[FRONTEND] add option to disable fp mul/add fusion (#2495)

By default, ptxas will enable fusion of mul/add to fma instructions. The
backend was also being configured unconditionally to enable this on
conversion from LLVM IR to PTX. This commit adds an option which can be
used to disable the FP fusion behavior in both locations.

include/triton/Target/PTX/PTXTranslation.h

@@ -10,7 +10,8 @@ class Module;
 namespace triton {
 
 // Translate TritonGPU IR to PTX code.
-std::string translateLLVMIRToPTX(llvm::Module &module, int cc, int version);
+std::string translateLLVMIRToPTX(llvm::Module &module, int cc, int version,
+                                 bool enable_fp_fusion);
 
 } // namespace triton