Rearrange unpacking clauses to optimize towards typical usage scenarios

This should give around 20% speedup.
lexmag · Jan 3, 2021 · a4bafb7 · a4bafb7
1 parent 89ee44f
commit a4bafb7
Show file tree

Hide file tree

Showing 2 changed files with 63 additions and 56 deletions.
diff --git a/lib/msgpax/unpacker.ex b/lib/msgpax/unpacker.ex
@@ -47,34 +47,83 @@ defmodule Msgpax.Unpacker do
     unpack(buffer, [], Map.new(options), [], 0, 1)
   end
 
-  primitives = %{
+  defunpack = fn clauses ->
+    for {formats, value} <- clauses, format <- formats do
+      defp unpack(<<unquote(format), rest::bits>>, result, options, outer, index, count) do
+        unpack_continue(rest, [unquote(value) | result], options, outer, index, count)
+      end
+    end
+  end
+
+  defunpack.(%{
     [quote(do: <<0xC0>>)] => nil,
     [quote(do: <<0xC2>>)] => false,
     [quote(do: <<0xC3>>)] => true,
-    # Strings
     [
+      # Strings
       quote(do: <<0b101::3, length::5, value::size(length)-bytes>>),
       quote(do: <<0xD9, length::8, value::size(length)-bytes>>),
-      quote(do: <<0xDA, length::16, value::size(length)-bytes>>),
-      quote(do: <<0xDB, length::32, value::size(length)-bytes>>),
-
       # Floats
       quote(do: <<0xCA, value::32-float>>),
       quote(do: <<0xCB, value::64-float>>),
-
       # Integers
       quote(do: <<0::1, value::7>>),
       quote(do: <<0xCC, value::8>>),
       quote(do: <<0xCD, value::16>>),
+      quote(do: <<0xD0, value::8-signed>>),
+      quote(do: <<0xD1, value::16-signed>>)
+    ] => quote(do: value),
+    # Negative fixint
+    [quote(do: <<0b111::3, value::5>>)] => quote(do: value - 0b100000)
+  })
+
+  maps = [
+    quote(do: <<0b1000::4, length::4>>),
+    quote(do: <<0xDE, length::16>>),
+    quote(do: <<0xDF, length::32>>)
+  ]
+
+  for format <- maps do
+    defp unpack(<<unquote(format), rest::bits>>, result, options, outer, index, count) do
+      case var!(length, __MODULE__) do
+        0 ->
+          unpack_continue(rest, [%{} | result], options, outer, index, count)
+
+        length ->
+          unpack(rest, result, options, [:map, index, count | outer], 0, length * 2)
+      end
+    end
+  end
+
+  lists = [
+    quote(do: <<0b1001::4, length::4>>),
+    quote(do: <<0xDC, length::16>>),
+    quote(do: <<0xDD, length::32>>)
+  ]
+
+  for format <- lists do
+    defp unpack(<<unquote(format), rest::bits>>, result, options, outer, index, count) do
+      case var!(length, __MODULE__) do
+        0 ->
+          unpack_continue(rest, [[] | result], options, outer, index, count)
+
+        length ->
+          unpack(rest, result, options, [:list, index, count | outer], 0, length)
+      end
+    end
+  end
+
+  defunpack.(%{
+    [
+      # Strings
+      quote(do: <<0xDA, length::16, value::size(length)-bytes>>),
+      quote(do: <<0xDB, length::32, value::size(length)-bytes>>),
+      # Integers
       quote(do: <<0xCE, value::32>>),
       quote(do: <<0xCF, value::64>>),
-      quote(do: <<0xD0, value::8-signed>>),
-      quote(do: <<0xD1, value::16-signed>>),
       quote(do: <<0xD2, value::32-signed>>),
       quote(do: <<0xD3, value::64-signed>>)
     ] => quote(do: value),
-    # Negative fixint
-    [quote(do: <<0b111::3, value::5>>)] => quote(do: value - 0b100000),
     # Extensions
     [
       quote(do: <<0xD4, type, content::1-bytes>>),
@@ -92,6 +141,7 @@ defmodule Msgpax.Unpacker do
       quote(do: <<0xC5, length::16, content::size(length)-bytes>>),
       quote(do: <<0xC6, length::32, content::size(length)-bytes>>)
     ] => quote(do: unpack_binary(content, var!(options))),
+    # NaN and ±infinity
     [
       quote(do: <<0xCA, 0::1, 0xFF, 0::23>>),
       quote(do: <<0xCB, 0::1, 0xFF, 0b111::3, 0::52>>)
@@ -104,50 +154,7 @@ defmodule Msgpax.Unpacker do
       quote(do: <<0xCA, _sign::1, 0xFF, _fraction::23>>),
       quote(do: <<0xCB, _sign::1, 0xFF, 0b111::3, _fraction::52>>)
     ] => quote(do: unpack_float(NaN, var!(options)))
-  }
-
-  for {formats, value} <- primitives,
-      format <- formats do
-    defp unpack(<<unquote(format), rest::bits>>, result, options, outer, index, count) do
-      unpack_continue(rest, [unquote(value) | result], options, outer, index, count)
-    end
-  end
-
-  lists = [
-    quote(do: <<0b1001::4, length::4>>),
-    quote(do: <<0xDC, length::16>>),
-    quote(do: <<0xDD, length::32>>)
-  ]
-
-  for format <- lists do
-    defp unpack(<<unquote(format), rest::bits>>, result, options, outer, index, count) do
-      case var!(length, __MODULE__) do
-        0 ->
-          unpack_continue(rest, [[] | result], options, outer, index, count)
-
-        length ->
-          unpack(rest, result, options, [:list, index, count | outer], 0, length)
-      end
-    end
-  end
-
-  maps = [
-    quote(do: <<0b1000::4, length::4>>),
-    quote(do: <<0xDE, length::16>>),
-    quote(do: <<0xDF, length::32>>)
-  ]
-
-  for format <- maps do
-    defp unpack(<<unquote(format), rest::bits>>, result, options, outer, index, count) do
-      case var!(length, __MODULE__) do
-        0 ->
-          unpack_continue(rest, [%{} | result], options, outer, index, count)
-
-        length ->
-          unpack(rest, result, options, [:map, index, count | outer], 0, length * 2)
-      end
-    end
-  end
+  })
 
   defp unpack(<<byte, _::bits>>, _result, _options, _outer, _index, _count) do
     throw({:invalid_format, byte})

diff --git a/test/msgpax_test.exs b/test/msgpax_test.exs
@@ -108,11 +108,11 @@ defmodule MsgpaxTest do
     assert_format 42.1, <<203>>
 
     for {packed, value} <- %{
-          # 32-bit.
+          # 32-bit
           <<202, 0x7FC00000::32>> => Msgpax.NaN,
           <<202, 0x7F800000::32>> => Msgpax.Infinity,
           <<202, 0xFF800000::32>> => Msgpax.NegInfinity,
-          # 64-bit.
+          # 64-bit
           <<203, 0x7FF8::16, 0::48>> => Msgpax.NaN,
           <<203, 0x7FF0::16, 0::48>> => Msgpax.Infinity,
           <<203, 0xFFF0::16, 0::48>> => Msgpax.NegInfinity