Hi Nikhil, While today's compilers are still largely very intricate code for Turing machines, this will certainly change during your career. It seems you'll be using GPUs for AI-assisted construction of optimal program graphs and immediately testing the performance of code fragments instead of relying on heuristics. The starting point for these technologies is polyhedral analysis - I recommend reading about it in addition to a classical compiler course.
Kind regards, Dmitry. пн, 12 мая 2025 г. в 16:31, Nikhil Patil via Gcc <gcc@gcc.gnu.org>: > Hi Richard, > > Thank you so much for the reply! > > You're absolutely right about using CPU threads. I’m just really curious > about whether GPU acceleration could somehow be explored for compilation, > even if it’s not traditionally well-suited. I know it might not be > practical, but I wanted to understand *why* it’s considered inefficient and > learn through experimentation. > > I’m still a student and just starting to explore compilers more seriously, > so any resources you could recommend to help me learn — especially from the > basics — would be incredibly helpful. > > Thanks again for taking the time to respond! > > > On Mon, 12 May 2025 at 18:51, Richard Biener <richard.guent...@gmail.com> > wrote: > > > On Mon, May 12, 2025 at 2:55 PM Nikhil Patil via Gcc <gcc@gcc.gnu.org> > > wrote: > > > > > > Hi GCC Team, > > > > > > I'm fairly new to the world of compilers and trying to understand how > > they > > > work in more depth. Recently, I started exploring the idea of > > *parallelizing > > > the internal steps of compilation* — such as parsing, code generation, > > and > > > optimization — instead of the usual serial approach, and potentially > > > leveraging *GPU acceleration* (like CUDA) for this. > > > > > > I was wondering if this concept has been explored in GCC, or if there > are > > > any existing resources, discussions, or directions I could look into? > > > > > > Apologies if this isn’t the right channel — I’m still getting familiar > > with > > > the community and couldn’t find another communication method. Please > let > > me > > > know if there’s a better place to ask such questions. > > > > There were two Google Summer of Code projects exploring parallel > > compilation using CPU threads. I don't think using a GPU is a good tool > > for any of the sub-problems a compiler solves. On the high-level GCC > > allows more efficient parallel compilation of multiple source file > projects > > when using LTO. > > > > Richard. > > > > > Best regards, > > > *Nikhil Patil* > > >
