[llvm-mc] Add --runs option for benchmarking (#151149)

Add support for measuring decode times in llvm-mc tool. Add command line
options to enable time-trace profiling (similar to llc or opt) and add
option `runs` to run the decoder several times on each instruction.

Author: jurahul

llvm/test/tools/llvm-mc/disassembler-profile.test

@@ -0,0 +1,12 @@
+# REQUIRES: aarch64-registered-target
+# RUN: rm -rf %t.json
+# RUN: llvm-mc -triple=aarch64 -disassemble -o /dev/null %s -runs=1000 -time-trace -time-trace-file=%t.json
+# RUN: FileCheck --input-file %t.json %s
+
+# Note: Test input taken from llvm/test/MC/Disassembler/AArch64/udf.txt
+
+# CHECK: "name":"Total getInstruction"
+# CHECK: "args":{"count":3,"avg ms":{{.*}}}
+[0x00,0x00,0x00,0x00]
+[0x01,0x02,0x00,0x00]
+[0xff,0xff,0x00,0x00]

llvm/tools/llvm-mc/Disassembler.cpp

@@ -24,6 +24,7 @@
 #include "llvm/MC/TargetRegistry.h"
 #include "llvm/Support/MemoryBuffer.h"
 #include "llvm/Support/SourceMgr.h"
+#include "llvm/Support/TimeProfiler.h"
 #include "llvm/Support/raw_ostream.h"
 #include "llvm/TargetParser/Triple.h"
 
@@ -32,24 +33,29 @@ using namespace llvm;
 typedef std::pair<std::vector<unsigned char>, std::vector<const char *>>
     ByteArrayTy;
 
-static bool PrintInsts(const MCDisassembler &DisAsm, const ByteArrayTy &Bytes,
+static MCDisassembler::DecodeStatus getInstruction(const MCDisassembler &DisAsm,
+                                                   const MCSubtargetInfo &STI,
+                                                   MCInst &Inst, uint64_t &Size,
+                                                   ArrayRef<uint8_t> Bytes,
+                                                   uint64_t Address) {
+  if (STI.getTargetTriple().getArch() == Triple::hexagon)
+    return DisAsm.getInstructionBundle(Inst, Size, Bytes, Address, nulls());
+  return DisAsm.getInstruction(Inst, Size, Bytes, Address, nulls());
+}
+
+static bool printInsts(const MCDisassembler &DisAsm, const ByteArrayTy &Bytes,
                        SourceMgr &SM, MCStreamer &Streamer, bool InAtomicBlock,
-                       const MCSubtargetInfo &STI) {
+                       const MCSubtargetInfo &STI, unsigned NumBenchmarkRuns) {
   ArrayRef<uint8_t> Data(Bytes.first);
 
   // Disassemble it to strings.
   uint64_t Size;
-  uint64_t Index;
 
-  for (Index = 0; Index < Bytes.first.size(); Index += Size) {
-    MCInst Inst;
+  for (uint64_t Index = 0; Index < Bytes.first.size(); Index += Size) {
 
-    MCDisassembler::DecodeStatus S;
-    if (STI.getTargetTriple().getArch() == Triple::hexagon)
-      S = DisAsm.getInstructionBundle(Inst, Size, Data.slice(Index), Index,
-                                      nulls());
-    else
-      S = DisAsm.getInstruction(Inst, Size, Data.slice(Index), Index, nulls());
+    MCInst Inst;
+    MCDisassembler::DecodeStatus S =
+        getInstruction(DisAsm, STI, Inst, Size, Data.slice(Index), Index);
     switch (S) {
     case MCDisassembler::Fail:
       SM.PrintMessage(SMLoc::getFromPointer(Bytes.second[Index]),
@@ -74,6 +80,18 @@ static bool PrintInsts(const MCDisassembler &DisAsm, const ByteArrayTy &Bytes,
       Streamer.emitInstruction(Inst, STI);
       break;
     }
+
+    if (S == MCDisassembler::Success && NumBenchmarkRuns != 0) {
+      // Benchmark mode, collect timing for decoding the instruction several
+      // times.
+      MCInst BMInst;
+      TimeTraceScope timeScope("getInstruction");
+      for (unsigned I = 0; I < NumBenchmarkRuns; ++I) {
+        BMInst.clear();
+        BMInst.setOpcode(0);
+        S = getInstruction(DisAsm, STI, BMInst, Size, Data.slice(Index), Index);
+      }
+    }
   }
 
   return false;
@@ -151,7 +169,7 @@ int Disassembler::disassemble(const Target &T, const std::string &Triple,
                               MCSubtargetInfo &STI, MCStreamer &Streamer,
                               MemoryBuffer &Buffer, SourceMgr &SM,
                               MCContext &Ctx, const MCTargetOptions &MCOptions,
-                              bool HexBytes) {
+                              bool HexBytes, unsigned NumBenchmarkRuns) {
   std::unique_ptr<const MCRegisterInfo> MRI(T.createMCRegInfo(Triple));
   if (!MRI) {
     errs() << "error: no register info for target " << Triple << "\n";
@@ -207,8 +225,8 @@ int Disassembler::disassemble(const Target &T, const std::string &Triple,
     ErrorOccurred |= byteArrayFromString(ByteArray, Str, SM, HexBytes);
 
     if (!ByteArray.first.empty())
-      ErrorOccurred |=
-          PrintInsts(*DisAsm, ByteArray, SM, Streamer, InAtomicBlock, STI);
+      ErrorOccurred |= printInsts(*DisAsm, ByteArray, SM, Streamer,
+                                  InAtomicBlock, STI, NumBenchmarkRuns);
   }
 
   if (InAtomicBlock) {

llvm/tools/llvm-mc/Disassembler.h

@@ -32,7 +32,8 @@ class Disassembler {
   static int disassemble(const Target &T, const std::string &Triple,
                          MCSubtargetInfo &STI, MCStreamer &Streamer,
                          MemoryBuffer &Buffer, SourceMgr &SM, MCContext &Ctx,
-                         const MCTargetOptions &MCOptions, bool HexBytes);
+                         const MCTargetOptions &MCOptions, bool HexBytes,
+                         unsigned NumBenchmarkRuns);
 };
 
 } // namespace llvm

llvm/tools/llvm-mc/llvm-mc.cpp

@@ -12,6 +12,7 @@
 //===----------------------------------------------------------------------===//
 
 #include "Disassembler.h"
+#include "llvm/ADT/ScopeExit.h"
 #include "llvm/DWARFCFIChecker/DWARFCFIFunctionFrameAnalyzer.h"
 #include "llvm/DWARFCFIChecker/DWARFCFIFunctionFrameStreamer.h"
 #include "llvm/MC/MCAsmBackend.h"
@@ -37,6 +38,7 @@
 #include "llvm/Support/MemoryBuffer.h"
 #include "llvm/Support/SourceMgr.h"
 #include "llvm/Support/TargetSelect.h"
+#include "llvm/Support/TimeProfiler.h"
 #include "llvm/Support/ToolOutputFile.h"
 #include "llvm/Support/WithColor.h"
 #include "llvm/TargetParser/Host.h"
@@ -240,6 +242,23 @@ static cl::opt<ActionType> Action(
                           "Colored disassembly of strings of hex bytes")),
     cl::cat(MCCategory));
 
+static cl::opt<unsigned>
+    NumBenchmarkRuns("runs", cl::desc("Number of runs for benchmarking"),
+                     cl::cat(MCCategory));
+
+static cl::opt<bool> TimeTrace("time-trace", cl::desc("Record time trace"));
+
+static cl::opt<unsigned> TimeTraceGranularity(
+    "time-trace-granularity",
+    cl::desc(
+        "Minimum time granularity (in microseconds) traced by time profiler"),
+    cl::init(500), cl::Hidden);
+
+static cl::opt<std::string>
+    TimeTraceFile("time-trace-file",
+                  cl::desc("Specify time trace file destination"),
+                  cl::value_desc("filename"));
+
 static const Target *GetTarget(const char *ProgName) {
   // Figure out the target triple.
   if (TripleName.empty())
@@ -371,6 +390,20 @@ int main(int argc, char **argv) {
 
   cl::HideUnrelatedOptions({&MCCategory, &getColorCategory()});
   cl::ParseCommandLineOptions(argc, argv, "llvm machine code playground\n");
+
+  if (TimeTrace)
+    timeTraceProfilerInitialize(TimeTraceGranularity, argv[0]);
+
+  auto TimeTraceScopeExit = make_scope_exit([]() {
+    if (!TimeTrace)
+      return;
+    if (auto E = timeTraceProfilerWrite(TimeTraceFile, OutputFilename)) {
+      logAllUnhandledErrors(std::move(E), errs());
+      return;
+    }
+    timeTraceProfilerCleanup();
+  });
+
   MCTargetOptions MCOptions = mc::InitMCTargetOptionsFromFlags();
   MCOptions.CompressDebugSections = CompressDebugSections.getValue();
   MCOptions.ShowMCInst = ShowInst;
@@ -620,13 +653,15 @@ int main(int argc, char **argv) {
   }
   if (disassemble)
     Res = Disassembler::disassemble(*TheTarget, TripleName, *STI, *Str, *Buffer,
-                                    SrcMgr, Ctx, MCOptions, HexBytes);
+                                    SrcMgr, Ctx, MCOptions, HexBytes,
+                                    NumBenchmarkRuns);
 
   // Keep output if no errors.
   if (Res == 0) {
     Out->keep();
     if (DwoOut)
       DwoOut->keep();
   }
+
   return Res;
 }