[Triton-MLIR] tt.dot operands now must have DotOperand layout; also added prefetch pass prototype (#712)

Co-authored-by: Jokeren <kerenzhou@openai.com>
Co-authored-by: Phil Tillet <phil@openai.com>
Co-authored-by: Superjomn <yanchunwei@outlook.com>

python/src/triton.cc

@@ -1255,6 +1255,10 @@ void init_triton_ir(py::module &&m) {
            [](mlir::PassManager &self, int numStages) {
              self.addPass(mlir::createTritonGPUPipelinePass(numStages));
            })
+      .def("add_tritongpu_prefetch_pass",
+           [](mlir::PassManager &self) {
+             self.addPass(mlir::createTritonGPUPrefetchPass());
+           })
       .def("add_triton_gpu_combine_pass",
            [](mlir::PassManager &self) {
              self.addPass(mlir::createTritonGPUCombineOpsPass());