Introduce a pass to insert vzeroupper instructions to avoid AVX to

SSE transition penalty. The pass is enabled through the "x86-use-vzeroupper"
llc command line option. This is only the first step (very naive and
conservative one) to sketch out the idea, but proper DFA is coming next
to allow smarter decisions. Comments and ideas now and in further commits
will be very appreciated.

llvm-svn: 138317

1 files changed, 16 insertions, 0 deletions

diff --git a/llvm/lib/Target/X86/X86TargetMachine.cpp b/llvm/lib/Target/X86/X86TargetMachine.cpp
index 569c0408f30..95e7021dce5 100644
--- a/llvm/lib/Target/X86/X86TargetMachine.cpp
+++ b/llvm/lib/Target/X86/X86TargetMachine.cpp
@@ -16,6 +16,7 @@
 #include "llvm/PassManager.h"
 #include "llvm/CodeGen/MachineFunction.h"
 #include "llvm/CodeGen/Passes.h"
+#include "llvm/Support/CommandLine.h"
 #include "llvm/Support/FormattedStream.h"
 #include "llvm/Target/TargetOptions.h"
 #include "llvm/Target/TargetRegistry.h"
@@ -92,6 +93,16 @@ X86TargetMachine::X86TargetMachine(const Target &T, StringRef TT,
 }
 
 //===----------------------------------------------------------------------===//
+// Command line options for x86
+//===----------------------------------------------------------------------===//
+bool UseVZeroUpper;
+
+static cl::opt<bool, true>
+VZeroUpper("x86-use-vzeroupper",
+  cl::desc("Minimize AVX to SSE transition penalty"),
+  cl::location(UseVZeroUpper), cl::init(false));
+
+//===----------------------------------------------------------------------===//
 // Pass Pipeline Configuration
 //===----------------------------------------------------------------------===//
 
@@ -125,6 +136,11 @@ bool X86TargetMachine::addPreEmitPass(PassManagerBase &PM,
     PM.add(createSSEDomainFixPass());
     return true;
   }
+
+  if (Subtarget.hasAVX() && UseVZeroUpper) {
+    PM.add(createX86IssueVZeroUpperPass());
+    return true;
+  }
   return false;
 }