Need a Backend Compiler

Question

I have created a compiler that produces an intermediate code. I do not have the time to write the backend for my project.

Is there any software I could use to evaluate the intermediate code produced? Where could I download this software?

The output looks something like this :

    t1 = 0.67596e-7
    sum = t1

    t1 = 2
    t2 = 3
    t3 = t2 + t1
    i = t3

L0:
    t1 = sum
    t2 = 20
t3 = compare(t1 <= t2)
    t4 = sum
    t5 = 12
t6 = compare(t4 ~= t5)
t7 = t3 | t6
    t8 = sum
    t9 = 20
t10 = compare(t8 > t9)
t11 = t7 & t10
if t11 true then goto L1 else goto L2
L1: 
    t1 = 2
    t2 = sum
    t3 = t2 + t1
    sum = t3

    t1 = 1
    t2 = i
    t3 = t2 + t1
    i = t3
    goto L0
L2:

Thanks for reading.

Answer 1

Your emitted code looks pretty close to C - whay not have your compiler emit C and use a C compiler as the backend - that's the way the original C++ cfront compiler worked.

Answer 2

Have a look at the llvm compiler infrastructure project. It is used in a lot of real world projects. The LLVM is a low level virtual machine, easy to create code for and easy to translate to native code.

Answer 3

Why not target another high level language as the intermediate output, for example you could generate C code, and use any C compiler to compile that. This has been used by a number of high level languages in the past.

Another slightly lower level option would be to generate IL code and use the .NET runtime. If the compiler is written in .NET you can use the Reflection.Emit namespace to generate a .NET assembly which can then be run in any .NET environment.

Answer 4

You may have better luck selecting a compiler stack and using the intermediate code supported by that stack. For example, gcc uses RTL, register transfer language as a common denominator for all languages supported by gcc. I suspect that .net has something similar that's used for things like IronPython.

If you want to use something that supplies a virtual machine for evaluating code, the SPIM MIPS vm supports the Motorola 68000 architecture and is really easy to work with.

Answer 5

NekoVM is designed to accept a high level IL similar to yours and then run it. Neil Butterworth's suggestion to use C is also a good choice.

Answer 6

If you can tolerate a dependency on the JVM, then it is pretty simple to generate byte-code using ASM.

If you can't, then gnu lightning can be used for on-the-fly code generation (smaller and simpler than LLVM, but without the optimisations).

Answer 7

Since your code is C-ish, you could try C--. It a C like language for compiler backends. IIRC, it's used by one of the major Haskell compilers.

Answer 8

Code generation is my business :-)

Comments on a few options:

• CLR:

  • Pro: industrial support
  • Con: you have to buy into their type system pretty much completely; depending on what you want to do with types, this may not matter
  • Con: Only Windows platform is really prime-time quality

• LLVM:

  • Pro: enthusiastic user community with charismatic leader
  • Pro: many interesting performance improvements
  • Con: somewhat complex interface
  • Con: history of holes in the engineering; as LLVM matures expect the holes in the engineering to be plugged by adding to the complexity of the interface

• C--

  • Pro: target is an actual written language, not an API; you can easily inspect, debug, and edit your C-- code
  • Pro: design is reasonably mature and reasonably clean
  • Pro: supports accurate garbage collection
  • Pro: most users report it is very easy to use
  • Con: very small development team
  • Con: as of early 2009, supports only three hardware platforms (x86, PPC, ARM)
  • Con: does not ship with a garbage collector
  • Con: future of project is uncertain

• C as target language

  • Pro: looks easy
  • Con: nearly impossible to get decent performance
  • Con: will drive you nuts in the long run; ask the long line of people who have tried to compile Haskell, ML, Modula-3, Scheme and more using this technique. At some point every one of these people gave up and built their own native code generator.

Summary: anything except C is a reasonable choice. For the best combination of flexibility, quality, and expected longevity, I'd probably recommend LLVM. But your example code is very close to C--, so that may be an advantage.

Full disclosure: I am affiliated with the C-- project.

Answer 9

I'd recommend spitting out Pascal code and using Free pascal.

It's natively compiled, compilation is blazing fast and you can target a plethora of platforms.