[Triton-MLIR][BACKEND] Add elementwise ops and tests (#804)

Co-authored-by: Keren Zhou <kerenzhou@openai.com>

include/triton/Target/LLVMIR/LLVMIRTranslation.h

@@ -1,6 +1,7 @@
 #ifndef TRITON_TARGET_LLVMIRTRANSLATION_H
 #define TRITON_TARGET_LLVMIRTRANSLATION_H
 #include <memory>
+#include <vector>
 
 namespace llvm {
 class Module;
@@ -14,6 +15,11 @@ class ModuleOp;
 namespace mlir {
 namespace triton {
 
+// add external dependent libs
+void addExternalLibs(mlir::ModuleOp &module,
+                     const std::vector<std::string> &names,
+                     const std::vector<std::string> &paths);
+
 // Translate TritonGPU dialect to LLVMIR, return null if failed.
 std::unique_ptr<llvm::Module>
 translateTritonGPUToLLVMIR(llvm::LLVMContext *llvmContext,