[MLIR][CUDA-RUNNER] Add CL options to pass SM version and index-bitwidth

[navdeepkk]

Summary

Add a CL option to pass sm version in MLIR CUDA Runner. There are currently
no checks on the compatibility/validity of PTX and SM version. This is just naive addition
to check the working of WMMA tensor core operations. Add CL option to pass index-bitwidth
for LLVM lowering passes on the device side.

[navdeepkk]

Changes in this diff:-

1.) Add CL option to pass index-bitwidth for LLVM lowering passes
  on the device side.

[navdeepkk]

Closing revision, As recent changes(removal of mlir-cuda-runner and others) have baked the functionality of this patch into convert-gpu-to-nvvm and gpu-to-cubin.

mlir/tools/mlir-cuda-runner/mlir-cuda-runner.cpp

OwnedBlob result =
       std::make_unique<std::vector<char>>(cubinAsChar, cubinAsChar + cubinSize);
 
   // This will also destroy the cubin data.
   RETURN_ON_CUDA_ERROR(cuLinkDestroy(linkState), "cuLinkDestroy");
 
   return result;
 }
 
+static llvm::cl::opt<std::string>
+    clSMVersion("sm", llvm::cl::desc("SM version to target"),
+                llvm::cl::init("sm_35"));
+
+static llvm::cl::opt<unsigned>
+    clIndexWidth("index-bitwidth",
+                 llvm::cl::desc("Bitwidth of index type to use for lowering"),
+                 llvm::cl::init(32));
+
 static LogicalResult runMLIRPasses(ModuleOp m) {
   PassManager pm(m.getContext());
   applyPassManagerCLOptions(pm);
 
   const char gpuBinaryAnnotation[] = "nvvm.cubin";
   pm.addPass(createGpuKernelOutliningPass());
   auto &kernelPm = pm.nest<gpu::GPUModuleOp>();
   kernelPm.addPass(createStripDebugInfoPass());
-  kernelPm.addPass(createLowerGpuOpsToNVVMOpsPass());
+  kernelPm.addPass(createLowerGpuOpsToNVVMOpsPass(clIndexWidth.getValue()));
   kernelPm.addPass(createConvertGPUKernelToBlobPass(
       translateModuleToNVVMIR, compilePtxToCubin, "nvptx64-nvidia-cuda",
-      "sm_35", "+ptx60", gpuBinaryAnnotation));
+      clSMVersion.getValue(), "+ptx60", gpuBinaryAnnotation));
   pm.addPass(createGpuToLLVMConversionPass(gpuBinaryAnnotation));
 
   return pm.run(m);
 }
 
 int main(int argc, char **argv) {
   registerPassManagerCLOptions();
   llvm::InitLLVM y(argc, argv);