00:00:00.044, CPU: 00:00:00.046, GC gen0: 0, gen1: 0, gen2: 0
-data
-|> Stream.ofArray
-|> Stream.filter (fun x -> x % 2L = 0L)
-|> Stream.map (fun x -> x + 1L)
-|> Stream.sum
-
 // Real: 00:00:00.264, CPU: 00:00:00.265, GC gen0: 0, gen1: 0, gen2: 0
 data
 |> Seq.filter (fun x -> x % 2L = 0L)
@@ -63,22 +54,29 @@ data
 |> Array.map (fun x -> x + 1L)
 |> Array.sum
 
+// Real: 00:00:00.044, CPU: 00:00:00.046, GC gen0: 0, gen1: 0, gen2: 0
+data
+|> Nessos.Streams.Stream.ofArray
+|> Nessos.Streams.Stream.filter (fun x -> x % 2L = 0L)
+|> Nessos.Streams.Stream.map (fun x -> x + 1L)
+|> Nessos.Streams.Stream.sum
+
 // Parallel
 open FSharp.Collections.ParallelSeq
 
-// Real: 00:00:00.017, CPU: 00:00:00.078, GC gen0: 0, gen1: 0, gen2: 0
-data
-|> ParStream.ofArray
-|> ParStream.filter (fun x -> x % 2L = 0L)
-|> ParStream.map (fun x -> x + 1L)
-|> ParStream.sum
-
 // Real: 00:00:00.045, CPU: 00:00:00.187, GC gen0: 0, gen1: 0, gen2: 0
 data
 |> PSeq.filter (fun x -> x % 2L = 0L)
 |> PSeq.map (fun x -> x + 1L)
 |> PSeq.sum
 
+// Real: 00:00:00.017, CPU: 00:00:00.078, GC gen0: 0, gen1: 0, gen2: 0
+data
+|> Nessos.Streams.ParStream.ofArray
+|> Nessos.Streams.ParStream.filter (fun x -> x % 2L = 0L)
+|> Nessos.Streams.ParStream.map (fun x -> x + 1L)
+|> Nessos.Streams.ParStream.sum
+
 (**
 ## References