[MLIR][GPU] Add debug output to enable dumping GPU assembly

- Set the DEBUG_TYPE of SerializeToBlob to serialize-to-blob
- Add debug output to print the assembly or PTX for GPU modules before
  they are assembled and linked

Note that, as SerializeToBlob is a superclass of SerializeToCubin and
SerializeToHsaco, --debug-only=serialize-to-blom will dump the
intermediate compiler result for both of these passes.

In addition, if LLVM options such as --stop-after are used to control
the GPU kernel compilation process, the debug output will contain the
appropriate intermediate IR.

Reviewed By: herhut

Differential Revision: https://reviews.llvm.org/D117519

mlir/include/mlir/Dialect/GPU/Passes.h

@@ -102,6 +102,13 @@ void registerGpuSerializeToCubinPass();
 /// annotation.
 void registerGpuSerializeToHsacoPass();
 
+/// Create an instance of the GPU kernel function to HSAco binary serialization
+/// pass.
+std::unique_ptr<Pass> createGpuSerializeToHsacoPass(StringRef triple,
+                                                    StringRef arch,
+                                                    StringRef features,
+                                                    int optLevel);
+
 /// Generate the code for registering passes.
 #define GEN_PASS_REGISTRATION
 #include "mlir/Dialect/GPU/Passes.h.inc"

mlir/lib/Dialect/GPU/Transforms/SerializeToBlob.cpp

@@ -21,6 +21,10 @@
 #include "llvm/Support/TargetSelect.h"
 #include "llvm/Target/TargetMachine.h"
 
+#include <string>
+
+#define DEBUG_TYPE "serialize-to-blob"
+
 using namespace mlir;
 
 std::string gpu::getDefaultGpuBinaryAnnotation() { return "gpu.binary"; }
@@ -76,6 +80,12 @@ void gpu::SerializeToBlobPass::runOnOperation() {
 
   std::string targetISA = std::move(maybeTargetISA.getValue());
 
+  LLVM_DEBUG({
+    llvm::dbgs() << "ISA for module: " << getOperation().getNameAttr() << "\n";
+    llvm::dbgs() << targetISA << "\n";
+    llvm::dbgs().flush();
+  });
+
   // Serialize the target ISA.
   std::unique_ptr<std::vector<char>> blob = serializeISA(targetISA);
   if (!blob)

mlir/lib/Dialect/GPU/Transforms/SerializeToHsaco.cpp

@@ -479,6 +479,17 @@ void mlir::registerGpuSerializeToHsacoPass() {
                                                       "", 2);
       });
 }
+
+/// Create an instance of the GPU kernel function to HSAco binary serialization
+/// pass.
+std::unique_ptr<Pass> mlir::createGpuSerializeToHsacoPass(StringRef triple,
+                                                          StringRef arch,
+                                                          StringRef features,
+                                                          int optLevel) {
+  return std::make_unique<SerializeToHsacoPass>(triple, arch, features,
+                                                optLevel);
+}
+
 #else  // MLIR_GPU_TO_HSACO_PASS_ENABLE
 void mlir::registerGpuSerializeToHsacoPass() {}
 #endif // MLIR_GPU_TO_HSACO_PASS_ENABLE