Hacker News new | past | comments | ask | show | jobs | submit login

Well, as I mentioned on a parallel thread, I know of a project taking around 3 hours to compile, that doesn't look very optimized to me.



You know of a project that takes 3 hours of time to preprocess with clang?

I have serious doubts. Overall compilation time is kind of irrelevant to this discussion, because Warp is just a preprocessor.

While you can get some X speedup to gcc by replacing the preprocessor, X, as a factor of overall compilation time, is usually 0.2-0.5 in most cases, depending on size of file.

I expect the gains warp gets over gcc overall from preprocessing to be similar to those clang gets over gcc overall from preprocessing.

(Though it depends on the size of files being compiled, etc).

Most companies that want actual fast overall compilation and have the resources, build caching distributed compilation infrastructure (Google, Facebook).

As mentioned, if warp is really that much faster than clang's preprocessor that it mattered, clang would be fixed :)


Faire enough, I meant full build times.


What is this even supposed to mean? Is the majority of that time spent preprocessing? How do other compilers perform? Your anecdote alone is absolutely useless.


What's stated doesn't say anything about optimization. It could be that it takes a long time to run processes that take a long time.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: