MergedLoadStoreMotion pass

Merges equivalent loads on both sides of a hammock/diamond
and hoists into into the header.
Merges equivalent stores on both sides of a hammock/diamond
and sinks it to the footer.
Can enable if conversion and tolerate better load misses
and store operand latencies.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213396 91177308-0d34-0410-b5e6-96231b3b80d8

bindings/ocaml/transforms/scalar/scalar_opts_ocaml.c

@@ -134,6 +134,12 @@ CAMLprim value llvm_add_tail_call_elimination(LLVMPassManagerRef PM) {
   return Val_unit;
 }
 
+/* [<Llvm.PassManager.any] Llvm.PassManager.t -> unit */
+CAMLprim value llvm_add_merged_load_store_motion(LLVMPassManagerRef PM) {
+  LLVMAddMergedLoadStoreMotionPass(PM);
+  return Val_unit;
+}
+
 /* [<Llvm.PassManager.any] Llvm.PassManager.t -> unit */
 CAMLprim value llvm_add_gvn(LLVMPassManagerRef PM) {
   LLVMAddGVNPass(PM);

include/llvm-c/Transforms/Scalar.h

@@ -44,6 +44,9 @@ void LLVMAddDeadStoreEliminationPass(LLVMPassManagerRef PM);
 /** See llvm::createScalarizerPass function. */
 void LLVMAddScalarizerPass(LLVMPassManagerRef PM);
 
+/** See llvm::createMergedLoadStoreMotionPass function. */
+void LLVMAddMergedLoadStoreMotionPass(LLVMPassManagerRef PM);
+
 /** See llvm::createGVNPass function. */
 void LLVMAddGVNPass(LLVMPassManagerRef PM);
 

include/llvm/InitializePasses.h

@@ -196,6 +196,7 @@ void initializeMachineVerifierPassPass(PassRegistry&);
 void initializeMemCpyOptPass(PassRegistry&);
 void initializeMemDepPrinterPass(PassRegistry&);
 void initializeMemoryDependenceAnalysisPass(PassRegistry&);
+void initializeMergedLoadStoreMotionPass(PassRegistry &);
 void initializeMetaRenamerPass(PassRegistry&);
 void initializeMergeFunctionsPass(PassRegistry&);
 void initializeModuleDebugInfoPrinterPass(PassRegistry&);

include/llvm/LinkAllPasses.h

@@ -134,6 +134,7 @@ namespace {
       (void) llvm::createConstantHoistingPass();
       (void) llvm::createCodeGenPreparePass();
       (void) llvm::createEarlyCSEPass();
+      (void)llvm::createMergedLoadStoreMotionPass();
       (void) llvm::createGVNPass();
       (void) llvm::createMemCpyOptPass();
       (void) llvm::createLoopDeletionPass();

include/llvm/Transforms/Scalar.h

@@ -286,6 +286,13 @@ extern char &LCSSAID;
 //
 FunctionPass *createEarlyCSEPass();
 
+//===----------------------------------------------------------------------===//
+//
+// MergedLoadStoreMotion - This pass merges loads and stores in diamonds. Loads
+// are hoisted into the header, while stores sink into the footer.
+//
+FunctionPass *createMergedLoadStoreMotionPass();
+
 //===----------------------------------------------------------------------===//
 //
 // GVN - This pass performs global value numbering and redundant load

lib/LTO/LTOCodeGenerator.cpp

@@ -107,6 +107,7 @@ void LTOCodeGenerator::initializeLTOPasses() {
   initializeFunctionAttrsPass(R);
   initializeGlobalsModRefPass(R);
   initializeLICMPass(R);
+  initializeMergedLoadStoreMotionPass(R);
   initializeGVNPass(R);
   initializeMemCpyOptPass(R);
   initializeDCEPass(R);

lib/Transforms/IPO/PassManagerBuilder.cpp

@@ -207,8 +207,10 @@ void PassManagerBuilder::populateModulePassManager(PassManagerBase &MPM) {
     MPM.add(createSimpleLoopUnrollPass());    // Unroll small loops
   addExtensionsToPM(EP_LoopOptimizerEnd, MPM);
 
-  if (OptLevel > 1)
+  if (OptLevel > 1) {
+    MPM.add(createMergedLoadStoreMotionPass()); // Merge load/stores in diamond
     MPM.add(createGVNPass());                 // Remove redundancies
+  }
   MPM.add(createMemCpyOptPass());             // Remove memcpy / form memset
   MPM.add(createSCCPPass());                  // Constant prop with SCCP
 
@@ -346,6 +348,7 @@ void PassManagerBuilder::populateLTOPassManager(PassManagerBase &PM,
   PM.add(createGlobalsModRefPass()); // IP alias analysis.
 
   PM.add(createLICMPass());                 // Hoist loop invariants.
+  PM.add(createMergedLoadStoreMotionPass()); // Merge load/stores in diamonds
   PM.add(createGVNPass(DisableGVNLoadPRE)); // Remove redundancies.
   PM.add(createMemCpyOptPass());            // Remove dead memcpys.
 

lib/Transforms/Scalar/CMakeLists.txt

@@ -22,6 +22,7 @@ add_llvm_library(LLVMScalarOpts
   LoopUnswitch.cpp
   LowerAtomic.cpp
   MemCpyOptimizer.cpp
+  MergedLoadStoreMotion.cpp
   PartiallyInlineLibCalls.cpp
   Reassociate.cpp
   Reg2Mem.cpp