StringSwitch efficiency

Okay, now I'm just being lazy. What is the simplest way to compile a module and produce a code listing? I'm on Windows with Visual Studio.

Compiler Explorer can provide LLVM’s trunk libraries (choose it from “Libraries” in the compiler tab) eg: https://godbolt.org/z/5jEj3E

Holy cow! I did not know about Compiler Explorer. Thanks!

If the argument is unknown, it seems that SROA can essentially
optimize a long chain of bcmp(string comparison) calls to dispatches
on the argument length.
Then SimplfyCFGOpt::SimplifyBranchOnICmpChain can generate a switch
instruction which will eventually lower to a jump table.
ExpandMemCmp.cpp can generate efficient comparisons for these strings.
There are cases where multiple memory loads are produced while a
memory load can be reused,
but overall the produced code looks pretty good.

int foo(llvm::StringRef s) {
    return llvm::StringSwitch<int>(s)
      .Case("int", 5)
      .Case("bit", 3)
      .Case("bits", 7)
      .Case("string", 8)
      .Case("list", 4)
      .Case("code", 2)
      .Case("dag", 1)
      .Case("class", 4)
      .Case("def", 3)
      .Case("foreach", 9)
      .Case("defm", 1)
      .Case("defset", 0)
      .Case("multiclass", 1)
      .Case("field", 4)
      .Case("let", 3)
      .Case("in", 9)
      .Case("defvar", 1)
      .Case("if", 0)
      .Case("then", 3)
      .Case("else", 7)
      .Default(5);
}

I used Compiler Explorer to look at the compiled code and, indeed, it is quite good. The person who wrote that particular StringSwitch put the keywords in frequency-of-use order, so that the sequence of same-length comparisons is checking for the more common cases first.

Awesome, yay for composition and smart compilers :slight_smile:

-Chris