Eliminate use of GetArg1() in the select_val instruction

Instead of having one i_select_val_sfI instruction that uses
the GetArg1() macro to fetch the controlling expression, use
three separate instructions for each of the register types.
That will save one word when selecting on the {x,0} register.
It should also be slightly faster since a conditional branch
is eliminated.

Although it seems that the BEAM compiler will never generate
a constant controlling expression (even with optimizations
turned off), we still make sure that they will work by
evaluating the select_val instruction at load time.

Handle the select_tuple_arity instruction in the same way.

1 files changed, 22 insertions, 4 deletions

diff --git a/erts/emulator/beam/beam_emu.c b/erts/emulator/beam/beam_emu.c
index 09f6fc8cef..6cd7fa78e0 100644
--- a/erts/emulator/beam/beam_emu.c
+++ b/erts/emulator/beam/beam_emu.c
@@ -1865,9 +1865,19 @@ void process_main(void)
  {
      Eterm select_val;
 
- OpCase(i_select_tuple_arity_sfI):
-     GetArg1(0, select_val);
+ OpCase(i_select_tuple_arity_xfI):
+     select_val = xb(Arg(0));
+     goto do_select_tuple_arity;
 
+ OpCase(i_select_tuple_arity_yfI):
+     select_val = yb(Arg(0));
+     goto do_select_tuple_arity;
+
+ OpCase(i_select_tuple_arity_rfI):
+     select_val = r(0);
+     I--;
+
+ do_select_tuple_arity:
      if (is_tuple(select_val)) {
 	 select_val = *tuple_val(select_val);
 	 goto do_binary_search;
@@ -1875,9 +1885,17 @@ void process_main(void)
      SET_I((BeamInstr *) Arg(1));
      Goto(*I);
 
+ OpCase(i_select_val_xfI):
+     select_val = xb(Arg(0));
+     goto do_binary_search;
 
- OpCase(i_select_val_sfI):
-     GetArg1(0, select_val);
+ OpCase(i_select_val_yfI):
+     select_val = yb(Arg(0));
+     goto do_binary_search;
+     
+ OpCase(i_select_val_rfI):
+     select_val = r(0);
+     I--;
 
  do_binary_search:
  {