Enable LocalCSE by default (#4089)

Enable it in -O3 and -Os and higher.

This helps very little on output from LLVM, but also it does not alter
compile times much anyhow. On code that has not been run through
an optimizing compiler already, this can help quite a lot, e.g., 15% of
code size on some wasm GC samples.

This will not normally help with speed, as optimizing VMs do such
things anyhow. However, this can help baseline compilers and
interpreters and so forth.

1 files changed, 11 insertions, 11 deletions

diff --git a/test/wasm2js/unary-ops.2asm.js.opt b/test/wasm2js/unary-ops.2asm.js.opt
index 027c9164d..ecf227753 100644
--- a/test/wasm2js/unary-ops.2asm.js.opt
+++ b/test/wasm2js/unary-ops.2asm.js.opt
@@ -46,23 +46,23 @@ function asmFunc(env) {
  
  function legalstub$2($0, $1_1, $2, $3) {
   var $4 = 0, $5 = 0, $6_1 = 0, $7_1 = 0;
-  $7_1 = $2;
+  $5 = $0;
   $4 = $1_1;
   while (1) {
-   $2 = $5;
    $1_1 = $6_1;
-   if ($0 | $4) {
-    $1_1 = $0;
-    $0 = $1_1 - 1 & $1_1;
-    $4 = $4 - ($1_1 >>> 0 < 1) & $4;
-    $5 = $5 + 1 | 0;
-    $6_1 = $5 >>> 0 < 1 ? $6_1 + 1 | 0 : $6_1;
+   $0 = $7_1;
+   if ($5 | $4) {
+    $0 = $5;
+    $5 = $0 - 1 & $0;
+    $4 = $4 - ($0 >>> 0 < 1) & $4;
+    $6_1 = $6_1 + 1 | 0;
+    $7_1 = $6_1 >>> 0 < 1 ? $7_1 + 1 | 0 : $7_1;
     continue;
    }
    break;
   };
-  i64toi32_i32$HIGH_BITS = $1_1;
-  return ($2 | 0) == ($7_1 | 0) & ($3 | 0) == (i64toi32_i32$HIGH_BITS | 0);
+  i64toi32_i32$HIGH_BITS = $0;
+  return ($1_1 | 0) == ($2 | 0) & ($3 | 0) == (i64toi32_i32$HIGH_BITS | 0);
  }
  
  function legalstub$3($0, $1_1, $2) {
@@ -87,7 +87,7 @@ function asmFunc(env) {
  function legalstub$9($0, $1_1, $2, $3) {
   var $4 = 0, $5 = 0, $6_1 = 0, $7_1 = 0;
   __inlined_func$__wasm_ctz_i64 : {
-   if ($0 | $1_1) {
+   if ($1_1 | $0) {
     $4 = $1_1 - 1 | 0;
     $5 = $4 + 1 | 0;
     $6_1 = $4;