Function Dispatch Tables in C (2019)

Post by boffinAudio » Thu, Jun 23, 2022, 8:18 AM UTC

dzaima wrote: post #47

boffinAudio wrote: post #37

Earlier quoted context omitted.

The call-table functions also get inlined, so the advantage is shared by both approaches, and yet the call-table produces more efficient code (no cmp/jump traps to fall into...) See my comment here for the C code and Assembly that demonstrates this: https://news.ycombinator.com/item?id=31834241

I can't reproduce neither clang nor gcc with -O3 inlining functions in a call table: https://godbolt.org/z/E7Tj31vox Nor do I see any of that kind of behavior in your linked assembly - call_table clearly contains "call *%r8", which is an indirect call to a function, definitely not inlined. Here's a more complete test: https://godbolt.org/z/dT45aKTe1 - putting the operation in a loop (to be able to more clearly see th…

I'm using gcc under Ubuntu/WSL on Windows, so it may be some compiler thing.

But nevertheless, thank you for taking the time to follow up on this, because it is a very interesting subject to me personally and a lot has been learned through efforts such as yours, in this thread.

Post by dzaima » Thu, Jun 23, 2022, 9:55 AM UTC

boffinAudio wrote: post #51

dzaima wrote: post #47

Earlier quoted context omitted.

I can't reproduce neither clang nor gcc with -O3 inlining functions in a call table: https://godbolt.org/z/E7Tj31vox Nor do I see any of that kind of behavior in your linked assembly - call_table clearly contains "call *%r8", which is an indirect call to a function, definitely not inlined. Here's a more complete test: https://godbolt.org/z/dT45aKTe1 - putting the operation in a loop (to be able to more clearly see th…

I'm using gcc under Ubuntu/WSL on Windows, so it may be some compiler thing. But nevertheless, thank you for taking the time to follow up on this, because it is a very interesting subject to me personally and a lot has been learned through efforts such as yours, in this thread.

Windows/WSL shouldn't change anything, gcc will have the same set of optimizations everywhere. There's just no inlining of the functions going on. There may be differences in the dispatching between different compilers & optimization levels, but none inline functions from a function list.

It'd be a pretty messy optimization, as it'd need to check whether all functions can be inlined instead of just any single one (and increase the requirements if the call otherwise would be a tail-call, which cost a decent bit less), and it'd make it impossible to predictably do an actual function jump table if the compiler sometimes just didn't.

Re: Function Dispatch Tables in C (2019)

Re: Function Dispatch Tables in C (2019)