On Mon, 12 Feb 2007, Reid Spencer wrote: >> - Result = new CallInst(F, Params); >> + Result = new CallInst(F, &Params[0], Params.size()); > > I don't see why this series of changes is better for performance.
It isn't. > Is there something I'm missing here? Nope, it's just an API cleanup. The perf improvements will happen when this change is complete. -Chris >> if (isTailCall) cast<CallInst>(Result)->setTailCall(); >> if (CallingConv) cast<CallInst>(Result)->setCallingConv(CallingConv); >> break; >> @@ -756,7 +756,7 @@ >> Params.push_back(getValue(Oprnds[i], Oprnds[i+1])); >> } >> >> - Result = new InvokeInst(F, Normal, Except, Params); >> + Result = new InvokeInst(F, Normal, Except, &Params[0], Params.size()); >> if (CallingConv) >> cast<InvokeInst>(Result)->setCallingConv(CallingConv); >> break; >> } >> >> >> >> _______________________________________________ >> llvm-commits mailing list >> llvm-commits@cs.uiuc.edu >> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits > -Chris -- http://nondot.org/sabre/ http://llvm.org/ _______________________________________________ llvm-commits mailing list llvm-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
