[MCA][NFCI] Minor changes to InstrBuilder and Instruction.

This is based on the assumption that most simulated instructions don't define
more than one or two registers. This is true for example on x86, where
most instruction definitions don't declare more than one register write.

The default code region size has been increased from 8 to 16. This is based on
the assumption that, for small microbenchmarks, the typical code snippet size is
often less than 16 instructions.

mca::Instruction now uses bitfields to pack flags.
No functional change intended.

include/llvm/MCA/InstrBuilder.h

@@ -63,7 +63,8 @@ public:
                const MCRegisterInfo &RI, const MCInstrAnalysis *IA);
 
   void clear() {
-    VariantDescriptors.shrink_and_clear();
+    Descriptors.clear();
+    VariantDescriptors.clear();
     FirstCallInst = true;
     FirstReturnInst = true;
   }

include/llvm/MCA/Instruction.h

@@ -346,7 +346,7 @@ struct ResourceUsage {
 
 /// An instruction descriptor
 struct InstrDesc {
-  SmallVector<WriteDescriptor, 4> Writes; // Implicit writes are at the end.
+  SmallVector<WriteDescriptor, 2> Writes; // Implicit writes are at the end.
   SmallVector<ReadDescriptor, 4> Reads;   // Implicit reads are at the end.
 
   // For every resource used by an instruction of this kind, this vector
@@ -370,16 +370,16 @@ struct InstrDesc {
   // subtarget when computing the reciprocal throughput.
   unsigned SchedClassID;
 
-  bool MayLoad;
+  unsigned MayLoad : 1;
-  bool MayStore;
+  unsigned MayStore : 1;
-  bool HasSideEffects;
+  unsigned HasSideEffects : 1;
-  bool BeginGroup;
+  unsigned BeginGroup : 1;
-  bool EndGroup;
+  unsigned EndGroup : 1;
-  bool RetireOOO;
+  unsigned RetireOOO : 1;
 
   // True if all buffered resources are in-order, and there is at least one
   // buffer which is a dispatch hazard (BufferSize = 0).
-  bool MustIssueImmediately;
+  unsigned MustIssueImmediately : 1;
 
   // A zero latency instruction doesn't consume any scheduler resources.
   bool isZeroLatency() const { return !MaxLatency && Resources.empty(); }
@@ -403,7 +403,7 @@ class InstructionBase {
 
   // Output dependencies.
   // One entry per each implicit and explicit register definition.
-  SmallVector<WriteState, 4> Defs;
+  SmallVector<WriteState, 2> Defs;
 
   // Input dependencies.
   // One entry per each implicit and explicit register use.

lib/MCA/CodeEmitter.cpp

@@ -15,8 +15,7 @@
 namespace llvm {
 namespace mca {
 
-CodeEmitter::EncodingInfo
+CodeEmitter::EncodingInfo CodeEmitter::getOrCreateEncodingInfo(unsigned MCID) {
-CodeEmitter::getOrCreateEncodingInfo(unsigned MCID) {
   EncodingInfo &EI = Encodings[MCID];
   if (EI.second)
     return EI;

lib/MCA/Context.cpp

@@ -44,8 +44,8 @@ Context::createDefaultPipeline(const PipelineOptions &Opts, SourceMgr &SrcMgr) {
 
   // Create the pipeline stages.
   auto Fetch = std::make_unique<EntryStage>(SrcMgr);
-  auto Dispatch = std::make_unique<DispatchStage>(STI, MRI, Opts.DispatchWidth,
+  auto Dispatch =
-                                                   *RCU, *PRF);
+      std::make_unique<DispatchStage>(STI, MRI, Opts.DispatchWidth, *RCU, *PRF);
   auto Execute =
       std::make_unique<ExecuteStage>(*HWS, Opts.EnableBottleneckAnalysis);
   auto Retire = std::make_unique<RetireStage>(*RCU, *PRF, *LSU);

tools/llvm-mca/CodeRegion.h

@@ -53,7 +53,7 @@ class CodeRegion {
   // An optional descriptor for this region.
   llvm::StringRef Description;
   // Instructions that form this region.
-  llvm::SmallVector<llvm::MCInst, 8> Instructions;
+  llvm::SmallVector<llvm::MCInst, 16> Instructions;
   // Source location range.
   llvm::SMLoc RangeStart;
   llvm::SMLoc RangeEnd;