Added optimizations for combining conditions together when their results are OR'ed

Added recognition of impossible comparisons via NZM and optimize them away
Recognize (x + -y) and transform to (x - y) for constants
Recognize (~x ) + 1 and transform to -x
Check and transform comparisons if theyre semantically equal to others
Detect comparisons of single-bit values with their only possible non-zero value and transform to true/false tests
Transform ==0 to IS_FALSE, !=0 to IS_TRUE
Truncate to int8 if operand for IS_TRUE/IS_FALSE has a nzm of 1
Reduced code generated for SubDidCarry slightly
Add special case for InstrEmit_srawix if mask == 1
Cut down the code generated for trap instructions, instead of naive or'ing or compare results do a switch and select the best condition
Rerun simplification pass until no changes, as some optimizations will enable others to be done
Enable rel32 call optimization by default

src/xenia/cpu/compiler/passes/context_promotion_pass.cc

@@ -21,6 +21,11 @@ DECLARE_bool(debug);
 DEFINE_bool(store_all_context_values, false,
             "Don't strip dead context stores to aid in debugging.", "CPU");
 
+DEFINE_bool(full_optimization_even_with_debug, false,
+            "For developer use to analyze the quality of the generated code, "
+            "not intended for actual debugging of the code",
+            "CPU");
+
 namespace xe {
 namespace cpu {
 namespace compiler {
@@ -77,7 +82,8 @@ bool ContextPromotionPass::Run(HIRBuilder* builder) {
   // Remove all dead stores.
   // This will break debugging as we can't recover this information when
   // trying to extract stack traces/register values, so we don't do that.
-  if (!cvars::debug && !cvars::store_all_context_values) {
+  if (cvars::full_optimization_even_with_debug ||
+      (!cvars::debug && !cvars::store_all_context_values)) {
     block = builder->first_block();
     while (block) {
       RemoveDeadStoresBlock(block);