rustc_codegen_gcc: Progress Report #25 (
from to on 01 Sep 2023 05:56


threaded - newest on 01 Sep 2023 07:22 next collapse

As far as I know, adding the support for restrict didn’t trigger any bugs in GCC

That’s very impressive for gcc. IIRC adding restrict to LLVM triiggered major bugs and miscompilations at least for the first two attempts. As they said they need to do a crater run to be sure, but even passing the initial smoke test is an achievement for gcc.

However, I’m surprised the code is “only” 3% faster using restric annotation. IIRC the speed-ups were about 5% for LLVM so maybe there is still some performance to gain on the gcc side? on 01 Sep 2023 14:47 collapse

To be fair, I heard that some bugs for restrict fixed in LLVM were also fixed in GCC as a result of finding them using the LLVM codegen of rustc.

For the 3%, this is in only one particular benchmark. Since I was curious, I compiled my test program with the LLVM codegen and disabled noalias: there’s also a 3% performance improvement. on 01 Sep 2023 07:25 next collapse

The devil is really is the details. One year ago, I felt that the progress was moving so fast I had the feeling that it was close to being finished, but it seems toat it’s much more complex than what I thought. Nonetheless, congratulation to all contributors! on 01 Sep 2023 07:27 collapse

On reddit, someone was always linking to the donation page for antonio. I liked that tradition and hope that someone will continue it on lemmy. on 01 Sep 2023 07:51 collapse

Be the change you want to see?