[AMDGPU] Enable scheduler mfma rewrite stage by default (#180751)

After performance testing, it was determined that the large number of
copies that are inserted outside the loop are more than offset by better
allocation within the loop as a result of the rewrite. Additionally, there is a
minor cleanup of the cost logic.

---------

Co-authored-by: Tony Linthicum <tlinthic@gmail.com>

llvm/lib/Target/AMDGPU/GCNSchedStrategy.cpp

@@ -101,7 +101,7 @@ static cl::opt<bool> PrintMaxRPRegUsageAfterScheduler(
 
 static cl::opt<bool> DisableRewriteMFMAFormSchedStage(
     "amdgpu-disable-rewrite-mfma-form-sched-stage", cl::Hidden,
-    cl::desc("Disable rewrie mfma rewrite scheduling stage"), cl::init(true));
+    cl::desc("Disable rewrite mfma rewrite scheduling stage"), cl::init(false));
 
 const unsigned ScheduleMetrics::ScaleFactor = 100;